From f8f67a788e4c8dc41b59d6f22631172fb4a431df Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Thu, 25 Jun 2020 19:55:45 +0300 Subject: [PATCH 001/121] allow to turn on fsync on inserts and merges --- src/Disks/DiskLocal.cpp | 17 +++++++- src/Disks/DiskLocal.h | 2 + src/Disks/DiskMemory.cpp | 5 +++ src/Disks/DiskMemory.h | 2 + src/Disks/IDisk.h | 3 ++ src/Disks/S3/DiskS3.cpp | 5 +++ src/Disks/S3/DiskS3.h | 2 + .../MergeTree/IMergeTreeDataPartWriter.cpp | 11 +++-- .../MergeTree/IMergeTreeDataPartWriter.h | 6 +-- .../MergeTree/MergeTreeDataMergerMutator.cpp | 41 +++++++++++++------ .../MergeTree/MergeTreeDataMergerMutator.h | 6 ++- .../MergeTreeDataPartWriterCompact.cpp | 4 +- .../MergeTreeDataPartWriterCompact.h | 2 +- .../MergeTree/MergeTreeDataPartWriterWide.cpp | 4 +- .../MergeTree/MergeTreeDataPartWriterWide.h | 2 +- .../MergeTree/MergeTreeDataWriter.cpp | 7 +++- src/Storages/MergeTree/MergeTreeSettings.h | 3 ++ .../MergeTree/MergedBlockOutputStream.cpp | 7 ++-- .../MergeTree/MergedBlockOutputStream.h | 1 + .../MergedColumnOnlyOutputStream.cpp | 9 ++-- .../MergeTree/MergedColumnOnlyOutputStream.h | 2 +- 21 files changed, 108 insertions(+), 33 deletions(-) diff --git a/src/Disks/DiskLocal.cpp b/src/Disks/DiskLocal.cpp index 68f5ee99a7a..c67bac7ffe2 100644 --- a/src/Disks/DiskLocal.cpp +++ b/src/Disks/DiskLocal.cpp @@ -8,7 +8,7 @@ #include #include - +#include namespace DB { @@ -19,6 +19,9 @@ namespace ErrorCodes extern const int EXCESSIVE_ELEMENT_IN_CONFIG; extern const int PATH_ACCESS_DENIED; extern const int INCORRECT_DISK_INDEX; + extern const int FILE_DOESNT_EXIST; + extern const int CANNOT_OPEN_FILE; + extern const int CANNOT_FSYNC; } std::mutex DiskLocal::reservation_mutex; @@ -188,6 +191,18 @@ void DiskLocal::moveDirectory(const String & from_path, const String & to_path) Poco::File(disk_path + from_path).renameTo(disk_path + to_path); } +void DiskLocal::sync(const String & path) const +{ + String full_path = disk_path + path; + int fd = ::open(full_path.c_str(), O_RDONLY); + if (-1 == fd) + throwFromErrnoWithPath("Cannot open file " + full_path, full_path, + errno == ENOENT ? ErrorCodes::FILE_DOESNT_EXIST : ErrorCodes::CANNOT_OPEN_FILE); + + if (-1 == fsync(fd)) + throwFromErrnoWithPath("Cannot fsync " + full_path, full_path, ErrorCodes::CANNOT_FSYNC); +} + DiskDirectoryIteratorPtr DiskLocal::iterateDirectory(const String & path) { return std::make_unique(disk_path, path); diff --git a/src/Disks/DiskLocal.h b/src/Disks/DiskLocal.h index 61a3994b655..743ba2ceb10 100644 --- a/src/Disks/DiskLocal.h +++ b/src/Disks/DiskLocal.h @@ -59,6 +59,8 @@ public: void moveDirectory(const String & from_path, const String & to_path) override; + void sync(const String & path) const override; + DiskDirectoryIteratorPtr iterateDirectory(const String & path) override; void createFile(const String & path) override; diff --git a/src/Disks/DiskMemory.cpp b/src/Disks/DiskMemory.cpp index 3e43d159ba5..5b3350e40f7 100644 --- a/src/Disks/DiskMemory.cpp +++ b/src/Disks/DiskMemory.cpp @@ -261,6 +261,11 @@ void DiskMemory::moveDirectory(const String & /*from_path*/, const String & /*to throw Exception("Method moveDirectory is not implemented for memory disks", ErrorCodes::NOT_IMPLEMENTED); } +void DiskMemory::sync(const String & /*path*/) const +{ + throw Exception("Method sync is not implemented for memory disks", ErrorCodes::NOT_IMPLEMENTED); +} + DiskDirectoryIteratorPtr DiskMemory::iterateDirectory(const String & path) { std::lock_guard lock(mutex); diff --git a/src/Disks/DiskMemory.h b/src/Disks/DiskMemory.h index b0c1d30c61d..8a3ddf05aa7 100644 --- a/src/Disks/DiskMemory.h +++ b/src/Disks/DiskMemory.h @@ -52,6 +52,8 @@ public: void moveDirectory(const String & from_path, const String & to_path) override; + void sync(const String & path) const override; + DiskDirectoryIteratorPtr iterateDirectory(const String & path) override; void createFile(const String & path) override; diff --git a/src/Disks/IDisk.h b/src/Disks/IDisk.h index 011c75402f4..8de77a560d1 100644 --- a/src/Disks/IDisk.h +++ b/src/Disks/IDisk.h @@ -105,6 +105,9 @@ public: /// Move directory from `from_path` to `to_path`. virtual void moveDirectory(const String & from_path, const String & to_path) = 0; + /// Do fsync on directory. + virtual void sync(const String & path) const = 0; + /// Return iterator to the contents of the specified directory. virtual DiskDirectoryIteratorPtr iterateDirectory(const String & path) = 0; diff --git a/src/Disks/S3/DiskS3.cpp b/src/Disks/S3/DiskS3.cpp index 71b5991f770..292f6567df4 100644 --- a/src/Disks/S3/DiskS3.cpp +++ b/src/Disks/S3/DiskS3.cpp @@ -466,6 +466,11 @@ void DiskS3::clearDirectory(const String & path) remove(it->path()); } +void DiskS3::sync(const String & /*path*/) const +{ + throw Exception("Method sync is not implemented for S3 disks", ErrorCodes::NOT_IMPLEMENTED); +} + void DiskS3::moveFile(const String & from_path, const String & to_path) { if (exists(to_path)) diff --git a/src/Disks/S3/DiskS3.h b/src/Disks/S3/DiskS3.h index 5fa8e8358a6..09132367ae8 100644 --- a/src/Disks/S3/DiskS3.h +++ b/src/Disks/S3/DiskS3.h @@ -58,6 +58,8 @@ public: void moveDirectory(const String & from_path, const String & to_path) override { moveFile(from_path, to_path); } + void sync(const String & path) const override; + DiskDirectoryIteratorPtr iterateDirectory(const String & path) override; void moveFile(const String & from_path, const String & to_path) override; diff --git a/src/Storages/MergeTree/IMergeTreeDataPartWriter.cpp b/src/Storages/MergeTree/IMergeTreeDataPartWriter.cpp index 73ac7fc0064..03ae2166504 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPartWriter.cpp +++ b/src/Storages/MergeTree/IMergeTreeDataPartWriter.cpp @@ -308,7 +308,8 @@ void IMergeTreeDataPartWriter::calculateAndSerializeSkipIndices( skip_index_data_mark = skip_index_current_data_mark; } -void IMergeTreeDataPartWriter::finishPrimaryIndexSerialization(MergeTreeData::DataPart::Checksums & checksums) +void IMergeTreeDataPartWriter::finishPrimaryIndexSerialization( + MergeTreeData::DataPart::Checksums & checksums, bool sync) { bool write_final_mark = (with_final_mark && data_written); if (write_final_mark && compute_granularity) @@ -330,12 +331,14 @@ void IMergeTreeDataPartWriter::finishPrimaryIndexSerialization(MergeTreeData::Da index_stream->next(); checksums.files["primary.idx"].file_size = index_stream->count(); checksums.files["primary.idx"].file_hash = index_stream->getHash(); - index_stream = nullptr; + if (sync) + index_stream->sync(); + index_stream.reset(); } } void IMergeTreeDataPartWriter::finishSkipIndicesSerialization( - MergeTreeData::DataPart::Checksums & checksums) + MergeTreeData::DataPart::Checksums & checksums, bool sync) { for (size_t i = 0; i < skip_indices.size(); ++i) { @@ -348,6 +351,8 @@ void IMergeTreeDataPartWriter::finishSkipIndicesSerialization( { stream->finalize(); stream->addToChecksums(checksums); + if (sync) + stream->sync(); } skip_indices_streams.clear(); diff --git a/src/Storages/MergeTree/IMergeTreeDataPartWriter.h b/src/Storages/MergeTree/IMergeTreeDataPartWriter.h index 2f849e7c895..eebdb880a66 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPartWriter.h +++ b/src/Storages/MergeTree/IMergeTreeDataPartWriter.h @@ -102,9 +102,9 @@ public: void initSkipIndices(); void initPrimaryIndex(); - virtual void finishDataSerialization(IMergeTreeDataPart::Checksums & checksums) = 0; - void finishPrimaryIndexSerialization(MergeTreeData::DataPart::Checksums & checksums); - void finishSkipIndicesSerialization(MergeTreeData::DataPart::Checksums & checksums); + virtual void finishDataSerialization(IMergeTreeDataPart::Checksums & checksums, bool sync) = 0; + void finishPrimaryIndexSerialization(MergeTreeData::DataPart::Checksums & checksums, bool sync); + void finishSkipIndicesSerialization(MergeTreeData::DataPart::Checksums & checksum, bool sync); protected: /// Count index_granularity for block and store in `index_granularity` diff --git a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp index 00830dd78c2..ccd7f234925 100644 --- a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp +++ b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp @@ -576,6 +576,13 @@ public: } }; +static bool needSyncPart(const size_t input_rows, size_t input_bytes, const MergeTreeSettings & settings) +{ + return ((settings.min_rows_to_sync_after_merge && input_rows >= settings.min_rows_to_sync_after_merge) + || (settings.min_compressed_bytes_to_sync_after_merge && input_bytes >= settings.min_compressed_bytes_to_sync_after_merge)); +} + + /// parts should be sorted. MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTemporaryPart( const FutureMergedMutatedPart & future_part, @@ -648,6 +655,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTempor } size_t sum_input_rows_upper_bound = merge_entry->total_rows_count; + size_t sum_compressed_bytes_upper_bound = merge_entry->total_size_bytes_compressed; MergeAlgorithm merge_alg = chooseMergeAlgorithm(parts, sum_input_rows_upper_bound, gathering_columns, deduplicate, need_remove_expired_values); LOG_DEBUG(log, "Selected MergeAlgorithm: {}", ((merge_alg == MergeAlgorithm::Vertical) ? "Vertical" : "Horizontal")); @@ -803,7 +811,6 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTempor if (need_remove_expired_values) merged_stream = std::make_shared(merged_stream, data, metadata_snapshot, new_data_part, time_of_merge, force_ttl); - if (metadata_snapshot->hasSecondaryIndices()) { const auto & indices = metadata_snapshot->getSecondaryIndices(); @@ -863,6 +870,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTempor if (need_remove_expired_values && ttl_merges_blocker.isCancelled()) throw Exception("Cancelled merging parts with expired TTL", ErrorCodes::ABORTED); + bool need_sync = needSyncPart(sum_input_rows_upper_bound, sum_compressed_bytes_upper_bound, *data_settings); MergeTreeData::DataPart::Checksums checksums_gathered_columns; /// Gather ordinary columns @@ -942,7 +950,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTempor throw Exception("Cancelled merging parts", ErrorCodes::ABORTED); column_gathered_stream.readSuffix(); - auto changed_checksums = column_to.writeSuffixAndGetChecksums(new_data_part, checksums_gathered_columns); + auto changed_checksums = column_to.writeSuffixAndGetChecksums(new_data_part, checksums_gathered_columns, need_sync); checksums_gathered_columns.add(std::move(changed_checksums)); if (rows_written != column_elems_written) @@ -979,9 +987,12 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTempor } if (merge_alg != MergeAlgorithm::Vertical) - to.writeSuffixAndFinalizePart(new_data_part); + to.writeSuffixAndFinalizePart(new_data_part, need_sync); else - to.writeSuffixAndFinalizePart(new_data_part, &storage_columns, &checksums_gathered_columns); + to.writeSuffixAndFinalizePart(new_data_part, need_sync, &storage_columns, &checksums_gathered_columns); + + if (need_sync) + new_data_part->volume->getDisk()->sync(new_part_tmp_path); return new_data_part; } @@ -1081,7 +1092,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mutatePartToTempor /// Don't change granularity type while mutating subset of columns auto mrk_extension = source_part->index_granularity_info.is_adaptive ? getAdaptiveMrkExtension(new_data_part->getType()) : getNonAdaptiveMrkExtension(); - + bool need_sync = needSyncPart(source_part->rows_count, source_part->getBytesOnDisk(), *data_settings); bool need_remove_expired_values = false; if (in && shouldExecuteTTL(metadata_snapshot, in->getHeader().getNamesAndTypesList().getNames(), commands_for_part)) @@ -1099,7 +1110,8 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mutatePartToTempor time_of_mutation, compression_codec, merge_entry, - need_remove_expired_values); + need_remove_expired_values, + need_sync); /// no finalization required, because mutateAllPartColumns use /// MergedBlockOutputStream which finilaze all part fields itself @@ -1154,7 +1166,8 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mutatePartToTempor time_of_mutation, compression_codec, merge_entry, - need_remove_expired_values); + need_remove_expired_values, + need_sync); } for (const auto & [rename_from, rename_to] : files_to_rename) @@ -1174,6 +1187,9 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mutatePartToTempor finalizeMutatedPart(source_part, new_data_part, need_remove_expired_values); } + if (need_sync) + new_data_part->volume->getDisk()->sync(new_part_tmp_path); + return new_data_part; } @@ -1599,7 +1615,8 @@ void MergeTreeDataMergerMutator::mutateAllPartColumns( time_t time_of_mutation, const CompressionCodecPtr & compression_codec, MergeListEntry & merge_entry, - bool need_remove_expired_values) const + bool need_remove_expired_values, + bool need_sync) const { if (mutating_stream == nullptr) throw Exception("Cannot mutate part columns with uninitialized mutations stream. It's a bug", ErrorCodes::LOGICAL_ERROR); @@ -1637,7 +1654,7 @@ void MergeTreeDataMergerMutator::mutateAllPartColumns( new_data_part->minmax_idx = std::move(minmax_idx); mutating_stream->readSuffix(); - out.writeSuffixAndFinalizePart(new_data_part); + out.writeSuffixAndFinalizePart(new_data_part, need_sync); } void MergeTreeDataMergerMutator::mutateSomePartColumns( @@ -1650,7 +1667,8 @@ void MergeTreeDataMergerMutator::mutateSomePartColumns( time_t time_of_mutation, const CompressionCodecPtr & compression_codec, MergeListEntry & merge_entry, - bool need_remove_expired_values) const + bool need_remove_expired_values, + bool need_sync) const { if (mutating_stream == nullptr) throw Exception("Cannot mutate part columns with uninitialized mutations stream. It's a bug", ErrorCodes::LOGICAL_ERROR); @@ -1684,10 +1702,9 @@ void MergeTreeDataMergerMutator::mutateSomePartColumns( mutating_stream->readSuffix(); - auto changed_checksums = out.writeSuffixAndGetChecksums(new_data_part, new_data_part->checksums); + auto changed_checksums = out.writeSuffixAndGetChecksums(new_data_part, new_data_part->checksums, need_sync); new_data_part->checksums.add(std::move(changed_checksums)); - } void MergeTreeDataMergerMutator::finalizeMutatedPart( diff --git a/src/Storages/MergeTree/MergeTreeDataMergerMutator.h b/src/Storages/MergeTree/MergeTreeDataMergerMutator.h index 121cc770d51..23b8d7f681b 100644 --- a/src/Storages/MergeTree/MergeTreeDataMergerMutator.h +++ b/src/Storages/MergeTree/MergeTreeDataMergerMutator.h @@ -189,7 +189,8 @@ private: time_t time_of_mutation, const CompressionCodecPtr & codec, MergeListEntry & merge_entry, - bool need_remove_expired_values) const; + bool need_remove_expired_values, + bool need_sync) const; /// Mutate some columns of source part with mutation_stream void mutateSomePartColumns( @@ -202,7 +203,8 @@ private: time_t time_of_mutation, const CompressionCodecPtr & codec, MergeListEntry & merge_entry, - bool need_remove_expired_values) const; + bool need_remove_expired_values, + bool need_sync) const; /// Initialize and write to disk new part fields like checksums, columns, /// etc. diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp index f7a3ad75cf5..79800204a3b 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp @@ -141,7 +141,7 @@ void MergeTreeDataPartWriterCompact::writeColumnSingleGranule(const ColumnWithTy column.type->serializeBinaryBulkStateSuffix(serialize_settings, state); } -void MergeTreeDataPartWriterCompact::finishDataSerialization(IMergeTreeDataPart::Checksums & checksums) +void MergeTreeDataPartWriterCompact::finishDataSerialization(IMergeTreeDataPart::Checksums & checksums, bool sync) { if (columns_buffer.size() != 0) writeBlock(header.cloneWithColumns(columns_buffer.releaseColumns())); @@ -158,6 +158,8 @@ void MergeTreeDataPartWriterCompact::finishDataSerialization(IMergeTreeDataPart: stream->finalize(); stream->addToChecksums(checksums); + if (sync) + stream->sync(); stream.reset(); } diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.h b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.h index 8183c038c4c..dde7deafc58 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.h +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.h @@ -20,7 +20,7 @@ public: void write(const Block & block, const IColumn::Permutation * permutation, const Block & primary_key_block, const Block & skip_indexes_block) override; - void finishDataSerialization(IMergeTreeDataPart::Checksums & checksums) override; + void finishDataSerialization(IMergeTreeDataPart::Checksums & checksums, bool sync) override; protected: void fillIndexGranularity(size_t index_granularity_for_block, size_t rows_in_block) override; diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp index e71ea4d4b94..fcd0249b10c 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp @@ -264,7 +264,7 @@ void MergeTreeDataPartWriterWide::writeColumn( next_index_offset = current_row - total_rows; } -void MergeTreeDataPartWriterWide::finishDataSerialization(IMergeTreeDataPart::Checksums & checksums) +void MergeTreeDataPartWriterWide::finishDataSerialization(IMergeTreeDataPart::Checksums & checksums, bool sync) { const auto & global_settings = storage.global_context.getSettingsRef(); IDataType::SerializeBinaryBulkSettings serialize_settings; @@ -295,6 +295,8 @@ void MergeTreeDataPartWriterWide::finishDataSerialization(IMergeTreeDataPart::Ch { stream.second->finalize(); stream.second->addToChecksums(checksums); + if (sync) + stream.second->sync(); } column_streams.clear(); diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.h b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.h index f5a9d17f63c..4286065a3ca 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.h +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.h @@ -23,7 +23,7 @@ public: void write(const Block & block, const IColumn::Permutation * permutation, const Block & primary_key_block, const Block & skip_indexes_block) override; - void finishDataSerialization(IMergeTreeDataPart::Checksums & checksums) override; + void finishDataSerialization(IMergeTreeDataPart::Checksums & checksums, bool sync) override; IDataType::OutputStreamGetter createStreamGetter(const String & name, WrittenOffsetColumns & offset_columns); diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.cpp b/src/Storages/MergeTree/MergeTreeDataWriter.cpp index 099480aca2f..cf8860b7f04 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.cpp +++ b/src/Storages/MergeTree/MergeTreeDataWriter.cpp @@ -303,10 +303,15 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataWriter::writeTempPart(BlockWithPa const auto & index_factory = MergeTreeIndexFactory::instance(); MergedBlockOutputStream out(new_data_part, metadata_snapshot, columns, index_factory.getMany(metadata_snapshot->getSecondaryIndices()), compression_codec); + bool sync_on_insert = data.getSettings()->sync_after_insert; out.writePrefix(); out.writeWithPermutation(block, perm_ptr); - out.writeSuffixAndFinalizePart(new_data_part); + out.writeSuffixAndFinalizePart(new_data_part, sync_on_insert); + + /// Sync part directory. + if (sync_on_insert) + new_data_part->volume->getDisk()->sync(full_path); ProfileEvents::increment(ProfileEvents::MergeTreeDataWriterRows, block.rows()); ProfileEvents::increment(ProfileEvents::MergeTreeDataWriterUncompressedBytes, block.bytes()); diff --git a/src/Storages/MergeTree/MergeTreeSettings.h b/src/Storages/MergeTree/MergeTreeSettings.h index f2d2a7cc3d4..da2c9ee49ee 100644 --- a/src/Storages/MergeTree/MergeTreeSettings.h +++ b/src/Storages/MergeTree/MergeTreeSettings.h @@ -43,6 +43,9 @@ struct MergeTreeSettings : public SettingsCollection M(SettingSeconds, old_parts_lifetime, 8 * 60, "How many seconds to keep obsolete parts.", 0) \ M(SettingSeconds, temporary_directories_lifetime, 86400, "How many seconds to keep tmp_-directories.", 0) \ M(SettingSeconds, lock_acquire_timeout_for_background_operations, DBMS_DEFAULT_LOCK_ACQUIRE_TIMEOUT_SEC, "For background operations like merges, mutations etc. How many seconds before failing to acquire table locks.", 0) \ + M(SettingUInt64, min_rows_to_sync_after_merge, 0, "Minimal number of rows to do fsync for part after merge (0 - disabled)", 0) \ + M(SettingUInt64, min_compressed_bytes_to_sync_after_merge, 0, "Minimal number of compressed bytes to do fsync for part after merge (0 - disabled)", 0) \ + M(SettingBool, sync_after_insert, false, "Do fsync for every inserted part. Significantly decreases performance of inserts, not recommended to use with wide parts.", 0) \ \ /** Inserts settings. */ \ M(SettingUInt64, parts_to_delay_insert, 150, "If table contains at least that many active parts in single partition, artificially slow down insert into table.", 0) \ diff --git a/src/Storages/MergeTree/MergedBlockOutputStream.cpp b/src/Storages/MergeTree/MergedBlockOutputStream.cpp index e776a35f21f..5e15084aa7d 100644 --- a/src/Storages/MergeTree/MergedBlockOutputStream.cpp +++ b/src/Storages/MergeTree/MergedBlockOutputStream.cpp @@ -85,6 +85,7 @@ void MergedBlockOutputStream::writeSuffix() void MergedBlockOutputStream::writeSuffixAndFinalizePart( MergeTreeData::MutableDataPartPtr & new_part, + bool sync, const NamesAndTypesList * total_columns_list, MergeTreeData::DataPart::Checksums * additional_column_checksums) { @@ -95,9 +96,9 @@ void MergedBlockOutputStream::writeSuffixAndFinalizePart( checksums = std::move(*additional_column_checksums); /// Finish columns serialization. - writer->finishDataSerialization(checksums); - writer->finishPrimaryIndexSerialization(checksums); - writer->finishSkipIndicesSerialization(checksums); + writer->finishDataSerialization(checksums, sync); + writer->finishPrimaryIndexSerialization(checksums, sync); + writer->finishSkipIndicesSerialization(checksums, sync); NamesAndTypesList part_columns; if (!total_columns_list) diff --git a/src/Storages/MergeTree/MergedBlockOutputStream.h b/src/Storages/MergeTree/MergedBlockOutputStream.h index 1a8bf9da822..002ef78a9af 100644 --- a/src/Storages/MergeTree/MergedBlockOutputStream.h +++ b/src/Storages/MergeTree/MergedBlockOutputStream.h @@ -46,6 +46,7 @@ public: /// Finilize writing part and fill inner structures void writeSuffixAndFinalizePart( MergeTreeData::MutableDataPartPtr & new_part, + bool sync = false, const NamesAndTypesList * total_columns_list = nullptr, MergeTreeData::DataPart::Checksums * additional_column_checksums = nullptr); diff --git a/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp b/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp index 1faadd0d720..e767fb3f155 100644 --- a/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp +++ b/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp @@ -63,12 +63,15 @@ void MergedColumnOnlyOutputStream::writeSuffix() } MergeTreeData::DataPart::Checksums -MergedColumnOnlyOutputStream::writeSuffixAndGetChecksums(MergeTreeData::MutableDataPartPtr & new_part, MergeTreeData::DataPart::Checksums & all_checksums) +MergedColumnOnlyOutputStream::writeSuffixAndGetChecksums( + MergeTreeData::MutableDataPartPtr & new_part, + MergeTreeData::DataPart::Checksums & all_checksums, + bool sync) { /// Finish columns serialization. MergeTreeData::DataPart::Checksums checksums; - writer->finishDataSerialization(checksums); - writer->finishSkipIndicesSerialization(checksums); + writer->finishDataSerialization(checksums, sync); + writer->finishSkipIndicesSerialization(checksums, sync); auto columns = new_part->getColumns(); diff --git a/src/Storages/MergeTree/MergedColumnOnlyOutputStream.h b/src/Storages/MergeTree/MergedColumnOnlyOutputStream.h index 902138ced9d..507a964ede0 100644 --- a/src/Storages/MergeTree/MergedColumnOnlyOutputStream.h +++ b/src/Storages/MergeTree/MergedColumnOnlyOutputStream.h @@ -27,7 +27,7 @@ public: void write(const Block & block) override; void writeSuffix() override; MergeTreeData::DataPart::Checksums - writeSuffixAndGetChecksums(MergeTreeData::MutableDataPartPtr & new_part, MergeTreeData::DataPart::Checksums & all_checksums); + writeSuffixAndGetChecksums(MergeTreeData::MutableDataPartPtr & new_part, MergeTreeData::DataPart::Checksums & all_checksums, bool sync = false); private: Block header; From b2aa565a37076230af2ceaa32ee21fa351d37931 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Sat, 27 Jun 2020 00:55:48 +0300 Subject: [PATCH 002/121] allow to turn on fsync on inserts, merges and fetches --- src/Common/FileSyncGuard.h | 41 +++++++++++++++++++ src/Disks/DiskLocal.cpp | 35 ++++++++++------ src/Disks/DiskLocal.h | 6 ++- src/Disks/DiskMemory.cpp | 20 ++++++--- src/Disks/DiskMemory.h | 6 ++- src/Disks/IDisk.h | 12 ++++-- src/Disks/S3/DiskS3.cpp | 21 +++++++--- src/Disks/S3/DiskS3.h | 6 ++- src/Storages/MergeTree/DataPartsExchange.cpp | 16 +++++++- src/Storages/MergeTree/DataPartsExchange.h | 1 + src/Storages/MergeTree/IMergeTreeDataPart.cpp | 5 +++ .../MergeTree/MergeTreeDataMergerMutator.cpp | 15 ++++--- .../MergeTree/MergeTreeDataWriter.cpp | 12 +++--- src/Storages/MergeTree/MergeTreeSettings.h | 2 + 14 files changed, 154 insertions(+), 44 deletions(-) create mode 100644 src/Common/FileSyncGuard.h diff --git a/src/Common/FileSyncGuard.h b/src/Common/FileSyncGuard.h new file mode 100644 index 00000000000..5ec9b1d0c98 --- /dev/null +++ b/src/Common/FileSyncGuard.h @@ -0,0 +1,41 @@ +#pragma once + +#include + +namespace DB +{ + +/// Helper class, that recieves file descriptor and does fsync for it in destructor. +/// It's used to keep descriptor open, while doing some operations with it, and do fsync at the end. +/// Guaranties of sequence 'close-reopen-fsync' may depend on kernel version. +/// Source: linux-fsdevel mailing-list https://marc.info/?l=linux-fsdevel&m=152535409207496 +class FileSyncGuard +{ +public: + /// NOTE: If you have already opened descriptor, it's preffered to use + /// this constructor instead of construnctor with path. + FileSyncGuard(const DiskPtr & disk_, int fd_) : disk(disk_), fd(fd_) {} + + FileSyncGuard(const DiskPtr & disk_, const String & path) + : disk(disk_), fd(disk_->open(path, O_RDONLY)) {} + + ~FileSyncGuard() + { + try + { + disk->sync(fd); + disk->close(fd); + } + catch (...) + { + tryLogCurrentException(__PRETTY_FUNCTION__); + } + } + +private: + DiskPtr disk; + int fd = -1; +}; + +} + diff --git a/src/Disks/DiskLocal.cpp b/src/Disks/DiskLocal.cpp index c67bac7ffe2..f85b69baf5e 100644 --- a/src/Disks/DiskLocal.cpp +++ b/src/Disks/DiskLocal.cpp @@ -22,6 +22,7 @@ namespace ErrorCodes extern const int FILE_DOESNT_EXIST; extern const int CANNOT_OPEN_FILE; extern const int CANNOT_FSYNC; + extern const int CANNOT_CLOSE_FILE; } std::mutex DiskLocal::reservation_mutex; @@ -191,18 +192,6 @@ void DiskLocal::moveDirectory(const String & from_path, const String & to_path) Poco::File(disk_path + from_path).renameTo(disk_path + to_path); } -void DiskLocal::sync(const String & path) const -{ - String full_path = disk_path + path; - int fd = ::open(full_path.c_str(), O_RDONLY); - if (-1 == fd) - throwFromErrnoWithPath("Cannot open file " + full_path, full_path, - errno == ENOENT ? ErrorCodes::FILE_DOESNT_EXIST : ErrorCodes::CANNOT_OPEN_FILE); - - if (-1 == fsync(fd)) - throwFromErrnoWithPath("Cannot fsync " + full_path, full_path, ErrorCodes::CANNOT_FSYNC); -} - DiskDirectoryIteratorPtr DiskLocal::iterateDirectory(const String & path) { return std::make_unique(disk_path, path); @@ -299,6 +288,28 @@ void DiskLocal::copy(const String & from_path, const std::shared_ptr & to IDisk::copy(from_path, to_disk, to_path); /// Copy files through buffers. } +int DiskLocal::open(const String & path, mode_t mode) const +{ + String full_path = disk_path + path; + int fd = ::open(full_path.c_str(), mode); + if (-1 == fd) + throwFromErrnoWithPath("Cannot open file " + full_path, full_path, + errno == ENOENT ? ErrorCodes::FILE_DOESNT_EXIST : ErrorCodes::CANNOT_OPEN_FILE); + return fd; +} + +void DiskLocal::close(int fd) const +{ + if (-1 == ::close(fd)) + throw Exception("Cannot close file", ErrorCodes::CANNOT_CLOSE_FILE); +} + +void DiskLocal::sync(int fd) const +{ + if (-1 == ::fsync(fd)) + throw Exception("Cannot fsync", ErrorCodes::CANNOT_FSYNC); +} + DiskPtr DiskLocalReservation::getDisk(size_t i) const { if (i != 0) diff --git a/src/Disks/DiskLocal.h b/src/Disks/DiskLocal.h index 743ba2ceb10..d70ac06c18b 100644 --- a/src/Disks/DiskLocal.h +++ b/src/Disks/DiskLocal.h @@ -59,8 +59,6 @@ public: void moveDirectory(const String & from_path, const String & to_path) override; - void sync(const String & path) const override; - DiskDirectoryIteratorPtr iterateDirectory(const String & path) override; void createFile(const String & path) override; @@ -101,6 +99,10 @@ public: void createHardLink(const String & src_path, const String & dst_path) override; + int open(const String & path, mode_t mode) const override; + void close(int fd) const override; + void sync(int fd) const override; + private: bool tryReserve(UInt64 bytes); diff --git a/src/Disks/DiskMemory.cpp b/src/Disks/DiskMemory.cpp index 5b3350e40f7..a7f1df04e1f 100644 --- a/src/Disks/DiskMemory.cpp +++ b/src/Disks/DiskMemory.cpp @@ -261,11 +261,6 @@ void DiskMemory::moveDirectory(const String & /*from_path*/, const String & /*to throw Exception("Method moveDirectory is not implemented for memory disks", ErrorCodes::NOT_IMPLEMENTED); } -void DiskMemory::sync(const String & /*path*/) const -{ - throw Exception("Method sync is not implemented for memory disks", ErrorCodes::NOT_IMPLEMENTED); -} - DiskDirectoryIteratorPtr DiskMemory::iterateDirectory(const String & path) { std::lock_guard lock(mutex); @@ -413,6 +408,21 @@ void DiskMemory::setReadOnly(const String &) throw Exception("Method setReadOnly is not implemented for memory disks", ErrorCodes::NOT_IMPLEMENTED); } +int DiskMemory::open(const String & /*path*/, mode_t /*mode*/) const +{ + throw Exception("Method open is not implemented for memory disks", ErrorCodes::NOT_IMPLEMENTED); +} + +void DiskMemory::close(int /*fd*/) const +{ + throw Exception("Method close is not implemented for memory disks", ErrorCodes::NOT_IMPLEMENTED); +} + +void DiskMemory::sync(int /*fd*/) const +{ + throw Exception("Method sync is not implemented for memory disks", ErrorCodes::NOT_IMPLEMENTED); +} + using DiskMemoryPtr = std::shared_ptr; diff --git a/src/Disks/DiskMemory.h b/src/Disks/DiskMemory.h index 8a3ddf05aa7..7f111fe5e7d 100644 --- a/src/Disks/DiskMemory.h +++ b/src/Disks/DiskMemory.h @@ -52,8 +52,6 @@ public: void moveDirectory(const String & from_path, const String & to_path) override; - void sync(const String & path) const override; - DiskDirectoryIteratorPtr iterateDirectory(const String & path) override; void createFile(const String & path) override; @@ -92,6 +90,10 @@ public: void createHardLink(const String & src_path, const String & dst_path) override; + int open(const String & path, mode_t mode) const override; + void close(int fd) const override; + void sync(int fd) const override; + private: void createDirectoriesImpl(const String & path); void replaceFileImpl(const String & from_path, const String & to_path); diff --git a/src/Disks/IDisk.h b/src/Disks/IDisk.h index 8de77a560d1..bc5c9381643 100644 --- a/src/Disks/IDisk.h +++ b/src/Disks/IDisk.h @@ -105,9 +105,6 @@ public: /// Move directory from `from_path` to `to_path`. virtual void moveDirectory(const String & from_path, const String & to_path) = 0; - /// Do fsync on directory. - virtual void sync(const String & path) const = 0; - /// Return iterator to the contents of the specified directory. virtual DiskDirectoryIteratorPtr iterateDirectory(const String & path) = 0; @@ -174,6 +171,15 @@ public: /// Create hardlink from `src_path` to `dst_path`. virtual void createHardLink(const String & src_path, const String & dst_path) = 0; + + /// Wrapper for POSIX open + virtual int open(const String & path, mode_t mode) const = 0; + + /// Wrapper for POSIX close + virtual void close(int fd) const = 0; + + /// Wrapper for POSIX fsync + virtual void sync(int fd) const = 0; }; using DiskPtr = std::shared_ptr; diff --git a/src/Disks/S3/DiskS3.cpp b/src/Disks/S3/DiskS3.cpp index 292f6567df4..3e0fb05ed6f 100644 --- a/src/Disks/S3/DiskS3.cpp +++ b/src/Disks/S3/DiskS3.cpp @@ -29,6 +29,7 @@ namespace ErrorCodes extern const int CANNOT_SEEK_THROUGH_FILE; extern const int UNKNOWN_FORMAT; extern const int INCORRECT_DISK_INDEX; + extern const int NOT_IMPLEMENTED; } namespace @@ -466,11 +467,6 @@ void DiskS3::clearDirectory(const String & path) remove(it->path()); } -void DiskS3::sync(const String & /*path*/) const -{ - throw Exception("Method sync is not implemented for S3 disks", ErrorCodes::NOT_IMPLEMENTED); -} - void DiskS3::moveFile(const String & from_path, const String & to_path) { if (exists(to_path)) @@ -669,6 +665,21 @@ void DiskS3::setReadOnly(const String & path) Poco::File(metadata_path + path).setReadOnly(true); } +int DiskS3::open(const String & /*path*/, mode_t /*mode*/) const +{ + throw Exception("Method open is not implemented for S3 disks", ErrorCodes::NOT_IMPLEMENTED); +} + +void DiskS3::close(int /*fd*/) const +{ + throw Exception("Method close is not implemented for S3 disks", ErrorCodes::NOT_IMPLEMENTED); +} + +void DiskS3::sync(int /*fd*/) const +{ + throw Exception("Method sync is not implemented for S3 disks", ErrorCodes::NOT_IMPLEMENTED); +} + DiskS3Reservation::~DiskS3Reservation() { try diff --git a/src/Disks/S3/DiskS3.h b/src/Disks/S3/DiskS3.h index 09132367ae8..cbf161da561 100644 --- a/src/Disks/S3/DiskS3.h +++ b/src/Disks/S3/DiskS3.h @@ -58,8 +58,6 @@ public: void moveDirectory(const String & from_path, const String & to_path) override { moveFile(from_path, to_path); } - void sync(const String & path) const override; - DiskDirectoryIteratorPtr iterateDirectory(const String & path) override; void moveFile(const String & from_path, const String & to_path) override; @@ -98,6 +96,10 @@ public: void setReadOnly(const String & path) override; + int open(const String & path, mode_t mode) const override; + void close(int fd) const override; + void sync(int fd) const override; + private: bool tryReserve(UInt64 bytes); diff --git a/src/Storages/MergeTree/DataPartsExchange.cpp b/src/Storages/MergeTree/DataPartsExchange.cpp index 6796e630ff2..e7bb8206cd9 100644 --- a/src/Storages/MergeTree/DataPartsExchange.cpp +++ b/src/Storages/MergeTree/DataPartsExchange.cpp @@ -2,6 +2,7 @@ #include #include #include +#include #include #include #include @@ -224,9 +225,9 @@ MergeTreeData::MutableDataPartPtr Fetcher::fetchPart( int server_protocol_version = parse(in.getResponseCookie("server_protocol_version", "0")); ReservationPtr reservation; + size_t sum_files_size = 0; if (server_protocol_version >= REPLICATION_PROTOCOL_VERSION_WITH_PARTS_SIZE) { - size_t sum_files_size; readBinary(sum_files_size, in); if (server_protocol_version >= REPLICATION_PROTOCOL_VERSION_WITH_PARTS_SIZE_AND_TTL_INFOS) { @@ -247,7 +248,10 @@ MergeTreeData::MutableDataPartPtr Fetcher::fetchPart( reservation = data.makeEmptyReservationOnLargestDisk(); } - return downloadPart(part_name, replica_path, to_detached, tmp_prefix_, std::move(reservation), in); + bool sync = (data_settings->min_compressed_bytes_to_sync_after_fetch + && sum_files_size >= data_settings->min_compressed_bytes_to_sync_after_fetch); + + return downloadPart(part_name, replica_path, to_detached, tmp_prefix_, sync, std::move(reservation), in); } MergeTreeData::MutableDataPartPtr Fetcher::downloadPart( @@ -255,6 +259,7 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPart( const String & replica_path, bool to_detached, const String & tmp_prefix_, + bool sync, const ReservationPtr reservation, PooledReadWriteBufferFromHTTP & in) { @@ -276,6 +281,10 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPart( disk->createDirectories(part_download_path); + std::optional sync_guard; + if (data.getSettings()->sync_part_directory) + sync_guard.emplace(disk, part_download_path); + MergeTreeData::DataPart::Checksums checksums; for (size_t i = 0; i < files; ++i) { @@ -316,6 +325,9 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPart( if (file_name != "checksums.txt" && file_name != "columns.txt") checksums.addFile(file_name, file_size, expected_hash); + + if (sync) + hashing_out.sync(); } assertEOF(in); diff --git a/src/Storages/MergeTree/DataPartsExchange.h b/src/Storages/MergeTree/DataPartsExchange.h index c1aff6bdba5..e983d6deecf 100644 --- a/src/Storages/MergeTree/DataPartsExchange.h +++ b/src/Storages/MergeTree/DataPartsExchange.h @@ -71,6 +71,7 @@ private: const String & replica_path, bool to_detached, const String & tmp_prefix_, + bool sync, const ReservationPtr reservation, PooledReadWriteBufferFromHTTP & in); diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.cpp b/src/Storages/MergeTree/IMergeTreeDataPart.cpp index 61dfeed6b7c..ab9bb7879aa 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.cpp +++ b/src/Storages/MergeTree/IMergeTreeDataPart.cpp @@ -10,6 +10,7 @@ #include #include #include +#include #include #include @@ -664,6 +665,10 @@ void IMergeTreeDataPart::renameTo(const String & new_relative_path, bool remove_ String from = getFullRelativePath(); String to = storage.relative_data_path + new_relative_path + "/"; + std::optional sync_guard; + if (storage.getSettings()->sync_part_directory) + sync_guard.emplace(volume->getDisk(), to); + if (!volume->getDisk()->exists(from)) throw Exception("Part directory " + fullPath(volume->getDisk(), from) + " doesn't exist. Most likely it is logical error.", ErrorCodes::FILE_DOESNT_EXIST); diff --git a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp index ccd7f234925..9c8c4e3c1d5 100644 --- a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp +++ b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp @@ -30,6 +30,7 @@ #include #include #include +#include #include #include #include @@ -695,6 +696,10 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTempor gathering_column_names.clear(); } + std::optional sync_guard; + if (data.getSettings()->sync_part_directory) + sync_guard.emplace(disk, new_part_tmp_path); + /** Read from all parts, merge and write into a new one. * In passing, we calculate expression for sorting. */ @@ -991,9 +996,6 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTempor else to.writeSuffixAndFinalizePart(new_data_part, need_sync, &storage_columns, &checksums_gathered_columns); - if (need_sync) - new_data_part->volume->getDisk()->sync(new_part_tmp_path); - return new_data_part; } @@ -1089,6 +1091,10 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mutatePartToTempor disk->createDirectories(new_part_tmp_path); + std::optional sync_guard; + if (data.getSettings()->sync_part_directory) + sync_guard.emplace(disk, new_part_tmp_path); + /// Don't change granularity type while mutating subset of columns auto mrk_extension = source_part->index_granularity_info.is_adaptive ? getAdaptiveMrkExtension(new_data_part->getType()) : getNonAdaptiveMrkExtension(); @@ -1187,9 +1193,6 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mutatePartToTempor finalizeMutatedPart(source_part, new_data_part, need_remove_expired_values); } - if (need_sync) - new_data_part->volume->getDisk()->sync(new_part_tmp_path); - return new_data_part; } diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.cpp b/src/Storages/MergeTree/MergeTreeDataWriter.cpp index cf8860b7f04..01f0b086cea 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.cpp +++ b/src/Storages/MergeTree/MergeTreeDataWriter.cpp @@ -12,6 +12,7 @@ #include #include #include +#include namespace ProfileEvents @@ -259,7 +260,12 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataWriter::writeTempPart(BlockWithPa new_data_part->volume->getDisk()->removeRecursive(full_path); } - new_data_part->volume->getDisk()->createDirectories(full_path); + const auto disk = new_data_part->volume->getDisk(); + disk->createDirectories(full_path); + + std::optional sync_guard; + if (data.getSettings()->sync_part_directory) + sync_guard.emplace(disk, full_path); /// If we need to calculate some columns to sort. if (metadata_snapshot->hasSortingKey() || metadata_snapshot->hasSecondaryIndices()) @@ -309,10 +315,6 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataWriter::writeTempPart(BlockWithPa out.writeWithPermutation(block, perm_ptr); out.writeSuffixAndFinalizePart(new_data_part, sync_on_insert); - /// Sync part directory. - if (sync_on_insert) - new_data_part->volume->getDisk()->sync(full_path); - ProfileEvents::increment(ProfileEvents::MergeTreeDataWriterRows, block.rows()); ProfileEvents::increment(ProfileEvents::MergeTreeDataWriterUncompressedBytes, block.bytes()); ProfileEvents::increment(ProfileEvents::MergeTreeDataWriterCompressedBytes, new_data_part->getBytesOnDisk()); diff --git a/src/Storages/MergeTree/MergeTreeSettings.h b/src/Storages/MergeTree/MergeTreeSettings.h index da2c9ee49ee..c559ce2804e 100644 --- a/src/Storages/MergeTree/MergeTreeSettings.h +++ b/src/Storages/MergeTree/MergeTreeSettings.h @@ -45,7 +45,9 @@ struct MergeTreeSettings : public SettingsCollection M(SettingSeconds, lock_acquire_timeout_for_background_operations, DBMS_DEFAULT_LOCK_ACQUIRE_TIMEOUT_SEC, "For background operations like merges, mutations etc. How many seconds before failing to acquire table locks.", 0) \ M(SettingUInt64, min_rows_to_sync_after_merge, 0, "Minimal number of rows to do fsync for part after merge (0 - disabled)", 0) \ M(SettingUInt64, min_compressed_bytes_to_sync_after_merge, 0, "Minimal number of compressed bytes to do fsync for part after merge (0 - disabled)", 0) \ + M(SettingUInt64, min_compressed_bytes_to_sync_after_fetch, 0, "Minimal number of compressed bytes to do fsync for part after fetch (0 - disabled)", 0) \ M(SettingBool, sync_after_insert, false, "Do fsync for every inserted part. Significantly decreases performance of inserts, not recommended to use with wide parts.", 0) \ + M(SettingBool, sync_part_directory, false, "Do fsync for part directory after all part operations (writes, renames, etc.).", 0) \ \ /** Inserts settings. */ \ M(SettingUInt64, parts_to_delay_insert, 150, "If table contains at least that many active parts in single partition, artificially slow down insert into table.", 0) \ From ca346ea13cd0ad0f02a29d59302584c826b52298 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Fri, 3 Jul 2020 02:41:37 +0300 Subject: [PATCH 003/121] rename fsync-related settings --- src/Storages/MergeTree/DataPartsExchange.cpp | 6 +++--- src/Storages/MergeTree/IMergeTreeDataPart.cpp | 2 +- src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp | 8 ++++---- src/Storages/MergeTree/MergeTreeDataWriter.cpp | 4 ++-- src/Storages/MergeTree/MergeTreeSettings.h | 10 +++++----- 5 files changed, 15 insertions(+), 15 deletions(-) diff --git a/src/Storages/MergeTree/DataPartsExchange.cpp b/src/Storages/MergeTree/DataPartsExchange.cpp index e7bb8206cd9..72b478cf587 100644 --- a/src/Storages/MergeTree/DataPartsExchange.cpp +++ b/src/Storages/MergeTree/DataPartsExchange.cpp @@ -248,8 +248,8 @@ MergeTreeData::MutableDataPartPtr Fetcher::fetchPart( reservation = data.makeEmptyReservationOnLargestDisk(); } - bool sync = (data_settings->min_compressed_bytes_to_sync_after_fetch - && sum_files_size >= data_settings->min_compressed_bytes_to_sync_after_fetch); + bool sync = (data_settings->min_compressed_bytes_to_fsync_after_fetch + && sum_files_size >= data_settings->min_compressed_bytes_to_fsync_after_fetch); return downloadPart(part_name, replica_path, to_detached, tmp_prefix_, sync, std::move(reservation), in); } @@ -282,7 +282,7 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPart( disk->createDirectories(part_download_path); std::optional sync_guard; - if (data.getSettings()->sync_part_directory) + if (data.getSettings()->fsync_part_directory) sync_guard.emplace(disk, part_download_path); MergeTreeData::DataPart::Checksums checksums; diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.cpp b/src/Storages/MergeTree/IMergeTreeDataPart.cpp index ab9bb7879aa..3d8cb6b7fc5 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.cpp +++ b/src/Storages/MergeTree/IMergeTreeDataPart.cpp @@ -666,7 +666,7 @@ void IMergeTreeDataPart::renameTo(const String & new_relative_path, bool remove_ String to = storage.relative_data_path + new_relative_path + "/"; std::optional sync_guard; - if (storage.getSettings()->sync_part_directory) + if (storage.getSettings()->fsync_part_directory) sync_guard.emplace(volume->getDisk(), to); if (!volume->getDisk()->exists(from)) diff --git a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp index 9c8c4e3c1d5..c39d1981031 100644 --- a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp +++ b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp @@ -579,8 +579,8 @@ public: static bool needSyncPart(const size_t input_rows, size_t input_bytes, const MergeTreeSettings & settings) { - return ((settings.min_rows_to_sync_after_merge && input_rows >= settings.min_rows_to_sync_after_merge) - || (settings.min_compressed_bytes_to_sync_after_merge && input_bytes >= settings.min_compressed_bytes_to_sync_after_merge)); + return ((settings.min_rows_to_fsync_after_merge && input_rows >= settings.min_rows_to_fsync_after_merge) + || (settings.min_compressed_bytes_to_fsync_after_merge && input_bytes >= settings.min_compressed_bytes_to_fsync_after_merge)); } @@ -697,7 +697,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTempor } std::optional sync_guard; - if (data.getSettings()->sync_part_directory) + if (data.getSettings()->fsync_part_directory) sync_guard.emplace(disk, new_part_tmp_path); /** Read from all parts, merge and write into a new one. @@ -1092,7 +1092,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mutatePartToTempor disk->createDirectories(new_part_tmp_path); std::optional sync_guard; - if (data.getSettings()->sync_part_directory) + if (data.getSettings()->fsync_part_directory) sync_guard.emplace(disk, new_part_tmp_path); /// Don't change granularity type while mutating subset of columns diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.cpp b/src/Storages/MergeTree/MergeTreeDataWriter.cpp index 01f0b086cea..23210fc604e 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.cpp +++ b/src/Storages/MergeTree/MergeTreeDataWriter.cpp @@ -264,7 +264,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataWriter::writeTempPart(BlockWithPa disk->createDirectories(full_path); std::optional sync_guard; - if (data.getSettings()->sync_part_directory) + if (data.getSettings()->fsync_part_directory) sync_guard.emplace(disk, full_path); /// If we need to calculate some columns to sort. @@ -309,7 +309,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataWriter::writeTempPart(BlockWithPa const auto & index_factory = MergeTreeIndexFactory::instance(); MergedBlockOutputStream out(new_data_part, metadata_snapshot, columns, index_factory.getMany(metadata_snapshot->getSecondaryIndices()), compression_codec); - bool sync_on_insert = data.getSettings()->sync_after_insert; + bool sync_on_insert = data.getSettings()->fsync_after_insert; out.writePrefix(); out.writeWithPermutation(block, perm_ptr); diff --git a/src/Storages/MergeTree/MergeTreeSettings.h b/src/Storages/MergeTree/MergeTreeSettings.h index c559ce2804e..eeee0c4b1e1 100644 --- a/src/Storages/MergeTree/MergeTreeSettings.h +++ b/src/Storages/MergeTree/MergeTreeSettings.h @@ -43,11 +43,11 @@ struct MergeTreeSettings : public SettingsCollection M(SettingSeconds, old_parts_lifetime, 8 * 60, "How many seconds to keep obsolete parts.", 0) \ M(SettingSeconds, temporary_directories_lifetime, 86400, "How many seconds to keep tmp_-directories.", 0) \ M(SettingSeconds, lock_acquire_timeout_for_background_operations, DBMS_DEFAULT_LOCK_ACQUIRE_TIMEOUT_SEC, "For background operations like merges, mutations etc. How many seconds before failing to acquire table locks.", 0) \ - M(SettingUInt64, min_rows_to_sync_after_merge, 0, "Minimal number of rows to do fsync for part after merge (0 - disabled)", 0) \ - M(SettingUInt64, min_compressed_bytes_to_sync_after_merge, 0, "Minimal number of compressed bytes to do fsync for part after merge (0 - disabled)", 0) \ - M(SettingUInt64, min_compressed_bytes_to_sync_after_fetch, 0, "Minimal number of compressed bytes to do fsync for part after fetch (0 - disabled)", 0) \ - M(SettingBool, sync_after_insert, false, "Do fsync for every inserted part. Significantly decreases performance of inserts, not recommended to use with wide parts.", 0) \ - M(SettingBool, sync_part_directory, false, "Do fsync for part directory after all part operations (writes, renames, etc.).", 0) \ + M(SettingUInt64, min_rows_to_fsync_after_merge, 0, "Minimal number of rows to do fsync for part after merge (0 - disabled)", 0) \ + M(SettingUInt64, min_compressed_bytes_to_fsync_after_merge, 0, "Minimal number of compressed bytes to do fsync for part after merge (0 - disabled)", 0) \ + M(SettingUInt64, min_compressed_bytes_to_fsync_after_fetch, 0, "Minimal number of compressed bytes to do fsync for part after fetch (0 - disabled)", 0) \ + M(SettingBool, fsync_after_insert, false, "Do fsync for every inserted part. Significantly decreases performance of inserts, not recommended to use with wide parts.", 0) \ + M(SettingBool, fsync_part_directory, false, "Do fsync for part directory after all part operations (writes, renames, etc.).", 0) \ \ /** Inserts settings. */ \ M(SettingUInt64, parts_to_delay_insert, 150, "If table contains at least that many active parts in single partition, artificially slow down insert into table.", 0) \ From 230938d3a3082fbf241c9d873571231a69a5f450 Mon Sep 17 00:00:00 2001 From: Amos Bird Date: Sat, 11 Jul 2020 15:12:42 +0800 Subject: [PATCH 004/121] Refactor joinGet and implement multi-key lookup. --- src/Functions/FunctionJoinGet.cpp | 83 +++++++++---------- src/Functions/FunctionJoinGet.h | 11 +-- src/Interpreters/HashJoin.cpp | 69 ++++++++------- src/Interpreters/HashJoin.h | 10 +-- src/Interpreters/misc.h | 2 +- .../0_stateless/01080_join_get_null.reference | 2 +- .../0_stateless/01080_join_get_null.sql | 12 +-- .../01400_join_get_with_multi_keys.reference | 1 + .../01400_join_get_with_multi_keys.sql | 9 ++ 9 files changed, 104 insertions(+), 95 deletions(-) create mode 100644 tests/queries/0_stateless/01400_join_get_with_multi_keys.reference create mode 100644 tests/queries/0_stateless/01400_join_get_with_multi_keys.sql diff --git a/src/Functions/FunctionJoinGet.cpp b/src/Functions/FunctionJoinGet.cpp index a33b70684a5..1badc689c6a 100644 --- a/src/Functions/FunctionJoinGet.cpp +++ b/src/Functions/FunctionJoinGet.cpp @@ -1,10 +1,10 @@ #include +#include #include #include #include #include -#include #include @@ -16,19 +16,35 @@ namespace ErrorCodes extern const int NUMBER_OF_ARGUMENTS_DOESNT_MATCH; } +template +void ExecutableFunctionJoinGet::execute(Block & block, const ColumnNumbers & arguments, size_t result, size_t) +{ + Block keys; + for (size_t i = 2; i < arguments.size(); ++i) + { + auto key = block.getByPosition(arguments[i]); + keys.insert(std::move(key)); + } + block.getByPosition(result) = join->joinGet(keys, result_block); +} + +template +ExecutableFunctionImplPtr FunctionJoinGet::prepare(const Block &, const ColumnNumbers &, size_t) const +{ + return std::make_unique>(join, Block{{return_type->createColumn(), return_type, attr_name}}); +} + static auto getJoin(const ColumnsWithTypeAndName & arguments, const Context & context) { - if (arguments.size() != 3) - throw Exception{"Function joinGet takes 3 arguments", ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH}; - String join_name; if (const auto * name_col = checkAndGetColumnConst(arguments[0].column.get())) { join_name = name_col->getValue(); } else - throw Exception{"Illegal type " + arguments[0].type->getName() + " of first argument of function joinGet, expected a const string.", - ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT}; + throw Exception( + "Illegal type " + arguments[0].type->getName() + " of first argument of function joinGet, expected a const string.", + ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); size_t dot = join_name.find('.'); String database_name; @@ -43,10 +59,12 @@ static auto getJoin(const ColumnsWithTypeAndName & arguments, const Context & co ++dot; } String table_name = join_name.substr(dot); + if (table_name.empty()) + throw Exception("joinGet does not allow empty table name", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); auto table = DatabaseCatalog::instance().getTable({database_name, table_name}, context); auto storage_join = std::dynamic_pointer_cast(table); if (!storage_join) - throw Exception{"Table " + join_name + " should have engine StorageJoin", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT}; + throw Exception("Table " + join_name + " should have engine StorageJoin", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); String attr_name; if (const auto * name_col = checkAndGetColumnConst(arguments[1].column.get())) @@ -54,57 +72,30 @@ static auto getJoin(const ColumnsWithTypeAndName & arguments, const Context & co attr_name = name_col->getValue(); } else - throw Exception{"Illegal type " + arguments[1].type->getName() - + " of second argument of function joinGet, expected a const string.", - ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT}; + throw Exception( + "Illegal type " + arguments[1].type->getName() + " of second argument of function joinGet, expected a const string.", + ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); return std::make_pair(storage_join, attr_name); } template FunctionBaseImplPtr JoinGetOverloadResolver::build(const ColumnsWithTypeAndName & arguments, const DataTypePtr &) const { + if (arguments.size() < 3) + throw Exception( + "Number of arguments for function " + getName() + " doesn't match: passed " + toString(arguments.size()) + + ", should be greater or equal to 3", + ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH); auto [storage_join, attr_name] = getJoin(arguments, context); auto join = storage_join->getJoin(); - DataTypes data_types(arguments.size()); - + DataTypes data_types(arguments.size() - 2); + for (size_t i = 2; i < arguments.size(); ++i) + data_types[i - 2] = arguments[i].type; + auto return_type = join->joinGetCheckAndGetReturnType(data_types, attr_name, or_null); auto table_lock = storage_join->lockForShare(context.getInitialQueryId(), context.getSettingsRef().lock_acquire_timeout); - for (size_t i = 0; i < arguments.size(); ++i) - data_types[i] = arguments[i].type; - - auto return_type = join->joinGetReturnType(attr_name, or_null); return std::make_unique>(table_lock, storage_join, join, attr_name, data_types, return_type); } -template -DataTypePtr JoinGetOverloadResolver::getReturnType(const ColumnsWithTypeAndName & arguments) const -{ - auto [storage_join, attr_name] = getJoin(arguments, context); - auto join = storage_join->getJoin(); - return join->joinGetReturnType(attr_name, or_null); -} - - -template -void ExecutableFunctionJoinGet::execute(Block & block, const ColumnNumbers & arguments, size_t result, size_t input_rows_count) -{ - auto ctn = block.getByPosition(arguments[2]); - if (isColumnConst(*ctn.column)) - ctn.column = ctn.column->cloneResized(1); - ctn.name = ""; // make sure the key name never collide with the join columns - Block key_block = {ctn}; - join->joinGet(key_block, attr_name, or_null); - auto & result_ctn = key_block.getByPosition(1); - if (isColumnConst(*ctn.column)) - result_ctn.column = ColumnConst::create(result_ctn.column, input_rows_count); - block.getByPosition(result) = result_ctn; -} - -template -ExecutableFunctionImplPtr FunctionJoinGet::prepare(const Block &, const ColumnNumbers &, size_t) const -{ - return std::make_unique>(join, attr_name); -} - void registerFunctionJoinGet(FunctionFactory & factory) { // joinGet diff --git a/src/Functions/FunctionJoinGet.h b/src/Functions/FunctionJoinGet.h index a82da589960..6b3b1202f60 100644 --- a/src/Functions/FunctionJoinGet.h +++ b/src/Functions/FunctionJoinGet.h @@ -13,14 +13,14 @@ template class ExecutableFunctionJoinGet final : public IExecutableFunctionImpl { public: - ExecutableFunctionJoinGet(HashJoinPtr join_, String attr_name_) - : join(std::move(join_)), attr_name(std::move(attr_name_)) {} + ExecutableFunctionJoinGet(HashJoinPtr join_, const Block & result_block_) + : join(std::move(join_)), result_block(result_block_) {} static constexpr auto name = or_null ? "joinGetOrNull" : "joinGet"; bool useDefaultImplementationForNulls() const override { return false; } - bool useDefaultImplementationForConstants() const override { return true; } bool useDefaultImplementationForLowCardinalityColumns() const override { return true; } + bool useDefaultImplementationForConstants() const override { return true; } void execute(Block & block, const ColumnNumbers & arguments, size_t result, size_t input_rows_count) override; @@ -28,7 +28,7 @@ public: private: HashJoinPtr join; - const String attr_name; + Block result_block; }; template @@ -77,13 +77,14 @@ public: String getName() const override { return name; } FunctionBaseImplPtr build(const ColumnsWithTypeAndName & arguments, const DataTypePtr &) const override; - DataTypePtr getReturnType(const ColumnsWithTypeAndName & arguments) const override; + DataTypePtr getReturnType(const ColumnsWithTypeAndName &) const override { return {}; } // Not used bool useDefaultImplementationForNulls() const override { return false; } bool useDefaultImplementationForLowCardinalityColumns() const override { return true; } bool isVariadic() const override { return true; } size_t getNumberOfArguments() const override { return 0; } + ColumnNumbers getArgumentsThatAreAlwaysConstant() const override { return {0, 1}; } private: const Context & context; diff --git a/src/Interpreters/HashJoin.cpp b/src/Interpreters/HashJoin.cpp index 27294a57675..ffc806b9e88 100644 --- a/src/Interpreters/HashJoin.cpp +++ b/src/Interpreters/HashJoin.cpp @@ -42,6 +42,7 @@ namespace ErrorCodes extern const int SYNTAX_ERROR; extern const int SET_SIZE_LIMIT_EXCEEDED; extern const int TYPE_MISMATCH; + extern const int NUMBER_OF_ARGUMENTS_DOESNT_MATCH; } namespace @@ -1109,27 +1110,34 @@ void HashJoin::joinBlockImplCross(Block & block, ExtraBlockPtr & not_processed) block = block.cloneWithColumns(std::move(dst_columns)); } -static void checkTypeOfKey(const Block & block_left, const Block & block_right) -{ - const auto & [c1, left_type_origin, left_name] = block_left.safeGetByPosition(0); - const auto & [c2, right_type_origin, right_name] = block_right.safeGetByPosition(0); - auto left_type = removeNullable(left_type_origin); - auto right_type = removeNullable(right_type_origin); - if (!left_type->equals(*right_type)) - throw Exception("Type mismatch of columns to joinGet by: " - + left_name + " " + left_type->getName() + " at left, " - + right_name + " " + right_type->getName() + " at right", - ErrorCodes::TYPE_MISMATCH); -} - - -DataTypePtr HashJoin::joinGetReturnType(const String & column_name, bool or_null) const +DataTypePtr HashJoin::joinGetCheckAndGetReturnType(const DataTypes & data_types, const String & column_name, bool or_null) const { std::shared_lock lock(data->rwlock); + size_t num_keys = data_types.size(); + if (right_table_keys.columns() != num_keys) + throw Exception( + "Number of arguments for function joinGet" + toString(or_null ? "OrNull" : "") + + " doesn't match: passed, should be equal to " + toString(num_keys), + ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH); + + for (size_t i = 0; i < num_keys; ++i) + { + const auto & left_type_origin = data_types[i]; + const auto & [c2, right_type_origin, right_name] = right_table_keys.safeGetByPosition(i); + auto left_type = removeNullable(left_type_origin); + auto right_type = removeNullable(right_type_origin); + if (!left_type->equals(*right_type)) + throw Exception( + "Type mismatch in joinGet key " + toString(i) + ": found type " + left_type->getName() + ", while the needed type is " + + right_type->getName(), + ErrorCodes::TYPE_MISMATCH); + } + if (!sample_block_with_columns_to_add.has(column_name)) throw Exception("StorageJoin doesn't contain column " + column_name, ErrorCodes::NO_SUCH_COLUMN_IN_TABLE); + auto elem = sample_block_with_columns_to_add.getByName(column_name); if (or_null) elem.type = makeNullable(elem.type); @@ -1138,34 +1146,33 @@ DataTypePtr HashJoin::joinGetReturnType(const String & column_name, bool or_null template -void HashJoin::joinGetImpl(Block & block, const Block & block_with_columns_to_add, const Maps & maps_) const +ColumnWithTypeAndName HashJoin::joinGetImpl(const Block & block, const Block & block_with_columns_to_add, const Maps & maps_) const { - joinBlockImpl( - block, {block.getByPosition(0).name}, block_with_columns_to_add, maps_); + // Assemble the key block with correct names. + Block keys; + for (size_t i = 0; i < block.columns(); ++i) + { + auto key = block.getByPosition(i); + key.name = key_names_right[i]; + keys.insert(std::move(key)); + } + + joinBlockImpl( + keys, key_names_right, block_with_columns_to_add, maps_); + return keys.getByPosition(keys.columns() - 1); } -// TODO: support composite key // TODO: return multiple columns as named tuple // TODO: return array of values when strictness == ASTTableJoin::Strictness::All -void HashJoin::joinGet(Block & block, const String & column_name, bool or_null) const +ColumnWithTypeAndName HashJoin::joinGet(const Block & block, const Block & block_with_columns_to_add) const { std::shared_lock lock(data->rwlock); - if (key_names_right.size() != 1) - throw Exception("joinGet only supports StorageJoin containing exactly one key", ErrorCodes::UNSUPPORTED_JOIN_KEYS); - - checkTypeOfKey(block, right_table_keys); - - auto elem = sample_block_with_columns_to_add.getByName(column_name); - if (or_null) - elem.type = makeNullable(elem.type); - elem.column = elem.type->createColumn(); - if ((strictness == ASTTableJoin::Strictness::Any || strictness == ASTTableJoin::Strictness::RightAny) && kind == ASTTableJoin::Kind::Left) { - joinGetImpl(block, {elem}, std::get(data->maps)); + return joinGetImpl(block, block_with_columns_to_add, std::get(data->maps)); } else throw Exception("joinGet only supports StorageJoin of type Left Any", ErrorCodes::INCOMPATIBLE_TYPE_OF_JOIN); diff --git a/src/Interpreters/HashJoin.h b/src/Interpreters/HashJoin.h index 67d83d27a6d..025f41ac28f 100644 --- a/src/Interpreters/HashJoin.h +++ b/src/Interpreters/HashJoin.h @@ -162,11 +162,11 @@ public: */ void joinBlock(Block & block, ExtraBlockPtr & not_processed) override; - /// Infer the return type for joinGet function - DataTypePtr joinGetReturnType(const String & column_name, bool or_null) const; + /// Check joinGet arguments and infer the return type. + DataTypePtr joinGetCheckAndGetReturnType(const DataTypes & data_types, const String & column_name, bool or_null) const; - /// Used by joinGet function that turns StorageJoin into a dictionary - void joinGet(Block & block, const String & column_name, bool or_null) const; + /// Used by joinGet function that turns StorageJoin into a dictionary. + ColumnWithTypeAndName joinGet(const Block & block, const Block & block_with_columns_to_add) const; /** Keep "totals" (separate part of dataset, see WITH TOTALS) to use later. */ @@ -383,7 +383,7 @@ private: void joinBlockImplCross(Block & block, ExtraBlockPtr & not_processed) const; template - void joinGetImpl(Block & block, const Block & block_with_columns_to_add, const Maps & maps_) const; + ColumnWithTypeAndName joinGetImpl(const Block & block, const Block & block_with_columns_to_add, const Maps & maps_) const; static Type chooseMethod(const ColumnRawPtrs & key_columns, Sizes & key_sizes); }; diff --git a/src/Interpreters/misc.h b/src/Interpreters/misc.h index 094dfbbbb81..cae2691ca1f 100644 --- a/src/Interpreters/misc.h +++ b/src/Interpreters/misc.h @@ -28,7 +28,7 @@ inline bool functionIsLikeOperator(const std::string & name) inline bool functionIsJoinGet(const std::string & name) { - return name == "joinGet" || startsWith(name, "dictGet"); + return startsWith(name, "joinGet"); } inline bool functionIsDictGet(const std::string & name) diff --git a/tests/queries/0_stateless/01080_join_get_null.reference b/tests/queries/0_stateless/01080_join_get_null.reference index bfde072a796..0cfbf08886f 100644 --- a/tests/queries/0_stateless/01080_join_get_null.reference +++ b/tests/queries/0_stateless/01080_join_get_null.reference @@ -1 +1 @@ -2 2 +2 diff --git a/tests/queries/0_stateless/01080_join_get_null.sql b/tests/queries/0_stateless/01080_join_get_null.sql index 71e7ddf8e75..9f782452d34 100644 --- a/tests/queries/0_stateless/01080_join_get_null.sql +++ b/tests/queries/0_stateless/01080_join_get_null.sql @@ -1,12 +1,12 @@ DROP TABLE IF EXISTS test_joinGet; -DROP TABLE IF EXISTS test_join_joinGet; -CREATE TABLE test_joinGet(id Int32, user_id Nullable(Int32)) Engine = Memory(); -CREATE TABLE test_join_joinGet(user_id Int32, name String) Engine = Join(ANY, LEFT, user_id); +CREATE TABLE test_joinGet(user_id Nullable(Int32), name String) Engine = Join(ANY, LEFT, user_id); -INSERT INTO test_join_joinGet VALUES (2, 'a'), (6, 'b'), (10, 'c'); +INSERT INTO test_joinGet VALUES (2, 'a'), (6, 'b'), (10, 'c'), (null, 'd'); -SELECT 2 id, toNullable(toInt32(2)) user_id WHERE joinGet(test_join_joinGet, 'name', user_id) != ''; +SELECT toNullable(toInt32(2)) user_id WHERE joinGet(test_joinGet, 'name', user_id) != ''; + +-- If the JOIN keys are Nullable fields, the rows where at least one of the keys has the value NULL are not joined. +SELECT cast(null AS Nullable(Int32)) user_id WHERE joinGet(test_joinGet, 'name', user_id) != ''; DROP TABLE test_joinGet; -DROP TABLE test_join_joinGet; diff --git a/tests/queries/0_stateless/01400_join_get_with_multi_keys.reference b/tests/queries/0_stateless/01400_join_get_with_multi_keys.reference new file mode 100644 index 00000000000..49d59571fbf --- /dev/null +++ b/tests/queries/0_stateless/01400_join_get_with_multi_keys.reference @@ -0,0 +1 @@ +0.1 diff --git a/tests/queries/0_stateless/01400_join_get_with_multi_keys.sql b/tests/queries/0_stateless/01400_join_get_with_multi_keys.sql new file mode 100644 index 00000000000..73068270762 --- /dev/null +++ b/tests/queries/0_stateless/01400_join_get_with_multi_keys.sql @@ -0,0 +1,9 @@ +DROP TABLE IF EXISTS test_joinGet; + +CREATE TABLE test_joinGet(a String, b String, c Float64) ENGINE = Join(any, left, a, b); + +INSERT INTO test_joinGet VALUES ('ab', '1', 0.1), ('ab', '2', 0.2), ('cd', '3', 0.3); + +SELECT joinGet(test_joinGet, 'c', 'ab', '1'); + +DROP TABLE test_joinGet; From 4331158d3051437f44c7fa1271e4673272cf8cac Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Mon, 24 Aug 2020 16:09:23 +0300 Subject: [PATCH 005/121] merge with master --- src/Disks/DiskDecorator.cpp | 15 +++++++++++++++ src/Disks/DiskDecorator.h | 3 +++ .../MergeTree/MergeTreeDataPartWriterInMemory.cpp | 2 +- .../MergeTree/MergeTreeDataPartWriterInMemory.h | 2 +- src/Storages/MergeTree/MergeTreeDataWriter.cpp | 11 ++++++----- 5 files changed, 26 insertions(+), 7 deletions(-) diff --git a/src/Disks/DiskDecorator.cpp b/src/Disks/DiskDecorator.cpp index e55534e347f..7f2ea58d7cf 100644 --- a/src/Disks/DiskDecorator.cpp +++ b/src/Disks/DiskDecorator.cpp @@ -165,4 +165,19 @@ void DiskDecorator::truncateFile(const String & path, size_t size) delegate->truncateFile(path, size); } +int DiskDecorator::open(const String & path, mode_t mode) const +{ + return delegate->open(path, mode); +} + +void DiskDecorator::close(int fd) const +{ + delegate->close(fd); +} + +void DiskDecorator::sync(int fd) const +{ + delegate->sync(fd); +} + } diff --git a/src/Disks/DiskDecorator.h b/src/Disks/DiskDecorator.h index 71bb100c576..f1ddfff4952 100644 --- a/src/Disks/DiskDecorator.h +++ b/src/Disks/DiskDecorator.h @@ -42,6 +42,9 @@ public: void setReadOnly(const String & path) override; void createHardLink(const String & src_path, const String & dst_path) override; void truncateFile(const String & path, size_t size) override; + int open(const String & path, mode_t mode) const override; + void close(int fd) const override; + void sync(int fd) const override; const String getType() const override { return delegate->getType(); } protected: diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.cpp index a7486158737..f0738a1130a 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.cpp @@ -70,7 +70,7 @@ void MergeTreeDataPartWriterInMemory::calculateAndSerializePrimaryIndex(const Bl } } -void MergeTreeDataPartWriterInMemory::finishDataSerialization(IMergeTreeDataPart::Checksums & checksums) +void MergeTreeDataPartWriterInMemory::finishDataSerialization(IMergeTreeDataPart::Checksums & checksums, bool) { /// If part is empty we still need to initialize block by empty columns. if (!part_in_memory->block) diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.h b/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.h index 92e4228a90d..6e59cdd08a9 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.h +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.h @@ -18,7 +18,7 @@ public: void write(const Block & block, const IColumn::Permutation * permutation, const Block & primary_key_block, const Block & skip_indexes_block) override; - void finishDataSerialization(IMergeTreeDataPart::Checksums & checksums) override; + void finishDataSerialization(IMergeTreeDataPart::Checksums & checksums, bool sync) override; void calculateAndSerializePrimaryIndex(const Block & primary_index_block) override; diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.cpp b/src/Storages/MergeTree/MergeTreeDataWriter.cpp index f3a72657be5..b05b970da3b 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.cpp +++ b/src/Storages/MergeTree/MergeTreeDataWriter.cpp @@ -251,6 +251,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataWriter::writeTempPart(BlockWithPa new_data_part->minmax_idx = std::move(minmax_idx); new_data_part->is_temp = true; + std::optional sync_guard; if (new_data_part->isStoredOnDisk()) { /// The name could be non-unique in case of stale files from previous runs. @@ -262,12 +263,12 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataWriter::writeTempPart(BlockWithPa new_data_part->volume->getDisk()->removeRecursive(full_path); } - const auto disk = new_data_part->volume->getDisk(); - disk->createDirectories(full_path); + const auto disk = new_data_part->volume->getDisk(); + disk->createDirectories(full_path); - std::optional sync_guard; - if (data.getSettings()->fsync_part_directory) - sync_guard.emplace(disk, full_path); + if (data.getSettings()->fsync_part_directory) + sync_guard.emplace(disk, full_path); + } /// If we need to calculate some columns to sort. if (metadata_snapshot->hasSortingKey() || metadata_snapshot->hasSecondaryIndices()) From 25140b9bd5b6421b84ef8586827cc49b9d015e7b Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Tue, 1 Sep 2020 04:39:36 +0300 Subject: [PATCH 006/121] fsync MergeTree format file --- src/Storages/MergeTree/MergeTreeData.cpp | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp index bbefba70c58..bc668659b6a 100644 --- a/src/Storages/MergeTree/MergeTreeData.cpp +++ b/src/Storages/MergeTree/MergeTreeData.cpp @@ -220,6 +220,8 @@ MergeTreeData::MergeTreeData( format_version = min_format_version; auto buf = version_file.second->writeFile(version_file.first); writeIntText(format_version.toUnderType(), *buf); + if (global_context.getSettingsRef().fsync_metadata) + buf->sync(); } else { From 927eb32e882d070ff5ff5446d5b9e0071e2c6f9d Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Tue, 1 Sep 2020 04:46:40 +0300 Subject: [PATCH 007/121] add test for durability (draft) --- utils/durability-test/create.sql | 1 + utils/durability-test/durability-test.sh | 154 +++++++++++++++++++++++ utils/durability-test/insert.sql | 1 + utils/durability-test/install.sh | 3 + utils/durability-test/sshd_config | 8 ++ utils/durability-test/startup.exp | 23 ++++ 6 files changed, 190 insertions(+) create mode 100644 utils/durability-test/create.sql create mode 100644 utils/durability-test/durability-test.sh create mode 100644 utils/durability-test/insert.sql create mode 100644 utils/durability-test/install.sh create mode 100644 utils/durability-test/sshd_config create mode 100755 utils/durability-test/startup.exp diff --git a/utils/durability-test/create.sql b/utils/durability-test/create.sql new file mode 100644 index 00000000000..1ec394100e2 --- /dev/null +++ b/utils/durability-test/create.sql @@ -0,0 +1 @@ +CREATE TABLE test (a Int, s String) ENGINE = MergeTree ORDER BY a; diff --git a/utils/durability-test/durability-test.sh b/utils/durability-test/durability-test.sh new file mode 100644 index 00000000000..1f47c900f49 --- /dev/null +++ b/utils/durability-test/durability-test.sh @@ -0,0 +1,154 @@ +#!/bin/bash + +URL=http://cloud-images.ubuntu.com/bionic/current +IMAGE=bionic-server-cloudimg-amd64.img +SSH_PORT=11022 +CLICKHOUSE_PORT=9090 +PASSWORD=root + +TABLE_NAME=$1 +CREATE_QUERY=$2 +INSERT_QUERY=$3 + +if [[ -z $TABLE_NAME || -z $CREATE_QUERY || -z $INSERT_QUERY ]]; then + echo "Required 3 arguments: table name, file with create query, file with insert query" + exit 1 +fi + +function run() +{ + sshpass -p $PASSWORD ssh -p $SSH_PORT root@localhost "$1" +} + +function copy() +{ + sshpass -p $PASSWORD scp -r -P $SSH_PORT $1 root@localhost:$2 +} + +function wait_vm_for_start() +{ + echo "Waiting until VM started..." + started=0 + for i in {0..100}; do + run "exit" + if [ $? -eq 0 ]; then + started=1 + break + fi + sleep 1s + done + + if ((started == 0)); then + echo "Can't start or connect to VM." + exit 1 + fi + + echo "Started VM" +} + +function wait_clickhouse_for_start() +{ + echo "Waiting until ClickHouse started..." + started=0 + for i in {0..15}; do + run "clickhouse client --query 'select 1'" + if [ $? -eq 0 ]; then + started=1 + break + fi + sleep 1s + done + + if ((started == 0)); then + echo "Can't start ClickHouse." + fi + + echo "Started ClickHouse" +} + +echo "Downloading image" +curl -O $URL/$IMAGE + +qemu-img resize $IMAGE +10G +virt-customize -a $IMAGE --root-password password:$PASSWORD +virt-copy-in -a $IMAGE sshd_config /etc/ssh + +echo "Starting VM" + +chmod +x ./startup.exp +./startup.exp > qemu.log 2>&1 & + +wait_vm_for_start + +echo "Preparing VM" + +# Resize partition +run "growpart /dev/sda 1 && resize2fs /dev/sda1" + +if [[ -z $CLICKHOUSE_BINARY ]]; then + CLICKHOUSE_BINARY=/usr/bin/clickhouse +fi + +if [[ -z $CLICKHOUSE_CONFIG_DIR ]]; then + CLICKHOUSE_CONFIG_DIR=/etc/clickhouse-server +fi + +echo "Using ClickHouse binary: " $CLICKHOUSE_BINARY +echo "Using ClickHouse config from: " $CLICKHOUSE_CONFIG_DIR + +copy $CLICKHOUSE_BINARY /usr/bin +copy $CLICKHOUSE_CONFIG_DIR /etc +run "mv /etc/$CLICKHOUSE_CONFIG_DIR /etc/clickhouse-server" + +echo "Prepared VM" +echo "Starting ClickHouse" + +run "clickhouse server --config-file=/etc/clickhouse-server/config.xml > clickhouse-server.log 2>&1" & + +wait_clickhouse_for_start + +echo "Started ClickHouse" + +query=`cat $CREATE_QUERY` +echo "Executing query:" $query +run "clickhouse client --query '$query'" + +query=`cat $INSERT_QUERY` +echo "Will run in a loop query: " $query +run "clickhouse benchmark <<< '$query'" & +echo "Running queries" + +pid=`pidof qemu-system-x86_64` +sec=$(( (RANDOM % 3) + 25 )) + +ms=$(( RANDOM % 1000 )) + +echo "Will kill VM in $sec.$ms sec" + +sleep $sec.$ms +kill -9 $pid + +echo "Restarting" + +./startup.exp > qemu.log 2>&1 & +wait_vm_for_start + +run "rm -r *data/system" +run "clickhouse server --config-file=/etc/clickhouse-server/config.xml > clickhouse-server.log 2>&1" & +wait_clickhouse_for_start + +result=`run "grep $TABLE_NAME clickhouse-server.log | grep 'Caught exception while loading metadata'"` +if [[ -n $result ]]; then + echo "FAIL. Can't attach table:" + echo $result + exit 1 +fi + +result=`run "grep $TABLE_NAME clickhouse-server.log | grep 'Considering to remove broken part'"` +if [[ -n $result ]]; then + echo "FAIL. Have broken parts:" + echo $result + exit 1 +fi + +echo OK diff --git a/utils/durability-test/insert.sql b/utils/durability-test/insert.sql new file mode 100644 index 00000000000..8982ad47228 --- /dev/null +++ b/utils/durability-test/insert.sql @@ -0,0 +1 @@ +INSERT INTO test SELECT number, toString(number) FROM numbers(10) diff --git a/utils/durability-test/install.sh b/utils/durability-test/install.sh new file mode 100644 index 00000000000..526cde6743f --- /dev/null +++ b/utils/durability-test/install.sh @@ -0,0 +1,3 @@ +#!/bin/bash + +apt update && apt install qemu-kvm qemu virt-manager virt-viewer libguestfs-tools sshpass expect diff --git a/utils/durability-test/sshd_config b/utils/durability-test/sshd_config new file mode 100644 index 00000000000..6ed06d3d8ad --- /dev/null +++ b/utils/durability-test/sshd_config @@ -0,0 +1,8 @@ +PermitRootLogin yes +PasswordAuthentication yes +ChallengeResponseAuthentication no +UsePAM yes +X11Forwarding yes +PrintMotd no +AcceptEnv LANG LC_* +Subsystem sftp /usr/lib/openssh/sftp-server diff --git a/utils/durability-test/startup.exp b/utils/durability-test/startup.exp new file mode 100755 index 00000000000..540cfc0e4b8 --- /dev/null +++ b/utils/durability-test/startup.exp @@ -0,0 +1,23 @@ +#!/usr/bin/expect -f + +# Wait enough (forever) until a long-time boot +set timeout -1 + +spawn qemu-system-x86_64 \ + -hda bionic-server-cloudimg-amd64.img \ + -cpu qemu64,+ssse3,+sse4.1,+sse4.2,+popcnt -smp 8 \ + -net nic -net user,hostfwd=tcp::11022-:22 \ + -m 4096 -nographic + +expect "login: " +send "root\n" + +expect "Password: " +send "root\n" + +# Without it ssh is not working on guest machine for some reason +expect "# " +send "dhclient && ssh-keygen -A && systemctl restart sshd.service\n" + +# Wait forever +expect "########" From 3cadc9033ae63d7faa851b1707b3c6f9ce1a36aa Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Tue, 1 Sep 2020 18:26:49 +0300 Subject: [PATCH 008/121] fsyncs for metadata files of part --- .../MergeTree/IMergeTreeDataPartWriter.h | 2 +- .../MergeTreeDataPartWriterOnDisk.cpp | 2 +- .../MergeTree/MergedBlockOutputStream.cpp | 13 +++++++-- .../MergeTree/MergedBlockOutputStream.h | 3 +- utils/durability-test/create_sync.sql | 1 + utils/durability-test/durability-test.sh | 28 ++++++++++--------- utils/durability-test/insert_sync.sql | 1 + 7 files changed, 32 insertions(+), 18 deletions(-) create mode 100644 utils/durability-test/create_sync.sql mode change 100644 => 100755 utils/durability-test/durability-test.sh create mode 100644 utils/durability-test/insert_sync.sql diff --git a/src/Storages/MergeTree/IMergeTreeDataPartWriter.h b/src/Storages/MergeTree/IMergeTreeDataPartWriter.h index 4d3602e732e..4a42a58a65b 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPartWriter.h +++ b/src/Storages/MergeTree/IMergeTreeDataPartWriter.h @@ -52,7 +52,7 @@ public: virtual void initPrimaryIndex() {} virtual void finishDataSerialization(IMergeTreeDataPart::Checksums & checksums, bool sync) = 0; - virtual void finishPrimaryIndexSerialization(MergeTreeData::DataPart::Checksums & /* checksums */, bool /* sync */) {} + virtual void finishPrimaryIndexSerialization(MergeTreeData::DataPart::Checksums & /* checksums */, bool /* sync */) {} virtual void finishSkipIndicesSerialization(MergeTreeData::DataPart::Checksums & /* checksums */, bool /* sync */) {} Columns releaseIndexColumns(); diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp index dbe41144573..8295b881d87 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp @@ -332,7 +332,7 @@ void MergeTreeDataPartWriterOnDisk::finishPrimaryIndexSerialization( checksums.files["primary.idx"].file_size = index_stream->count(); checksums.files["primary.idx"].file_hash = index_stream->getHash(); if (sync) - index_stream->sync(); + index_file_stream->sync(); index_stream = nullptr; } } diff --git a/src/Storages/MergeTree/MergedBlockOutputStream.cpp b/src/Storages/MergeTree/MergedBlockOutputStream.cpp index fdef5d69688..bdc6bade259 100644 --- a/src/Storages/MergeTree/MergedBlockOutputStream.cpp +++ b/src/Storages/MergeTree/MergedBlockOutputStream.cpp @@ -111,7 +111,7 @@ void MergedBlockOutputStream::writeSuffixAndFinalizePart( part_columns = *total_columns_list; if (new_part->isStoredOnDisk()) - finalizePartOnDisk(new_part, part_columns, checksums); + finalizePartOnDisk(new_part, part_columns, checksums, sync); new_part->setColumns(part_columns); new_part->rows_count = rows_count; @@ -126,7 +126,8 @@ void MergedBlockOutputStream::writeSuffixAndFinalizePart( void MergedBlockOutputStream::finalizePartOnDisk( const MergeTreeData::MutableDataPartPtr & new_part, NamesAndTypesList & part_columns, - MergeTreeData::DataPart::Checksums & checksums) + MergeTreeData::DataPart::Checksums & checksums, + bool sync) { if (storage.format_version >= MERGE_TREE_DATA_MIN_FORMAT_VERSION_WITH_CUSTOM_PARTITIONING || isCompactPart(new_part)) { @@ -143,6 +144,8 @@ void MergedBlockOutputStream::finalizePartOnDisk( count_out_hashing.next(); checksums.files["count.txt"].file_size = count_out_hashing.count(); checksums.files["count.txt"].file_hash = count_out_hashing.getHash(); + if (sync) + count_out->sync(); } if (!new_part->ttl_infos.empty()) @@ -153,6 +156,8 @@ void MergedBlockOutputStream::finalizePartOnDisk( new_part->ttl_infos.write(out_hashing); checksums.files["ttl.txt"].file_size = out_hashing.count(); checksums.files["ttl.txt"].file_hash = out_hashing.getHash(); + if (sync) + out->sync(); } removeEmptyColumnsFromPart(new_part, part_columns, checksums); @@ -161,12 +166,16 @@ void MergedBlockOutputStream::finalizePartOnDisk( /// Write a file with a description of columns. auto out = volume->getDisk()->writeFile(part_path + "columns.txt", 4096); part_columns.writeText(*out); + if (sync) + out->sync(); } { /// Write file with checksums. auto out = volume->getDisk()->writeFile(part_path + "checksums.txt", 4096); checksums.write(*out); + if (sync) + out->sync(); } } diff --git a/src/Storages/MergeTree/MergedBlockOutputStream.h b/src/Storages/MergeTree/MergedBlockOutputStream.h index 0b500b93f01..87ff9dd1ded 100644 --- a/src/Storages/MergeTree/MergedBlockOutputStream.h +++ b/src/Storages/MergeTree/MergedBlockOutputStream.h @@ -59,7 +59,8 @@ private: void finalizePartOnDisk( const MergeTreeData::MutableDataPartPtr & new_part, NamesAndTypesList & part_columns, - MergeTreeData::DataPart::Checksums & checksums); + MergeTreeData::DataPart::Checksums & checksums, + bool sync); private: NamesAndTypesList columns_list; diff --git a/utils/durability-test/create_sync.sql b/utils/durability-test/create_sync.sql new file mode 100644 index 00000000000..2cc88d2c943 --- /dev/null +++ b/utils/durability-test/create_sync.sql @@ -0,0 +1 @@ +CREATE TABLE test_sync (a Int, s String) ENGINE = MergeTree ORDER BY a SETTINGS fsync_after_insert = 1, min_compressed_bytes_to_fsync_after_merge = 1; diff --git a/utils/durability-test/durability-test.sh b/utils/durability-test/durability-test.sh old mode 100644 new mode 100755 index 1f47c900f49..c7f8936ec95 --- a/utils/durability-test/durability-test.sh +++ b/utils/durability-test/durability-test.sh @@ -17,12 +17,12 @@ fi function run() { - sshpass -p $PASSWORD ssh -p $SSH_PORT root@localhost "$1" + sshpass -p $PASSWORD ssh -p $SSH_PORT root@localhost "$1" 2>/dev/null } function copy() { - sshpass -p $PASSWORD scp -r -P $SSH_PORT $1 root@localhost:$2 + sshpass -p $PASSWORD scp -r -P $SSH_PORT $1 root@localhost:$2 2>/dev/null } function wait_vm_for_start() @@ -50,8 +50,8 @@ function wait_clickhouse_for_start() { echo "Waiting until ClickHouse started..." started=0 - for i in {0..15}; do - run "clickhouse client --query 'select 1'" + for i in {0..30}; do + run "clickhouse client --query 'select 1'" > /dev/null if [ $? -eq 0 ]; then started=1 break @@ -70,7 +70,7 @@ echo "Downloading image" curl -O $URL/$IMAGE qemu-img resize $IMAGE +10G -virt-customize -a $IMAGE --root-password password:$PASSWORD +virt-customize -a $IMAGE --root-password password:$PASSWORD > /dev/null 2>&1 virt-copy-in -a $IMAGE sshd_config /etc/ssh echo "Starting VM" @@ -93,8 +93,8 @@ if [[ -z $CLICKHOUSE_CONFIG_DIR ]]; then CLICKHOUSE_CONFIG_DIR=/etc/clickhouse-server fi -echo "Using ClickHouse binary: " $CLICKHOUSE_BINARY -echo "Using ClickHouse config from: " $CLICKHOUSE_CONFIG_DIR +echo "Using ClickHouse binary:" $CLICKHOUSE_BINARY +echo "Using ClickHouse config from:" $CLICKHOUSE_CONFIG_DIR copy $CLICKHOUSE_BINARY /usr/bin copy $CLICKHOUSE_CONFIG_DIR /etc @@ -104,23 +104,19 @@ echo "Prepared VM" echo "Starting ClickHouse" run "clickhouse server --config-file=/etc/clickhouse-server/config.xml > clickhouse-server.log 2>&1" & - wait_clickhouse_for_start -echo "Started ClickHouse" - query=`cat $CREATE_QUERY` echo "Executing query:" $query run "clickhouse client --query '$query'" query=`cat $INSERT_QUERY` echo "Will run in a loop query: " $query -run "clickhouse benchmark <<< '$query'" & +run "clickhouse benchmark <<< '$query' -c 8" & echo "Running queries" pid=`pidof qemu-system-x86_64` -sec=$(( (RANDOM % 3) + 25 )) - +sec=$(( (RANDOM % 5) + 25 )) ms=$(( RANDOM % 1000 )) echo "Will kill VM in $sec.$ms sec" @@ -130,6 +126,8 @@ kill -9 $pid echo "Restarting" +sleep 5s + ./startup.exp > qemu.log 2>&1 & wait_vm_for_start @@ -137,10 +135,12 @@ run "rm -r *data/system" run "clickhouse server --config-file=/etc/clickhouse-server/config.xml > clickhouse-server.log 2>&1" & wait_clickhouse_for_start +pid=`pidof qemu-system-x86_64` result=`run "grep $TABLE_NAME clickhouse-server.log | grep 'Caught exception while loading metadata'"` if [[ -n $result ]]; then echo "FAIL. Can't attach table:" echo $result + kill -9 $pid exit 1 fi @@ -148,7 +148,9 @@ result=`run "grep $TABLE_NAME clickhouse-server.log | grep 'Considering to remov if [[ -n $result ]]; then echo "FAIL. Have broken parts:" echo $result + kill -9 $pid exit 1 fi +kill -9 $pid echo OK diff --git a/utils/durability-test/insert_sync.sql b/utils/durability-test/insert_sync.sql new file mode 100644 index 00000000000..a1ad2ff4ea5 --- /dev/null +++ b/utils/durability-test/insert_sync.sql @@ -0,0 +1 @@ +INSERT INTO test_sync SELECT number, toString(number) FROM numbers(10) From 26d75f76026303b6f3769ab4ea39ff639ebe836a Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Wed, 2 Sep 2020 01:25:10 +0300 Subject: [PATCH 009/121] do fsync for WAL --- src/Storages/MergeTree/MergeTreeSettings.h | 2 ++ .../MergeTree/MergeTreeWriteAheadLog.cpp | 32 +++++++++++++++++-- .../MergeTree/MergeTreeWriteAheadLog.h | 10 +++++- 3 files changed, 41 insertions(+), 3 deletions(-) diff --git a/src/Storages/MergeTree/MergeTreeSettings.h b/src/Storages/MergeTree/MergeTreeSettings.h index 1341526c38b..edf03710974 100644 --- a/src/Storages/MergeTree/MergeTreeSettings.h +++ b/src/Storages/MergeTree/MergeTreeSettings.h @@ -43,6 +43,8 @@ struct Settings; M(UInt64, min_compressed_bytes_to_fsync_after_fetch, 0, "Minimal number of compressed bytes to do fsync for part after fetch (0 - disabled)", 0) \ M(Bool, fsync_after_insert, false, "Do fsync for every inserted part. Significantly decreases performance of inserts, not recommended to use with wide parts.", 0) \ M(Bool, fsync_part_directory, false, "Do fsync for part directory after all part operations (writes, renames, etc.).", 0) \ + M(UInt64, write_ahead_log_bytes_to_fsync, 100ULL * 1024 * 1024, "Amount of bytes, accumulated in WAL to do fsync.", 0) \ + M(UInt64, write_ahead_log_interval_ms_to_fsync, 100, "Interval in milliseconds after which fsync for WAL is being done.", 0) \ \ /** Inserts settings. */ \ M(UInt64, parts_to_delay_insert, 150, "If table contains at least that many active parts in single partition, artificially slow down insert into table.", 0) \ diff --git a/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp b/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp index eda8579c76a..6f220fc7d5d 100644 --- a/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp +++ b/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp @@ -4,6 +4,7 @@ #include #include #include +#include namespace DB { @@ -16,17 +17,23 @@ namespace ErrorCodes extern const int CORRUPTED_DATA; } - MergeTreeWriteAheadLog::MergeTreeWriteAheadLog( - const MergeTreeData & storage_, + MergeTreeData & storage_, const DiskPtr & disk_, const String & name_) : storage(storage_) , disk(disk_) , name(name_) , path(storage.getRelativeDataPath() + name_) + , pool(storage.global_context.getSchedulePool()) { init(); + sync_task = pool.createTask("MergeTreeWriteAheadLog::sync", [this] + { + std::lock_guard lock(write_mutex); + out->sync(); + sync_scheduled = false; + }); } void MergeTreeWriteAheadLog::init() @@ -38,6 +45,7 @@ void MergeTreeWriteAheadLog::init() block_out = std::make_unique(*out, 0, Block{}); min_block_number = std::numeric_limits::max(); max_block_number = -1; + bytes_at_last_sync = 0; } void MergeTreeWriteAheadLog::addPart(const Block & block, const String & part_name) @@ -53,6 +61,7 @@ void MergeTreeWriteAheadLog::addPart(const Block & block, const String & part_na writeStringBinary(part_name, *out); block_out->write(block); block_out->flush(); + sync(lock); auto max_wal_bytes = storage.getSettings()->write_ahead_log_max_bytes; if (out->count() > max_wal_bytes) @@ -66,6 +75,7 @@ void MergeTreeWriteAheadLog::dropPart(const String & part_name) writeIntBinary(static_cast(0), *out); writeIntBinary(static_cast(ActionType::DROP_PART), *out); writeStringBinary(part_name, *out); + sync(lock); } void MergeTreeWriteAheadLog::rotate(const std::lock_guard &) @@ -175,6 +185,24 @@ MergeTreeData::MutableDataPartsVector MergeTreeWriteAheadLog::restore(const Stor return result; } +void MergeTreeWriteAheadLog::sync(const std::lock_guard &) +{ + size_t bytes_to_sync = storage.getSettings()->write_ahead_log_bytes_to_fsync; + time_t time_to_sync = storage.getSettings()->write_ahead_log_interval_ms_to_fsync; + size_t current_bytes = out->count(); + + if (bytes_to_sync && current_bytes - bytes_at_last_sync > bytes_to_sync) + { + sync_task->schedule(); + bytes_at_last_sync = current_bytes; + } + else if (time_to_sync && !sync_scheduled) + { + sync_task->scheduleAfter(time_to_sync); + sync_scheduled = true; + } +} + std::optional MergeTreeWriteAheadLog::tryParseMinMaxBlockNumber(const String & filename) { diff --git a/src/Storages/MergeTree/MergeTreeWriteAheadLog.h b/src/Storages/MergeTree/MergeTreeWriteAheadLog.h index 2cc3c2b4181..43abf3c04be 100644 --- a/src/Storages/MergeTree/MergeTreeWriteAheadLog.h +++ b/src/Storages/MergeTree/MergeTreeWriteAheadLog.h @@ -3,6 +3,7 @@ #include #include #include +#include #include namespace DB @@ -31,7 +32,7 @@ public: constexpr static auto WAL_FILE_EXTENSION = ".bin"; constexpr static auto DEFAULT_WAL_FILE_NAME = "wal.bin"; - MergeTreeWriteAheadLog(const MergeTreeData & storage_, const DiskPtr & disk_, + MergeTreeWriteAheadLog(MergeTreeData & storage_, const DiskPtr & disk_, const String & name = DEFAULT_WAL_FILE_NAME); void addPart(const Block & block, const String & part_name); @@ -44,6 +45,7 @@ public: private: void init(); void rotate(const std::lock_guard & lock); + void sync(const std::lock_guard & lock); const MergeTreeData & storage; DiskPtr disk; @@ -56,6 +58,12 @@ private: Int64 min_block_number = std::numeric_limits::max(); Int64 max_block_number = -1; + BackgroundSchedulePool & pool; + BackgroundSchedulePoolTaskHolder sync_task; + + size_t bytes_at_last_sync = 0; + bool sync_scheduled = false; + mutable std::mutex write_mutex; }; From 23b9677879a2a0618b35032439650ec08e760c57 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Fri, 4 Sep 2020 08:46:58 +0300 Subject: [PATCH 010/121] Added a script to import git repository to ClickHouse --- src/Common/ShellCommand.cpp | 4 + src/IO/ReadBufferFromFile.cpp | 3 + src/IO/WriteBufferFromFile.cpp | 3 + utils/CMakeLists.txt | 1 + utils/git-to-clickhouse/CMakeLists.txt | 2 + utils/git-to-clickhouse/git-to-clickhouse.cpp | 638 ++++++++++++++++++ 6 files changed, 651 insertions(+) create mode 100644 utils/git-to-clickhouse/CMakeLists.txt create mode 100644 utils/git-to-clickhouse/git-to-clickhouse.cpp diff --git a/src/Common/ShellCommand.cpp b/src/Common/ShellCommand.cpp index 53ab2301a0a..127f95fef06 100644 --- a/src/Common/ShellCommand.cpp +++ b/src/Common/ShellCommand.cpp @@ -186,6 +186,10 @@ int ShellCommand::tryWait() { wait_called = true; + in.close(); + out.close(); + err.close(); + LOG_TRACE(getLogger(), "Will wait for shell command pid {}", pid); int status = 0; diff --git a/src/IO/ReadBufferFromFile.cpp b/src/IO/ReadBufferFromFile.cpp index 40f69625e68..226615c757e 100644 --- a/src/IO/ReadBufferFromFile.cpp +++ b/src/IO/ReadBufferFromFile.cpp @@ -77,6 +77,9 @@ ReadBufferFromFile::~ReadBufferFromFile() void ReadBufferFromFile::close() { + if (fd < 0) + return; + if (0 != ::close(fd)) throw Exception("Cannot close file", ErrorCodes::CANNOT_CLOSE_FILE); diff --git a/src/IO/WriteBufferFromFile.cpp b/src/IO/WriteBufferFromFile.cpp index b59a110edb4..4ade2e2c971 100644 --- a/src/IO/WriteBufferFromFile.cpp +++ b/src/IO/WriteBufferFromFile.cpp @@ -92,6 +92,9 @@ WriteBufferFromFile::~WriteBufferFromFile() /// Close file before destruction of object. void WriteBufferFromFile::close() { + if (fd < 0) + return; + next(); if (0 != ::close(fd)) diff --git a/utils/CMakeLists.txt b/utils/CMakeLists.txt index 0dd95388e7d..dd03afe9fb8 100644 --- a/utils/CMakeLists.txt +++ b/utils/CMakeLists.txt @@ -29,6 +29,7 @@ if (NOT DEFINED ENABLE_UTILS OR ENABLE_UTILS) add_subdirectory (convert-month-partitioned-parts) add_subdirectory (checksum-for-compressed-block) add_subdirectory (wal-dump) + add_subdirectory (git-to-clickhouse) endif () if (ENABLE_CODE_QUALITY) diff --git a/utils/git-to-clickhouse/CMakeLists.txt b/utils/git-to-clickhouse/CMakeLists.txt new file mode 100644 index 00000000000..0e46b68d471 --- /dev/null +++ b/utils/git-to-clickhouse/CMakeLists.txt @@ -0,0 +1,2 @@ +add_executable (git-to-clickhouse git-to-clickhouse.cpp) +target_link_libraries(git-to-clickhouse PRIVATE dbms boost::program_options) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp new file mode 100644 index 00000000000..42920328ad7 --- /dev/null +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -0,0 +1,638 @@ +#include +#include +#include + +#include + +#include +#include +#include +#include + +#include +#include +#include +#include +#include + + +namespace DB +{ + +namespace ErrorCodes +{ + extern const int INCORRECT_DATA; +} + +enum class LineType +{ + Empty, + Comment, + Punct, + Code, +}; + +void writeText(LineType type, WriteBuffer & out) +{ + switch (type) + { + case LineType::Empty: writeString("Empty", out); break; + case LineType::Comment: writeString("Comment", out); break; + case LineType::Punct: writeString("Punct", out); break; + case LineType::Code: writeString("Code", out); break; + } +} + +struct LineChange +{ + int8_t sign{}; /// 1 if added, -1 if deleted + uint16_t line_number_old{}; + uint16_t line_number_new{}; + uint16_t hunk_num{}; /// ordinal number of hunk in diff, starting with 0 + uint16_t hunk_start_line_number_old{}; + uint16_t hunk_start_line_number_new{}; + std::string hunk_context; /// The context (like a line with function name) as it is calculated by git + std::string line; /// Line content without leading whitespaces + uint8_t indent{}; /// The number of leading whitespaces or tabs * 4 + LineType line_type{}; + + void setLineInfo(std::string full_line) + { + indent = 0; + + const char * pos = full_line.data(); + const char * end = pos + full_line.size(); + + while (pos < end) + { + if (*pos == ' ') + ++indent; + else if (*pos == '\t') + indent += 4; + else + break; + ++pos; + } + + line.assign(pos, end); + + if (pos == end) + { + line_type = LineType::Empty; + } + else if (pos + 1 < end + && ((pos[0] == '/' && pos[1] == '/') + || (pos[0] == '*' && pos[1] == ' '))) /// This is not precise. + { + line_type = LineType::Comment; + } + else + { + while (pos < end) + { + if (isAlphaNumericASCII(*pos)) + { + line_type = LineType::Code; + break; + } + ++pos; + } + if (pos == end) + line_type = LineType::Punct; + } + } + + void writeTextWithoutNewline(WriteBuffer & out) const + { + writeText(sign, out); + writeChar('\t', out); + writeText(line_number_old, out); + writeChar('\t', out); + writeText(line_number_new, out); + writeChar('\t', out); + writeText(hunk_num, out); + writeChar('\t', out); + writeText(hunk_start_line_number_old, out); + writeChar('\t', out); + writeText(hunk_start_line_number_new, out); + writeChar('\t', out); + writeText(hunk_context, out); + writeChar('\t', out); + writeText(line, out); + writeChar('\t', out); + writeText(indent, out); + writeChar('\t', out); + writeText(line_type, out); + } +}; + +using LineChanges = std::vector; + +enum class FileChangeType +{ + Add, + Delete, + Modify, + Rename, + Copy, + Type, +}; + +void writeText(FileChangeType type, WriteBuffer & out) +{ + switch (type) + { + case FileChangeType::Add: writeString("Add", out); break; + case FileChangeType::Delete: writeString("Delete", out); break; + case FileChangeType::Modify: writeString("Modify", out); break; + case FileChangeType::Rename: writeString("Rename", out); break; + case FileChangeType::Copy: writeString("Copy", out); break; + case FileChangeType::Type: writeString("Type", out); break; + } +} + +struct FileChange +{ + FileChangeType change_type{}; + std::string new_file_path; + std::string old_file_path; + uint16_t lines_added{}; + uint16_t lines_deleted{}; + uint16_t hunks_added{}; + uint16_t hunks_removed{}; + uint16_t hunks_changed{}; + + void writeTextWithoutNewline(WriteBuffer & out) const + { + writeText(change_type, out); + writeChar('\t', out); + writeText(new_file_path, out); + writeChar('\t', out); + writeText(old_file_path, out); + writeChar('\t', out); + writeText(lines_added, out); + writeChar('\t', out); + writeText(lines_deleted, out); + writeChar('\t', out); + writeText(hunks_added, out); + writeChar('\t', out); + writeText(hunks_removed, out); + writeChar('\t', out); + writeText(hunks_changed, out); + } +}; + +struct FileChangeAndLineChanges +{ + FileChange file_change; + LineChanges line_changes; +}; + +struct Commit +{ + std::string hash; + std::string author_name; + std::string author_email; + time_t time{}; + std::string message; + uint32_t files_added{}; + uint32_t files_deleted{}; + uint32_t files_renamed{}; + uint32_t files_modified{}; + uint32_t lines_added{}; + uint32_t lines_deleted{}; + uint32_t hunks_added{}; + uint32_t hunks_removed{}; + uint32_t hunks_changed{}; + + void writeTextWithoutNewline(WriteBuffer & out) const + { + writeText(hash, out); + writeChar('\t', out); + writeText(author_name, out); + writeChar('\t', out); + writeText(author_email, out); + writeChar('\t', out); + writeText(time, out); + writeChar('\t', out); + writeText(message, out); + writeChar('\t', out); + writeText(files_added, out); + writeChar('\t', out); + writeText(files_deleted, out); + writeChar('\t', out); + writeText(files_renamed, out); + writeChar('\t', out); + writeText(files_modified, out); + writeChar('\t', out); + writeText(lines_added, out); + writeChar('\t', out); + writeText(lines_deleted, out); + writeChar('\t', out); + writeText(hunks_added, out); + writeChar('\t', out); + writeText(hunks_removed, out); + writeChar('\t', out); + writeText(hunks_changed, out); + } +}; + + +void skipUntilWhitespace(ReadBuffer & buf) +{ + while (!buf.eof()) + { + char * next_pos = find_first_symbols<'\t', '\n', ' '>(buf.position(), buf.buffer().end()); + buf.position() = next_pos; + + if (!buf.hasPendingData()) + continue; + + if (*buf.position() == '\t' || *buf.position() == '\n' || *buf.position() == ' ') + return; + } +} + +void skipUntilNextLine(ReadBuffer & buf) +{ + while (!buf.eof()) + { + char * next_pos = find_first_symbols<'\n'>(buf.position(), buf.buffer().end()); + buf.position() = next_pos; + + if (!buf.hasPendingData()) + continue; + + if (*buf.position() == '\n') + { + ++buf.position(); + return; + } + } +} + +void readStringUntilNextLine(std::string & s, ReadBuffer & buf) +{ + s.clear(); + while (!buf.eof()) + { + char * next_pos = find_first_symbols<'\n'>(buf.position(), buf.buffer().end()); + s.append(buf.position(), next_pos - buf.position()); + buf.position() = next_pos; + + if (!buf.hasPendingData()) + continue; + + if (*buf.position() == '\n') + { + ++buf.position(); + return; + } + } +} + + +struct Result +{ + WriteBufferFromFile commits{"commits.tsv"}; + WriteBufferFromFile file_changes{"file_changes.tsv"}; + WriteBufferFromFile line_changes{"line_changes.tsv"}; +}; + + +void processCommit(std::string hash, Result & result) +{ + std::string command = fmt::format( + "git show --raw --pretty='format:%at%x09%aN%x09%aE%x0A%s%x00' --patch --unified=0 {}", + hash); + + std::cerr << command << "\n"; + + auto commit_info = ShellCommand::execute(command); + auto & in = commit_info->out; + + Commit commit; + commit.hash = hash; + + readText(commit.time, in); + assertChar('\t', in); + readText(commit.author_name, in); + assertChar('\t', in); + readText(commit.author_email, in); + assertChar('\n', in); + readNullTerminated(commit.message, in); + + std::cerr << fmt::format("{}\t{}\n", toString(LocalDateTime(commit.time)), commit.message); + + if (!in.eof()) + assertChar('\n', in); + + /// File changes in form + /// :100644 100644 b90fe6bb94 3ffe4c380f M src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp + /// :100644 100644 828dedf6b5 828dedf6b5 R100 dbms/src/Functions/GeoUtils.h dbms/src/Functions/PolygonUtils.h + + std::map file_changes; + + while (checkChar(':', in)) + { + FileChange file_change; + + for (size_t i = 0; i < 4; ++i) + { + skipUntilWhitespace(in); + skipWhitespaceIfAny(in); + } + + char change_type; + readChar(change_type, in); + + int confidence; + switch (change_type) + { + case 'A': + file_change.change_type = FileChangeType::Add; + ++commit.files_added; + break; + case 'D': + file_change.change_type = FileChangeType::Delete; + ++commit.files_deleted; + break; + case 'M': + file_change.change_type = FileChangeType::Modify; + ++commit.files_modified; + break; + case 'R': + file_change.change_type = FileChangeType::Rename; + ++commit.files_renamed; + readText(confidence, in); + break; + case 'C': + file_change.change_type = FileChangeType::Copy; + readText(confidence, in); + break; + case 'T': + file_change.change_type = FileChangeType::Type; + break; + default: + throw Exception(ErrorCodes::INCORRECT_DATA, "Unexpected file change type: {}", change_type); + } + + skipWhitespaceIfAny(in); + + if (change_type == 'R' || change_type == 'C') + { + readText(file_change.old_file_path, in); + skipWhitespaceIfAny(in); + readText(file_change.new_file_path, in); + } + else + { + readText(file_change.new_file_path, in); + } + + assertChar('\n', in); + + file_changes.emplace( + file_change.new_file_path, + FileChangeAndLineChanges{ file_change, {} }); + } + + if (!in.eof()) + { + assertChar('\n', in); + + /// Diffs for every file in form of + /// --- a/src/Storages/StorageReplicatedMergeTree.cpp + /// +++ b/src/Storages/StorageReplicatedMergeTree.cpp + /// @@ -1387,2 +1387 @@ bool StorageReplicatedMergeTree::tryExecuteMerge(const LogEntry & entry) + /// - table_lock, entry.create_time, reserved_space, entry.deduplicate, + /// - entry.force_ttl); + /// + table_lock, entry.create_time, reserved_space, entry.deduplicate); + + std::string old_file_path; + std::string new_file_path; + FileChangeAndLineChanges * file_change_and_line_changes = nullptr; + LineChange line_change; + + while (!in.eof()) + { + if (checkString("@@ ", in)) + { + if (!file_change_and_line_changes) + { + auto file_name = new_file_path.empty() ? old_file_path : new_file_path; + auto it = file_changes.find(file_name); + if (file_changes.end() == it) + std::cerr << fmt::format("Warning: skipping bad file name {}\n", file_name); + else + file_change_and_line_changes = &it->second; + } + + if (file_change_and_line_changes) + { + uint16_t old_lines = 1; + uint16_t new_lines = 1; + + assertChar('-', in); + readText(line_change.hunk_start_line_number_old, in); + if (checkChar(',', in)) + readText(old_lines, in); + + assertString(" +", in); + readText(line_change.hunk_start_line_number_new, in); + if (checkChar(',', in)) + readText(new_lines, in); + + assertString(" @@", in); + if (checkChar(' ', in)) + readStringUntilNextLine(line_change.hunk_context, in); + else + assertChar('\n', in); + + ++line_change.hunk_num; + line_change.line_number_old = line_change.hunk_start_line_number_old; + line_change.line_number_new = line_change.hunk_start_line_number_new; + + if (old_lines && new_lines) + { + ++commit.hunks_changed; + ++file_change_and_line_changes->file_change.hunks_changed; + } + else if (old_lines) + { + ++commit.hunks_removed; + ++file_change_and_line_changes->file_change.hunks_removed; + } + else if (new_lines) + { + ++commit.hunks_added; + ++file_change_and_line_changes->file_change.hunks_added; + } + } + } + else if (checkChar('-', in)) + { + if (checkString("-- ", in)) + { + if (checkString("a/", in)) + { + readStringUntilNextLine(old_file_path, in); + line_change = LineChange{}; + file_change_and_line_changes = nullptr; + } + else if (checkString("/dev/null", in)) + { + old_file_path.clear(); + assertChar('\n', in); + line_change = LineChange{}; + file_change_and_line_changes = nullptr; + } + else + skipUntilNextLine(in); /// Actually it can be the line in diff. Skip it for simplicity. + } + else + { + if (file_change_and_line_changes) + { + ++commit.lines_deleted; + + line_change.sign = -1; + readStringUntilNextLine(line_change.line, in); + line_change.setLineInfo(line_change.line); + + file_change_and_line_changes->line_changes.push_back(line_change); + ++line_change.line_number_old; + } + } + } + else if (checkChar('+', in)) + { + if (checkString("++ ", in)) + { + if (checkString("b/", in)) + { + readStringUntilNextLine(new_file_path, in); + line_change = LineChange{}; + file_change_and_line_changes = nullptr; + } + else if (checkString("/dev/null", in)) + { + new_file_path.clear(); + assertChar('\n', in); + line_change = LineChange{}; + file_change_and_line_changes = nullptr; + } + else + skipUntilNextLine(in); /// Actually it can be the line in diff. Skip it for simplicity. + } + else + { + if (file_change_and_line_changes) + { + ++commit.lines_added; + + line_change.sign = 1; + readStringUntilNextLine(line_change.line, in); + line_change.setLineInfo(line_change.line); + + file_change_and_line_changes->line_changes.push_back(line_change); + ++line_change.line_number_new; + } + } + } + else + { + skipUntilNextLine(in); + } + } + } + + /// Write the result + + /// commits table + { + auto & out = result.commits; + + commit.writeTextWithoutNewline(out); + writeChar('\n', out); + } + + for (const auto & elem : file_changes) + { + const FileChange & file_change = elem.second.file_change; + + /// file_changes table + { + auto & out = result.file_changes; + + file_change.writeTextWithoutNewline(out); + writeChar('\t', out); + commit.writeTextWithoutNewline(out); + writeChar('\n', out); + } + + /// line_changes table + for (const auto & line_change : elem.second.line_changes) + { + auto & out = result.line_changes; + + line_change.writeTextWithoutNewline(out); + writeChar('\t', out); + file_change.writeTextWithoutNewline(out); + writeChar('\t', out); + commit.writeTextWithoutNewline(out); + writeChar('\n', out); + } + } +} + + +void processLog() +{ + Result result; + + std::string command = "git log --no-merges --pretty=%H"; + std::cerr << command << "\n"; + auto git_log = ShellCommand::execute(command); + + auto & in = git_log->out; + while (!in.eof()) + { + std::string hash; + readString(hash, in); + assertChar('\n', in); + + std::cerr << fmt::format("Processing commit {}\n", hash); + processCommit(std::move(hash), result); + } +} + + +} + +int main(int /*argc*/, char ** /*argv*/) +try +{ + using namespace DB; + +/* boost::program_options::options_description desc("Allowed options"); + desc.add_options()("help,h", "produce help message"); + + boost::program_options::variables_map options; + boost::program_options::store(boost::program_options::parse_command_line(argc, argv, desc), options); + + if (options.count("help") || argc != 2) + { + std::cout << "Usage: " << argv[0] << std::endl; + std::cout << desc << std::endl; + return 1; + }*/ + + processLog(); + return 0; +} +catch (...) +{ + std::cerr << DB::getCurrentExceptionMessage(true) << '\n'; + throw; +} From 338a6e20f60bb21c99ee2c4f261d96bc55ec4b97 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Fri, 4 Sep 2020 09:12:16 +0300 Subject: [PATCH 011/121] Added a script to import git repository to ClickHouse --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 97 +++++++++++++++++++ 1 file changed, 97 insertions(+) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index 42920328ad7..314bba0d5b4 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -16,6 +16,101 @@ #include +/** How to use: + +DROP DATABASE IF EXISTS git; +CREATE DATABASE git; + +CREATE TABLE git.commits +( + hash String, + author_name LowCardinality(String), + author_email LowCardinality(String), + time DateTime, + message String, + files_added UInt32, + files_deleted UInt32, + files_renamed UInt32, + files_modified UInt32, + lines_added UInt32, + lines_deleted UInt32, + hunks_added UInt32, + hunks_removed UInt32, + hunks_changed UInt32 +) ENGINE = MergeTree ORDER BY time; + +CREATE TABLE git.file_changes +( + change_type Enum('Add' = 1, 'Delete' = 2, 'Modify' = 3, 'Rename' = 4, 'Copy' = 5, 'Type' = 6), + new_file_path LowCardinality(String), + old_file_path LowCardinality(String), + lines_added UInt16, + lines_deleted UInt16, + hunks_added UInt16, + hunks_removed UInt16, + hunks_changed UInt16, + + commit_hash String, + author_name LowCardinality(String), + author_email LowCardinality(String), + time DateTime, + commit_message String, + commit_files_added UInt32, + commit_files_deleted UInt32, + commit_files_renamed UInt32, + commit_files_modified UInt32, + commit_lines_added UInt32, + commit_lines_deleted UInt32, + commit_hunks_added UInt32, + commit_hunks_removed UInt32, + commit_hunks_changed UInt32 +) ENGINE = MergeTree ORDER BY time; + +CREATE TABLE git.line_changes +( + sign Int8, + line_number_old UInt16, + line_number_new UInt16, + hunk_num UInt16, + hunk_start_line_number_old UInt16, + hunk_start_line_number_new UInt16, + hunk_context LowCardinality(String), + line LowCardinality(String), + indent UInt8, + line_type Enum('Empty' = 0, 'Comment' = 1, 'Punct' = 2, 'Code' = 3), + + file_change_type Enum('Add' = 1, 'Delete' = 2, 'Modify' = 3, 'Rename' = 4, 'Copy' = 5, 'Type' = 6), + new_file_path LowCardinality(String), + old_file_path LowCardinality(String), + file_lines_added UInt16, + file_lines_deleted UInt16, + file_hunks_added UInt16, + file_hunks_removed UInt16, + file_hunks_changed UInt16, + + commit_hash String, + author_name LowCardinality(String), + author_email LowCardinality(String), + time DateTime, + commit_message String, + commit_files_added UInt32, + commit_files_deleted UInt32, + commit_files_renamed UInt32, + commit_files_modified UInt32, + commit_lines_added UInt32, + commit_lines_deleted UInt32, + commit_hunks_added UInt32, + commit_hunks_removed UInt32, + commit_hunks_changed UInt32 +) ENGINE = MergeTree ORDER BY time; + +clickhouse-client --query "INSERT INTO git.commits FORMAT TSV" < commits.tsv +clickhouse-client --query "INSERT INTO git.file_changes FORMAT TSV" < file_changes.tsv +clickhouse-client --query "INSERT INTO git.line_changes FORMAT TSV" < line_changes.tsv + + */ + + namespace DB { @@ -495,6 +590,7 @@ void processCommit(std::string hash, Result & result) if (file_change_and_line_changes) { ++commit.lines_deleted; + ++file_change_and_line_changes->file_change.lines_deleted; line_change.sign = -1; readStringUntilNextLine(line_change.line, in); @@ -530,6 +626,7 @@ void processCommit(std::string hash, Result & result) if (file_change_and_line_changes) { ++commit.lines_added; + ++file_change_and_line_changes->file_change.lines_added; line_change.sign = 1; readStringUntilNextLine(line_change.line, in); From 7b95e56e8c902578f8fcebc5d9edeccce1eb35ee Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Sun, 6 Sep 2020 03:09:40 +0300 Subject: [PATCH 012/121] Advancements --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 187 +++++++++++++----- 1 file changed, 133 insertions(+), 54 deletions(-) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index 314bba0d5b4..d6264a63978 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -1,6 +1,11 @@ #include #include #include +#include +#include +#include + +#include #include @@ -16,7 +21,8 @@ #include -/** How to use: +static constexpr auto documentation = R"( +Prepare the database by executing the following queries: DROP DATABASE IF EXISTS git; CREATE DATABASE git; @@ -44,11 +50,11 @@ CREATE TABLE git.file_changes change_type Enum('Add' = 1, 'Delete' = 2, 'Modify' = 3, 'Rename' = 4, 'Copy' = 5, 'Type' = 6), new_file_path LowCardinality(String), old_file_path LowCardinality(String), - lines_added UInt16, - lines_deleted UInt16, - hunks_added UInt16, - hunks_removed UInt16, - hunks_changed UInt16, + lines_added UInt32, + lines_deleted UInt32, + hunks_added UInt32, + hunks_removed UInt32, + hunks_changed UInt32, commit_hash String, author_name LowCardinality(String), @@ -69,11 +75,11 @@ CREATE TABLE git.file_changes CREATE TABLE git.line_changes ( sign Int8, - line_number_old UInt16, - line_number_new UInt16, - hunk_num UInt16, - hunk_start_line_number_old UInt16, - hunk_start_line_number_new UInt16, + line_number_old UInt32, + line_number_new UInt32, + hunk_num UInt32, + hunk_start_line_number_old UInt32, + hunk_start_line_number_new UInt32, hunk_context LowCardinality(String), line LowCardinality(String), indent UInt8, @@ -82,11 +88,11 @@ CREATE TABLE git.line_changes file_change_type Enum('Add' = 1, 'Delete' = 2, 'Modify' = 3, 'Rename' = 4, 'Copy' = 5, 'Type' = 6), new_file_path LowCardinality(String), old_file_path LowCardinality(String), - file_lines_added UInt16, - file_lines_deleted UInt16, - file_hunks_added UInt16, - file_hunks_removed UInt16, - file_hunks_changed UInt16, + file_lines_added UInt32, + file_lines_deleted UInt32, + file_hunks_added UInt32, + file_hunks_removed UInt32, + file_hunks_changed UInt32, commit_hash String, author_name LowCardinality(String), @@ -104,12 +110,15 @@ CREATE TABLE git.line_changes commit_hunks_changed UInt32 ) ENGINE = MergeTree ORDER BY time; +Insert the data with the following commands: + clickhouse-client --query "INSERT INTO git.commits FORMAT TSV" < commits.tsv clickhouse-client --query "INSERT INTO git.file_changes FORMAT TSV" < file_changes.tsv clickhouse-client --query "INSERT INTO git.line_changes FORMAT TSV" < line_changes.tsv - */ +)"; +namespace po = boost::program_options; namespace DB { @@ -141,11 +150,11 @@ void writeText(LineType type, WriteBuffer & out) struct LineChange { int8_t sign{}; /// 1 if added, -1 if deleted - uint16_t line_number_old{}; - uint16_t line_number_new{}; - uint16_t hunk_num{}; /// ordinal number of hunk in diff, starting with 0 - uint16_t hunk_start_line_number_old{}; - uint16_t hunk_start_line_number_new{}; + uint32_t line_number_old{}; + uint32_t line_number_new{}; + uint32_t hunk_num{}; /// ordinal number of hunk in diff, starting with 0 + uint32_t hunk_start_line_number_old{}; + uint32_t hunk_start_line_number_new{}; std::string hunk_context; /// The context (like a line with function name) as it is calculated by git std::string line; /// Line content without leading whitespaces uint8_t indent{}; /// The number of leading whitespaces or tabs * 4 @@ -251,11 +260,11 @@ struct FileChange FileChangeType change_type{}; std::string new_file_path; std::string old_file_path; - uint16_t lines_added{}; - uint16_t lines_deleted{}; - uint16_t hunks_added{}; - uint16_t hunks_removed{}; - uint16_t hunks_changed{}; + uint32_t lines_added{}; + uint32_t lines_deleted{}; + uint32_t hunks_added{}; + uint32_t hunks_removed{}; + uint32_t hunks_changed{}; void writeTextWithoutNewline(WriteBuffer & out) const { @@ -395,13 +404,38 @@ struct Result }; -void processCommit(std::string hash, Result & result) +struct Options +{ + bool skip_commits_without_parents = true; + std::optional skip_paths; + std::unordered_set skip_commits; + size_t diff_size_limit = 0; + + Options(const po::variables_map & options) + { + skip_commits_without_parents = options["skip-commits-without-parents"].as(); + if (options.count("skip-paths")) + { + skip_paths.emplace(options["skip-paths"].as()); + } + if (options.count("skip-commit")) + { + auto vec = options["skip-commit"].as>(); + skip_commits.insert(vec.begin(), vec.end()); + } + diff_size_limit = options["diff-size-limit"].as(); + } +}; + + +void processCommit( + const Options & options, size_t commit_num, size_t total_commits, std::string hash, Result & result) { std::string command = fmt::format( - "git show --raw --pretty='format:%at%x09%aN%x09%aE%x0A%s%x00' --patch --unified=0 {}", + "git show --raw --pretty='format:%at%x09%aN%x09%aE%x09%P%x0A%s%x00' --patch --unified=0 {}", hash); - std::cerr << command << "\n"; + //std::cerr << command << "\n"; auto commit_info = ShellCommand::execute(command); auto & in = commit_info->out; @@ -414,10 +448,23 @@ void processCommit(std::string hash, Result & result) readText(commit.author_name, in); assertChar('\t', in); readText(commit.author_email, in); + assertChar('\t', in); + std::string parent_hash; + readString(parent_hash, in); assertChar('\n', in); readNullTerminated(commit.message, in); - std::cerr << fmt::format("{}\t{}\n", toString(LocalDateTime(commit.time)), commit.message); + std::string message_to_print = commit.message; + std::replace_if(message_to_print.begin(), message_to_print.end(), [](char c){ return std::iscntrl(c); }, ' '); + + fmt::print("{}% {} {} {}\n", + commit_num * 100 / total_commits, toString(LocalDateTime(commit.time)), hash, message_to_print); + + if (options.skip_commits_without_parents && commit_num != 0 && parent_hash.empty()) + { + std::cerr << "Warning: skipping commit without parents\n"; + return; + } if (!in.eof()) assertChar('\n', in); @@ -487,9 +534,12 @@ void processCommit(std::string hash, Result & result) assertChar('\n', in); - file_changes.emplace( - file_change.new_file_path, - FileChangeAndLineChanges{ file_change, {} }); + if (!(options.skip_paths && re2_st::RE2::PartialMatch(file_change.new_file_path, *options.skip_paths))) + { + file_changes.emplace( + file_change.new_file_path, + FileChangeAndLineChanges{ file_change, {} }); + } } if (!in.eof()) @@ -517,16 +567,14 @@ void processCommit(std::string hash, Result & result) { auto file_name = new_file_path.empty() ? old_file_path : new_file_path; auto it = file_changes.find(file_name); - if (file_changes.end() == it) - std::cerr << fmt::format("Warning: skipping bad file name {}\n", file_name); - else + if (file_changes.end() != it) file_change_and_line_changes = &it->second; } if (file_change_and_line_changes) { - uint16_t old_lines = 1; - uint16_t new_lines = 1; + uint32_t old_lines = 1; + uint32_t new_lines = 1; assertChar('-', in); readText(line_change.hunk_start_line_number_old, in); @@ -644,6 +692,9 @@ void processCommit(std::string hash, Result & result) } } + if (commit.lines_added + commit.lines_deleted > options.diff_size_limit) + return; + /// Write the result /// commits table @@ -684,14 +735,20 @@ void processCommit(std::string hash, Result & result) } -void processLog() +void processLog(const Options & options) { Result result; - std::string command = "git log --no-merges --pretty=%H"; - std::cerr << command << "\n"; + std::string command = "git log --reverse --no-merges --pretty=%H"; + fmt::print("{}\n", command); auto git_log = ShellCommand::execute(command); + /// Collect hashes in memory. This is inefficient but allows to display beautiful progress. + /// The number of commits is in order of single millions for the largest repositories, + /// so don't care about potential waste of ~100 MB of memory. + + std::vector hashes; + auto & in = git_log->out; while (!in.eof()) { @@ -699,33 +756,55 @@ void processLog() readString(hash, in); assertChar('\n', in); - std::cerr << fmt::format("Processing commit {}\n", hash); - processCommit(std::move(hash), result); + if (!options.skip_commits.count(hash)) + hashes.emplace_back(std::move(hash)); + } + + size_t num_commits = hashes.size(); + fmt::print("Total {} commits to process.\n", num_commits); + + for (size_t i = 0; i < num_commits; ++i) + { + processCommit(options, i, num_commits, hashes[i], result); } } } -int main(int /*argc*/, char ** /*argv*/) +int main(int argc, char ** argv) try { using namespace DB; -/* boost::program_options::options_description desc("Allowed options"); - desc.add_options()("help,h", "produce help message"); + po::options_description desc("Allowed options"); + desc.add_options() + ("help,h", "produce help message") + ("skip-commits-without-parents", po::value()->default_value(true), + "Skip commits without parents (except the initial commit)." + " These commits are usually erroneous but they can make sense in very rare cases.") + ("skip-paths", po::value(), + "Skip paths that matches regular expression (re2 syntax).") + ("skip-commit", po::value>(), + "Skip commit with specified hash. The option can be specified multiple times.") + ("diff-size-limit", po::value()->default_value(0), + "Skip commits whose diff size (number of added + removed lines) is larger than specified threshold") + ; - boost::program_options::variables_map options; - boost::program_options::store(boost::program_options::parse_command_line(argc, argv, desc), options); + po::variables_map options; + po::store(boost::program_options::parse_command_line(argc, argv, desc), options); - if (options.count("help") || argc != 2) + if (options.count("help")) { - std::cout << "Usage: " << argv[0] << std::endl; - std::cout << desc << std::endl; + std::cout << documentation << '\n' + << "Usage: " << argv[0] << '\n' + << desc << '\n' + << "\nExample:\n" + << "\n./git-to-clickhouse --diff-size-limit 100000 --skip-paths '^(contrib|docs?|website|libs/(libcityhash|liblz4|libdivide|libvectorclass|libdouble-conversion|libcpuid|libzstd|libfarmhash|libmetrohash|libpoco|libwidechar_width))/'\n"; return 1; - }*/ + } - processLog(); + processLog(options); return 0; } catch (...) From abe836a584aeaf71b0ba04b8c8cc670385519e94 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Sun, 6 Sep 2020 03:13:39 +0300 Subject: [PATCH 013/121] Remove emails as they are mostly useless --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 20 ++++++------------- 1 file changed, 6 insertions(+), 14 deletions(-) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index d6264a63978..9203efb0043 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -30,8 +30,7 @@ CREATE DATABASE git; CREATE TABLE git.commits ( hash String, - author_name LowCardinality(String), - author_email LowCardinality(String), + author LowCardinality(String), time DateTime, message String, files_added UInt32, @@ -57,8 +56,7 @@ CREATE TABLE git.file_changes hunks_changed UInt32, commit_hash String, - author_name LowCardinality(String), - author_email LowCardinality(String), + author LowCardinality(String), time DateTime, commit_message String, commit_files_added UInt32, @@ -95,8 +93,7 @@ CREATE TABLE git.line_changes file_hunks_changed UInt32, commit_hash String, - author_name LowCardinality(String), - author_email LowCardinality(String), + author LowCardinality(String), time DateTime, commit_message String, commit_files_added UInt32, @@ -295,8 +292,7 @@ struct FileChangeAndLineChanges struct Commit { std::string hash; - std::string author_name; - std::string author_email; + std::string author; time_t time{}; std::string message; uint32_t files_added{}; @@ -313,9 +309,7 @@ struct Commit { writeText(hash, out); writeChar('\t', out); - writeText(author_name, out); - writeChar('\t', out); - writeText(author_email, out); + writeText(author, out); writeChar('\t', out); writeText(time, out); writeChar('\t', out); @@ -445,9 +439,7 @@ void processCommit( readText(commit.time, in); assertChar('\t', in); - readText(commit.author_name, in); - assertChar('\t', in); - readText(commit.author_email, in); + readText(commit.author, in); assertChar('\t', in); std::string parent_hash; readString(parent_hash, in); From 09978decbdf40c95e7cd8855ad804a2ad31cc09d Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Sun, 6 Sep 2020 03:47:00 +0300 Subject: [PATCH 014/121] Adjustments --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 52 +++++++++++++------ 1 file changed, 35 insertions(+), 17 deletions(-) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index 9203efb0043..a81bc6679a7 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -4,6 +4,7 @@ #include #include #include +#include #include @@ -47,8 +48,9 @@ CREATE TABLE git.commits CREATE TABLE git.file_changes ( change_type Enum('Add' = 1, 'Delete' = 2, 'Modify' = 3, 'Rename' = 4, 'Copy' = 5, 'Type' = 6), - new_file_path LowCardinality(String), - old_file_path LowCardinality(String), + path LowCardinality(String), + old_path LowCardinality(String), + file_extension LowCardinality(String), lines_added UInt32, lines_deleted UInt32, hunks_added UInt32, @@ -84,8 +86,9 @@ CREATE TABLE git.line_changes line_type Enum('Empty' = 0, 'Comment' = 1, 'Punct' = 2, 'Code' = 3), file_change_type Enum('Add' = 1, 'Delete' = 2, 'Modify' = 3, 'Rename' = 4, 'Copy' = 5, 'Type' = 6), - new_file_path LowCardinality(String), - old_file_path LowCardinality(String), + path LowCardinality(String), + old_path LowCardinality(String), + file_extension LowCardinality(String), file_lines_added UInt32, file_lines_deleted UInt32, file_hunks_added UInt32, @@ -255,8 +258,9 @@ void writeText(FileChangeType type, WriteBuffer & out) struct FileChange { FileChangeType change_type{}; - std::string new_file_path; - std::string old_file_path; + std::string path; + std::string old_path; + std::string file_extension; uint32_t lines_added{}; uint32_t lines_deleted{}; uint32_t hunks_added{}; @@ -267,9 +271,11 @@ struct FileChange { writeText(change_type, out); writeChar('\t', out); - writeText(new_file_path, out); + writeText(path, out); writeChar('\t', out); - writeText(old_file_path, out); + writeText(old_path, out); + writeChar('\t', out); + writeText(file_extension, out); writeChar('\t', out); writeText(lines_added, out); writeChar('\t', out); @@ -422,11 +428,20 @@ struct Options }; +/// Rough snapshot of repository calculated by application of diffs. It's used to calculate blame info. +struct File +{ + std::vector lines; +}; + +using Snapshot = std::map; + + void processCommit( - const Options & options, size_t commit_num, size_t total_commits, std::string hash, Result & result) + const Options & options, size_t commit_num, size_t total_commits, std::string hash, Snapshot & /*snapshot*/, Result & result) { std::string command = fmt::format( - "git show --raw --pretty='format:%at%x09%aN%x09%aE%x09%P%x0A%s%x00' --patch --unified=0 {}", + "git show --raw --pretty='format:%at%x09%aN%x09%P%x0A%s%x00' --patch --unified=0 {}", hash); //std::cerr << command << "\n"; @@ -515,21 +530,23 @@ void processCommit( if (change_type == 'R' || change_type == 'C') { - readText(file_change.old_file_path, in); + readText(file_change.old_path, in); skipWhitespaceIfAny(in); - readText(file_change.new_file_path, in); + readText(file_change.path, in); } else { - readText(file_change.new_file_path, in); + readText(file_change.path, in); } + file_change.file_extension = std::filesystem::path(file_change.path).extension(); + assertChar('\n', in); - if (!(options.skip_paths && re2_st::RE2::PartialMatch(file_change.new_file_path, *options.skip_paths))) + if (!(options.skip_paths && re2_st::RE2::PartialMatch(file_change.path, *options.skip_paths))) { file_changes.emplace( - file_change.new_file_path, + file_change.path, FileChangeAndLineChanges{ file_change, {} }); } } @@ -755,9 +772,10 @@ void processLog(const Options & options) size_t num_commits = hashes.size(); fmt::print("Total {} commits to process.\n", num_commits); + Snapshot snapshot; for (size_t i = 0; i < num_commits; ++i) { - processCommit(options, i, num_commits, hashes[i], result); + processCommit(options, i, num_commits, hashes[i], snapshot, result); } } @@ -792,7 +810,7 @@ try << "Usage: " << argv[0] << '\n' << desc << '\n' << "\nExample:\n" - << "\n./git-to-clickhouse --diff-size-limit 100000 --skip-paths '^(contrib|docs?|website|libs/(libcityhash|liblz4|libdivide|libvectorclass|libdouble-conversion|libcpuid|libzstd|libfarmhash|libmetrohash|libpoco|libwidechar_width))/'\n"; + << "\n./git-to-clickhouse --diff-size-limit 100000 --skip-paths 'generated\\.cpp|^(contrib|docs?|website|libs/(libcityhash|liblz4|libdivide|libvectorclass|libdouble-conversion|libcpuid|libzstd|libfarmhash|libmetrohash|libpoco|libwidechar_width))/'\n"; return 1; } From d1f1326a1370abd5d837864d02851ef1b3b20745 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Sun, 6 Sep 2020 07:02:17 +0300 Subject: [PATCH 015/121] Concurrent processing + history --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 411 ++++++++++++------ 1 file changed, 283 insertions(+), 128 deletions(-) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index a81bc6679a7..6686c1ac480 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -4,6 +4,8 @@ #include #include #include +#include +#include #include #include @@ -80,11 +82,17 @@ CREATE TABLE git.line_changes hunk_num UInt32, hunk_start_line_number_old UInt32, hunk_start_line_number_new UInt32, + hunk_lines_added UInt32, + hunk_lines_deleted UInt32, hunk_context LowCardinality(String), line LowCardinality(String), indent UInt8, line_type Enum('Empty' = 0, 'Comment' = 1, 'Punct' = 2, 'Code' = 3), + prev_commit_hash String, + prev_author LowCardinality(String), + prev_time DateTime, + file_change_type Enum('Add' = 1, 'Delete' = 2, 'Modify' = 3, 'Rename' = 4, 'Copy' = 5, 'Type' = 6), path LowCardinality(String), old_path LowCardinality(String), @@ -128,6 +136,112 @@ namespace ErrorCodes extern const int INCORRECT_DATA; } + +struct Commit +{ + std::string hash; + std::string author; + LocalDateTime time{}; + std::string message; + uint32_t files_added{}; + uint32_t files_deleted{}; + uint32_t files_renamed{}; + uint32_t files_modified{}; + uint32_t lines_added{}; + uint32_t lines_deleted{}; + uint32_t hunks_added{}; + uint32_t hunks_removed{}; + uint32_t hunks_changed{}; + + void writeTextWithoutNewline(WriteBuffer & out) const + { + writeText(hash, out); + writeChar('\t', out); + writeText(author, out); + writeChar('\t', out); + writeText(time, out); + writeChar('\t', out); + writeText(message, out); + writeChar('\t', out); + writeText(files_added, out); + writeChar('\t', out); + writeText(files_deleted, out); + writeChar('\t', out); + writeText(files_renamed, out); + writeChar('\t', out); + writeText(files_modified, out); + writeChar('\t', out); + writeText(lines_added, out); + writeChar('\t', out); + writeText(lines_deleted, out); + writeChar('\t', out); + writeText(hunks_added, out); + writeChar('\t', out); + writeText(hunks_removed, out); + writeChar('\t', out); + writeText(hunks_changed, out); + } +}; + + +enum class FileChangeType +{ + Add, + Delete, + Modify, + Rename, + Copy, + Type, +}; + +void writeText(FileChangeType type, WriteBuffer & out) +{ + switch (type) + { + case FileChangeType::Add: writeString("Add", out); break; + case FileChangeType::Delete: writeString("Delete", out); break; + case FileChangeType::Modify: writeString("Modify", out); break; + case FileChangeType::Rename: writeString("Rename", out); break; + case FileChangeType::Copy: writeString("Copy", out); break; + case FileChangeType::Type: writeString("Type", out); break; + } +} + +struct FileChange +{ + FileChangeType change_type{}; + std::string path; + std::string old_path; + std::string file_extension; + uint32_t lines_added{}; + uint32_t lines_deleted{}; + uint32_t hunks_added{}; + uint32_t hunks_removed{}; + uint32_t hunks_changed{}; + + void writeTextWithoutNewline(WriteBuffer & out) const + { + writeText(change_type, out); + writeChar('\t', out); + writeText(path, out); + writeChar('\t', out); + writeText(old_path, out); + writeChar('\t', out); + writeText(file_extension, out); + writeChar('\t', out); + writeText(lines_added, out); + writeChar('\t', out); + writeText(lines_deleted, out); + writeChar('\t', out); + writeText(hunks_added, out); + writeChar('\t', out); + writeText(hunks_removed, out); + writeChar('\t', out); + writeText(hunks_changed, out); + } +}; + + enum class LineType { Empty, @@ -155,10 +269,15 @@ struct LineChange uint32_t hunk_num{}; /// ordinal number of hunk in diff, starting with 0 uint32_t hunk_start_line_number_old{}; uint32_t hunk_start_line_number_new{}; + uint32_t hunk_lines_added{}; + uint32_t hunk_lines_deleted{}; std::string hunk_context; /// The context (like a line with function name) as it is calculated by git std::string line; /// Line content without leading whitespaces uint8_t indent{}; /// The number of leading whitespaces or tabs * 4 LineType line_type{}; + std::string prev_commit_hash; + std::string prev_author; + LocalDateTime prev_time{}; void setLineInfo(std::string full_line) { @@ -220,6 +339,10 @@ struct LineChange writeChar('\t', out); writeText(hunk_start_line_number_new, out); writeChar('\t', out); + writeText(hunk_lines_added, out); + writeChar('\t', out); + writeText(hunk_lines_deleted, out); + writeChar('\t', out); writeText(hunk_context, out); writeChar('\t', out); writeText(line, out); @@ -227,120 +350,17 @@ struct LineChange writeText(indent, out); writeChar('\t', out); writeText(line_type, out); + writeChar('\t', out); + writeText(prev_commit_hash, out); + writeChar('\t', out); + writeText(prev_author, out); + writeChar('\t', out); + writeText(prev_time, out); } }; using LineChanges = std::vector; -enum class FileChangeType -{ - Add, - Delete, - Modify, - Rename, - Copy, - Type, -}; - -void writeText(FileChangeType type, WriteBuffer & out) -{ - switch (type) - { - case FileChangeType::Add: writeString("Add", out); break; - case FileChangeType::Delete: writeString("Delete", out); break; - case FileChangeType::Modify: writeString("Modify", out); break; - case FileChangeType::Rename: writeString("Rename", out); break; - case FileChangeType::Copy: writeString("Copy", out); break; - case FileChangeType::Type: writeString("Type", out); break; - } -} - -struct FileChange -{ - FileChangeType change_type{}; - std::string path; - std::string old_path; - std::string file_extension; - uint32_t lines_added{}; - uint32_t lines_deleted{}; - uint32_t hunks_added{}; - uint32_t hunks_removed{}; - uint32_t hunks_changed{}; - - void writeTextWithoutNewline(WriteBuffer & out) const - { - writeText(change_type, out); - writeChar('\t', out); - writeText(path, out); - writeChar('\t', out); - writeText(old_path, out); - writeChar('\t', out); - writeText(file_extension, out); - writeChar('\t', out); - writeText(lines_added, out); - writeChar('\t', out); - writeText(lines_deleted, out); - writeChar('\t', out); - writeText(hunks_added, out); - writeChar('\t', out); - writeText(hunks_removed, out); - writeChar('\t', out); - writeText(hunks_changed, out); - } -}; - -struct FileChangeAndLineChanges -{ - FileChange file_change; - LineChanges line_changes; -}; - -struct Commit -{ - std::string hash; - std::string author; - time_t time{}; - std::string message; - uint32_t files_added{}; - uint32_t files_deleted{}; - uint32_t files_renamed{}; - uint32_t files_modified{}; - uint32_t lines_added{}; - uint32_t lines_deleted{}; - uint32_t hunks_added{}; - uint32_t hunks_removed{}; - uint32_t hunks_changed{}; - - void writeTextWithoutNewline(WriteBuffer & out) const - { - writeText(hash, out); - writeChar('\t', out); - writeText(author, out); - writeChar('\t', out); - writeText(time, out); - writeChar('\t', out); - writeText(message, out); - writeChar('\t', out); - writeText(files_added, out); - writeChar('\t', out); - writeText(files_deleted, out); - writeChar('\t', out); - writeText(files_renamed, out); - writeChar('\t', out); - writeText(files_modified, out); - writeChar('\t', out); - writeText(lines_added, out); - writeChar('\t', out); - writeText(lines_deleted, out); - writeChar('\t', out); - writeText(hunks_added, out); - writeChar('\t', out); - writeText(hunks_removed, out); - writeChar('\t', out); - writeText(hunks_changed, out); - } -}; - void skipUntilWhitespace(ReadBuffer & buf) { @@ -407,13 +427,15 @@ struct Result struct Options { bool skip_commits_without_parents = true; + size_t threads = 1; std::optional skip_paths; std::unordered_set skip_commits; - size_t diff_size_limit = 0; + std::optional diff_size_limit; Options(const po::variables_map & options) { skip_commits_without_parents = options["skip-commits-without-parents"].as(); + threads = options["threads"].as(); if (options.count("skip-paths")) { skip_paths.emplace(options["skip-paths"].as()); @@ -423,36 +445,123 @@ struct Options auto vec = options["skip-commit"].as>(); skip_commits.insert(vec.begin(), vec.end()); } - diff_size_limit = options["diff-size-limit"].as(); + if (options.count("diff-size-limit")) + { + diff_size_limit = options["diff-size-limit"].as(); + } } }; /// Rough snapshot of repository calculated by application of diffs. It's used to calculate blame info. -struct File +struct FileBlame { - std::vector lines; + using Lines = std::list; + Lines lines; + Lines::iterator it; + size_t current_idx = 1; + + FileBlame() + { + it = lines.begin(); + } + + FileBlame & operator=(const FileBlame & rhs) + { + lines = rhs.lines; + it = lines.begin(); + current_idx = 1; + return *this; + } + + FileBlame(const FileBlame & rhs) + { + *this = rhs; + } + + void walk(uint32_t num) + { + if (current_idx < num) + { + while (current_idx < num && it != lines.end()) + { + ++current_idx; + ++it; + } + } + else if (current_idx > num) + { + --current_idx; + --it; + } + } + + const Commit * find(uint32_t num) + { + walk(num); + + if (current_idx == num && it != lines.end()) + return &*it; + return {}; + } + + void addLine(uint32_t num, Commit commit) + { + walk(num); + + while (it == lines.end() && current_idx < num) + { + lines.emplace_back(); + ++current_idx; + } + if (it == lines.end()) + { + lines.emplace_back(); + --it; + } + + lines.insert(it, commit); + } + + void removeLine(uint32_t num) + { + walk(num); + + if (current_idx == num) + it = lines.erase(it); + } }; -using Snapshot = std::map; +using Snapshot = std::map; + +struct FileChangeAndLineChanges +{ + FileChangeAndLineChanges(FileChange file_change_) : file_change(file_change_) {} + + FileChange file_change; + LineChanges line_changes; + + std::map deleted_lines; +}; void processCommit( - const Options & options, size_t commit_num, size_t total_commits, std::string hash, Snapshot & /*snapshot*/, Result & result) + std::unique_ptr & commit_info, + const Options & options, + size_t commit_num, + size_t total_commits, + std::string hash, + Snapshot & snapshot, + Result & result) { - std::string command = fmt::format( - "git show --raw --pretty='format:%at%x09%aN%x09%P%x0A%s%x00' --patch --unified=0 {}", - hash); - - //std::cerr << command << "\n"; - - auto commit_info = ShellCommand::execute(command); auto & in = commit_info->out; Commit commit; commit.hash = hash; - readText(commit.time, in); + time_t commit_time; + readText(commit_time, in); + commit.time = commit_time; assertChar('\t', in); readText(commit.author, in); assertChar('\t', in); @@ -465,7 +574,7 @@ void processCommit( std::replace_if(message_to_print.begin(), message_to_print.end(), [](char c){ return std::iscntrl(c); }, ' '); fmt::print("{}% {} {} {}\n", - commit_num * 100 / total_commits, toString(LocalDateTime(commit.time)), hash, message_to_print); + commit_num * 100 / total_commits, toString(commit.time), hash, message_to_print); if (options.skip_commits_without_parents && commit_num != 0 && parent_hash.empty()) { @@ -533,6 +642,8 @@ void processCommit( readText(file_change.old_path, in); skipWhitespaceIfAny(in); readText(file_change.path, in); + + snapshot[file_change.path] = snapshot[file_change.old_path]; } else { @@ -547,7 +658,7 @@ void processCommit( { file_changes.emplace( file_change.path, - FileChangeAndLineChanges{ file_change, {} }); + FileChangeAndLineChanges(file_change)); } } @@ -601,6 +712,9 @@ void processCommit( else assertChar('\n', in); + line_change.hunk_lines_added = new_lines; + line_change.hunk_lines_deleted = old_lines; + ++line_change.hunk_num; line_change.line_number_old = line_change.hunk_start_line_number_old; line_change.line_number_new = line_change.hunk_start_line_number_new; @@ -653,6 +767,16 @@ void processCommit( readStringUntilNextLine(line_change.line, in); line_change.setLineInfo(line_change.line); + FileBlame & file_snapshot = snapshot[old_file_path]; + if (const Commit * prev_commit = file_snapshot.find(line_change.line_number_old)) + { + line_change.prev_commit_hash = prev_commit->hash; + line_change.prev_author = prev_commit->author; + line_change.prev_time = prev_commit->time; + file_change_and_line_changes->deleted_lines[line_change.line_number_old] = *prev_commit; + file_snapshot.removeLine(line_change.line_number_old); + } + file_change_and_line_changes->line_changes.push_back(line_change); ++line_change.line_number_old; } @@ -689,6 +813,16 @@ void processCommit( readStringUntilNextLine(line_change.line, in); line_change.setLineInfo(line_change.line); + FileBlame & file_snapshot = snapshot[new_file_path]; + if (file_change_and_line_changes->deleted_lines.count(line_change.line_number_new)) + { + const auto & prev_commit = file_change_and_line_changes->deleted_lines[line_change.line_number_new]; + line_change.prev_commit_hash = prev_commit.hash; + line_change.prev_author = prev_commit.author; + line_change.prev_time = prev_commit.time; + } + file_snapshot.addLine(line_change.line_number_new, commit); + file_change_and_line_changes->line_changes.push_back(line_change); ++line_change.line_number_new; } @@ -701,7 +835,7 @@ void processCommit( } } - if (commit.lines_added + commit.lines_deleted > options.diff_size_limit) + if (options.diff_size_limit && commit.lines_added + commit.lines_deleted > *options.diff_size_limit) return; /// Write the result @@ -744,6 +878,16 @@ void processCommit( } +auto gitShow(const std::string & hash) +{ + std::string command = fmt::format( + "git show --raw --pretty='format:%at%x09%aN%x09%P%x0A%s%x00' --patch --unified=0 {}", + hash); + + return ShellCommand::execute(command); +} + + void processLog(const Options & options) { Result result; @@ -772,10 +916,19 @@ void processLog(const Options & options) size_t num_commits = hashes.size(); fmt::print("Total {} commits to process.\n", num_commits); + /// Will run multiple processes in parallel + size_t num_threads = options.threads; + + std::vector> show_commands(num_threads); + for (size_t i = 0; i < num_commits && i < num_threads; ++i) + show_commands[i] = gitShow(hashes[i]); + Snapshot snapshot; for (size_t i = 0; i < num_commits; ++i) { - processCommit(options, i, num_commits, hashes[i], snapshot, result); + processCommit(show_commands[i % num_threads], options, i, num_commits, hashes[i], snapshot, result); + if (i + num_threads < num_commits) + show_commands[i % num_threads] = gitShow(hashes[i + num_threads]); } } @@ -797,8 +950,10 @@ try "Skip paths that matches regular expression (re2 syntax).") ("skip-commit", po::value>(), "Skip commit with specified hash. The option can be specified multiple times.") - ("diff-size-limit", po::value()->default_value(0), + ("diff-size-limit", po::value(), "Skip commits whose diff size (number of added + removed lines) is larger than specified threshold") + ("threads", po::value()->default_value(std::thread::hardware_concurrency()), + "Number of threads to interact with git") ; po::variables_map options; From 3f29453c02ef3d3716927d81258218516b183d7b Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Sun, 6 Sep 2020 09:38:48 +0300 Subject: [PATCH 016/121] Roughly working blame --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 179 ++++++++++++++---- 1 file changed, 137 insertions(+), 42 deletions(-) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index 6686c1ac480..c1c27a82812 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -4,6 +4,7 @@ #include #include #include +#include #include #include #include @@ -13,6 +14,7 @@ #include #include +#include #include #include #include @@ -427,19 +429,26 @@ struct Result struct Options { bool skip_commits_without_parents = true; + bool skip_commits_with_duplicate_diffs = true; size_t threads = 1; std::optional skip_paths; + std::optional skip_commits_with_messages; std::unordered_set skip_commits; std::optional diff_size_limit; Options(const po::variables_map & options) { skip_commits_without_parents = options["skip-commits-without-parents"].as(); + skip_commits_with_duplicate_diffs = options["skip-commits-with-duplicate-diffs"].as(); threads = options["threads"].as(); if (options.count("skip-paths")) { skip_paths.emplace(options["skip-paths"].as()); } + if (options.count("skip-commits-with-messages")) + { + skip_commits_with_messages.emplace(options["skip-commits-with-messages"].as()); + } if (options.count("skip-commit")) { auto vec = options["skip-commit"].as>(); @@ -481,15 +490,12 @@ struct FileBlame void walk(uint32_t num) { - if (current_idx < num) + while (current_idx < num && it != lines.end()) { - while (current_idx < num && it != lines.end()) - { - ++current_idx; - ++it; - } + ++current_idx; + ++it; } - else if (current_idx > num) + while (current_idx > num) { --current_idx; --it; @@ -500,6 +506,8 @@ struct FileBlame { walk(num); +// std::cerr << "current_idx: " << current_idx << ", num: " << num << "\n"; + if (current_idx == num && it != lines.end()) return &*it; return {}; @@ -514,20 +522,17 @@ struct FileBlame lines.emplace_back(); ++current_idx; } - if (it == lines.end()) - { - lines.emplace_back(); - --it; - } - lines.insert(it, commit); + it = lines.insert(it, commit); } void removeLine(uint32_t num) { +// std::cerr << "Removing line " << num << ", current_idx: " << current_idx << "\n"; + walk(num); - if (current_idx == num) + if (current_idx == num && it != lines.end()) it = lines.erase(it); } }; @@ -540,10 +545,10 @@ struct FileChangeAndLineChanges FileChange file_change; LineChanges line_changes; - - std::map deleted_lines; }; +using DiffHashes = std::unordered_set; + void processCommit( std::unique_ptr & commit_info, @@ -552,6 +557,7 @@ void processCommit( size_t total_commits, std::string hash, Snapshot & snapshot, + DiffHashes & diff_hashes, Result & result) { auto & in = commit_info->out; @@ -570,6 +576,9 @@ void processCommit( assertChar('\n', in); readNullTerminated(commit.message, in); + if (options.skip_commits_with_messages && re2_st::RE2::PartialMatch(commit.message, *options.skip_commits_with_messages)) + return; + std::string message_to_print = commit.message; std::replace_if(message_to_print.begin(), message_to_print.end(), [](char c){ return std::iscntrl(c); }, ' '); @@ -643,7 +652,10 @@ void processCommit( skipWhitespaceIfAny(in); readText(file_change.path, in); - snapshot[file_change.path] = snapshot[file_change.old_path]; +// std::cerr << "Move from " << file_change.old_path << " to " << file_change.path << "\n"; + + if (file_change.path != file_change.old_path) + snapshot[file_change.path] = snapshot[file_change.old_path]; } else { @@ -706,6 +718,9 @@ void processCommit( if (checkChar(',', in)) readText(new_lines, in); + if (line_change.hunk_start_line_number_new == 0) + line_change.hunk_start_line_number_new = 1; + assertString(" @@", in); if (checkChar(' ', in)) readStringUntilNextLine(line_change.hunk_context, in); @@ -767,16 +782,6 @@ void processCommit( readStringUntilNextLine(line_change.line, in); line_change.setLineInfo(line_change.line); - FileBlame & file_snapshot = snapshot[old_file_path]; - if (const Commit * prev_commit = file_snapshot.find(line_change.line_number_old)) - { - line_change.prev_commit_hash = prev_commit->hash; - line_change.prev_author = prev_commit->author; - line_change.prev_time = prev_commit->time; - file_change_and_line_changes->deleted_lines[line_change.line_number_old] = *prev_commit; - file_snapshot.removeLine(line_change.line_number_old); - } - file_change_and_line_changes->line_changes.push_back(line_change); ++line_change.line_number_old; } @@ -813,16 +818,6 @@ void processCommit( readStringUntilNextLine(line_change.line, in); line_change.setLineInfo(line_change.line); - FileBlame & file_snapshot = snapshot[new_file_path]; - if (file_change_and_line_changes->deleted_lines.count(line_change.line_number_new)) - { - const auto & prev_commit = file_change_and_line_changes->deleted_lines[line_change.line_number_new]; - line_change.prev_commit_hash = prev_commit.hash; - line_change.prev_author = prev_commit.author; - line_change.prev_time = prev_commit.time; - } - file_snapshot.addLine(line_change.line_number_new, commit); - file_change_and_line_changes->line_changes.push_back(line_change); ++line_change.line_number_new; } @@ -838,6 +833,99 @@ void processCommit( if (options.diff_size_limit && commit.lines_added + commit.lines_deleted > *options.diff_size_limit) return; + /// Calculate hash of diff and skip duplicates + if (options.skip_commits_with_duplicate_diffs) + { + SipHash hasher; + + for (auto & elem : file_changes) + { + hasher.update(elem.second.file_change.change_type); + hasher.update(elem.second.file_change.old_path.size()); + hasher.update(elem.second.file_change.old_path); + hasher.update(elem.second.file_change.path.size()); + hasher.update(elem.second.file_change.path); + + hasher.update(elem.second.line_changes.size()); + for (auto & line_change : elem.second.line_changes) + { + hasher.update(line_change.sign); + hasher.update(line_change.line_number_old); + hasher.update(line_change.line_number_new); + hasher.update(line_change.indent); + hasher.update(line_change.line.size()); + hasher.update(line_change.line); + } + } + + UInt128 hash_of_diff; + hasher.get128(hash_of_diff.low, hash_of_diff.high); + + if (!diff_hashes.insert(hash_of_diff).second) + return; + } + + /// Update snapshot and blame info + + for (auto & elem : file_changes) + { +// std::cerr << elem.first << "\n"; + + FileBlame & file_snapshot = snapshot[elem.first]; + std::unordered_map deleted_lines; + + /// Obtain blame info from previous state of the snapshot + + for (auto & line_change : elem.second.line_changes) + { + if (line_change.sign == -1) + { + if (const Commit * prev_commit = file_snapshot.find(line_change.line_number_old); + prev_commit && prev_commit->time <= commit.time) + { + line_change.prev_commit_hash = prev_commit->hash; + line_change.prev_author = prev_commit->author; + line_change.prev_time = prev_commit->time; + deleted_lines[line_change.line_number_old] = *prev_commit; + } + else + { + // std::cerr << "Did not find line " << line_change.line_number_old << " from file " << elem.first << ": " << line_change.line << "\n"; + } + } + else if (line_change.sign == 1) + { + uint32_t this_line_in_prev_commit = line_change.hunk_start_line_number_old + + (line_change.line_number_new - line_change.hunk_start_line_number_new); + + if (deleted_lines.count(this_line_in_prev_commit)) + { + const auto & prev_commit = deleted_lines[this_line_in_prev_commit]; + if (prev_commit.time <= commit.time) + { + line_change.prev_commit_hash = prev_commit.hash; + line_change.prev_author = prev_commit.author; + line_change.prev_time = prev_commit.time; + } + } + } + } + + /// Update the snapshot + + for (const auto & line_change : elem.second.line_changes) + { + if (line_change.sign == -1) + { + file_snapshot.removeLine(line_change.line_number_new); + } + else if (line_change.sign == 1) + { + file_snapshot.addLine(line_change.line_number_new, commit); + } + } + } + /// Write the result /// commits table @@ -881,7 +969,7 @@ void processCommit( auto gitShow(const std::string & hash) { std::string command = fmt::format( - "git show --raw --pretty='format:%at%x09%aN%x09%P%x0A%s%x00' --patch --unified=0 {}", + "git show --raw --pretty='format:%ct%x09%aN%x09%P%x0A%s%x00' --patch --unified=0 {}", hash); return ShellCommand::execute(command); @@ -924,9 +1012,11 @@ void processLog(const Options & options) show_commands[i] = gitShow(hashes[i]); Snapshot snapshot; + DiffHashes diff_hashes; + for (size_t i = 0; i < num_commits; ++i) { - processCommit(show_commands[i % num_threads], options, i, num_commits, hashes[i], snapshot, result); + processCommit(show_commands[i % num_threads], options, i, num_commits, hashes[i], snapshot, diff_hashes, result); if (i + num_threads < num_commits) show_commands[i % num_threads] = gitShow(hashes[i + num_threads]); } @@ -946,10 +1036,15 @@ try ("skip-commits-without-parents", po::value()->default_value(true), "Skip commits without parents (except the initial commit)." " These commits are usually erroneous but they can make sense in very rare cases.") - ("skip-paths", po::value(), - "Skip paths that matches regular expression (re2 syntax).") + ("skip-commits-with-duplicate-diffs", po::value()->default_value(true), + "Skip commits with duplicate diffs." + " These commits are usually results of cherry-pick or merge after rebase.") ("skip-commit", po::value>(), "Skip commit with specified hash. The option can be specified multiple times.") + ("skip-paths", po::value(), + "Skip paths that matches regular expression (re2 syntax).") + ("skip-commits-with-messages", po::value(), + "Skip commits whose messages matches regular expression (re2 syntax).") ("diff-size-limit", po::value(), "Skip commits whose diff size (number of added + removed lines) is larger than specified threshold") ("threads", po::value()->default_value(std::thread::hardware_concurrency()), @@ -965,7 +1060,7 @@ try << "Usage: " << argv[0] << '\n' << desc << '\n' << "\nExample:\n" - << "\n./git-to-clickhouse --diff-size-limit 100000 --skip-paths 'generated\\.cpp|^(contrib|docs?|website|libs/(libcityhash|liblz4|libdivide|libvectorclass|libdouble-conversion|libcpuid|libzstd|libfarmhash|libmetrohash|libpoco|libwidechar_width))/'\n"; + << "\n./git-to-clickhouse --diff-size-limit 100000 --skip-paths 'generated\\.cpp|^(contrib|docs?|website|libs/(libcityhash|liblz4|libdivide|libvectorclass|libdouble-conversion|libcpuid|libzstd|libfarmhash|libmetrohash|libpoco|libwidechar_width))/' --skip-commits-with-messages '^Merge branch '\n"; return 1; } From 99c33612d65c627bbb9fc31d9d97906195d3cf53 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Sun, 6 Sep 2020 10:29:58 +0300 Subject: [PATCH 017/121] Better diagnostics --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index c1c27a82812..6b29708ead3 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -582,7 +582,7 @@ void processCommit( std::string message_to_print = commit.message; std::replace_if(message_to_print.begin(), message_to_print.end(), [](char c){ return std::iscntrl(c); }, ' '); - fmt::print("{}% {} {} {}\n", + std::cerr << fmt::format("{}% {} {} {}\n", commit_num * 100 / total_commits, toString(commit.time), hash, message_to_print); if (options.skip_commits_without_parents && commit_num != 0 && parent_hash.empty()) From 3ec9656aa21a3142d2898b7d259a4740a6691fd2 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Sun, 6 Sep 2020 10:38:39 +0300 Subject: [PATCH 018/121] Slightly more robust --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index 6b29708ead3..f3653bb282f 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -568,12 +568,10 @@ void processCommit( time_t commit_time; readText(commit_time, in); commit.time = commit_time; - assertChar('\t', in); - readText(commit.author, in); - assertChar('\t', in); + assertChar('\0', in); + readNullTerminated(commit.author, in); std::string parent_hash; - readString(parent_hash, in); - assertChar('\n', in); + readNullTerminated(parent_hash, in); readNullTerminated(commit.message, in); if (options.skip_commits_with_messages && re2_st::RE2::PartialMatch(commit.message, *options.skip_commits_with_messages)) @@ -969,7 +967,7 @@ void processCommit( auto gitShow(const std::string & hash) { std::string command = fmt::format( - "git show --raw --pretty='format:%ct%x09%aN%x09%P%x0A%s%x00' --patch --unified=0 {}", + "git show --raw --pretty='format:%ct%x00%aN%x00%P%x00%s%x00' --patch --unified=0 {}", hash); return ShellCommand::execute(command); From db58fa15aaf202318e043549440589797b51aa0a Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Mon, 7 Sep 2020 02:24:31 +0300 Subject: [PATCH 019/121] Some tweaks --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 21 ++++++++++++++----- 1 file changed, 16 insertions(+), 5 deletions(-) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index f3653bb282f..9e1ef14fcbf 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -435,6 +435,7 @@ struct Options std::optional skip_commits_with_messages; std::unordered_set skip_commits; std::optional diff_size_limit; + std::string stop_after_commit; Options(const po::variables_map & options) { @@ -458,6 +459,10 @@ struct Options { diff_size_limit = options["diff-size-limit"].as(); } + if (options.count("stop-after-commit")) + { + stop_after_commit = options["stop-after-commit"].as(); + } } }; @@ -828,7 +833,7 @@ void processCommit( } } - if (options.diff_size_limit && commit.lines_added + commit.lines_deleted > *options.diff_size_limit) + if (options.diff_size_limit && commit_num != 0 && commit.lines_added + commit.lines_deleted > *options.diff_size_limit) return; /// Calculate hash of diff and skip duplicates @@ -1015,6 +1020,10 @@ void processLog(const Options & options) for (size_t i = 0; i < num_commits; ++i) { processCommit(show_commands[i % num_threads], options, i, num_commits, hashes[i], snapshot, diff_hashes, result); + + if (!options.stop_after_commit.empty() && hashes[i] == options.stop_after_commit) + break; + if (i + num_threads < num_commits) show_commands[i % num_threads] = gitShow(hashes[i + num_threads]); } @@ -1043,10 +1052,12 @@ try "Skip paths that matches regular expression (re2 syntax).") ("skip-commits-with-messages", po::value(), "Skip commits whose messages matches regular expression (re2 syntax).") - ("diff-size-limit", po::value(), - "Skip commits whose diff size (number of added + removed lines) is larger than specified threshold") + ("diff-size-limit", po::value()->default_value(100000), + "Skip commits whose diff size (number of added + removed lines) is larger than specified threshold. Does not apply for initial commit.") + ("stop-after-commit", po::value(), + "Stop processing after specified commit hash.") ("threads", po::value()->default_value(std::thread::hardware_concurrency()), - "Number of threads to interact with git") + "Number of concurrent git subprocesses to spawn") ; po::variables_map options; @@ -1058,7 +1069,7 @@ try << "Usage: " << argv[0] << '\n' << desc << '\n' << "\nExample:\n" - << "\n./git-to-clickhouse --diff-size-limit 100000 --skip-paths 'generated\\.cpp|^(contrib|docs?|website|libs/(libcityhash|liblz4|libdivide|libvectorclass|libdouble-conversion|libcpuid|libzstd|libfarmhash|libmetrohash|libpoco|libwidechar_width))/' --skip-commits-with-messages '^Merge branch '\n"; + << "\n./git-to-clickhouse --skip-paths 'generated\\.cpp|^(contrib|docs?|website|libs/(libcityhash|liblz4|libdivide|libvectorclass|libdouble-conversion|libcpuid|libzstd|libfarmhash|libmetrohash|libpoco|libwidechar_width))/' --skip-commits-with-messages '^Merge branch '\n"; return 1; } From 684a910395cc37203453d1faa09ab839d3a4f32a Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Mon, 7 Sep 2020 03:17:26 +0300 Subject: [PATCH 020/121] Polish --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 803 ++++++++++-------- 1 file changed, 451 insertions(+), 352 deletions(-) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index 9e1ef14fcbf..6e43853d6ba 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -277,10 +277,14 @@ struct LineChange std::string line; /// Line content without leading whitespaces uint8_t indent{}; /// The number of leading whitespaces or tabs * 4 LineType line_type{}; + /// Information from the history (blame). std::string prev_commit_hash; std::string prev_author; LocalDateTime prev_time{}; + /** Classify line to empty / code / comment / single punctuation char. + * Very rough and mostly suitable for our C++ style. + */ void setLineInfo(std::string full_line) { indent = 0; @@ -306,8 +310,9 @@ struct LineChange line_type = LineType::Empty; } else if (pos + 1 < end - && ((pos[0] == '/' && pos[1] == '/') - || (pos[0] == '*' && pos[1] == ' '))) /// This is not precise. + && ((pos[0] == '/' && (pos[1] == '/' || pos[1] == '*')) + || (pos[0] == '*' && pos[1] == ' ') /// This is not precise. + || (pos[0] == '#' && pos[1] == ' '))) { line_type = LineType::Comment; } @@ -363,6 +368,18 @@ struct LineChange using LineChanges = std::vector; +struct FileDiff +{ + FileDiff(FileChange file_change_) : file_change(file_change_) {} + + FileChange file_change; + LineChanges line_changes; +}; + +using CommitDiff = std::map; + + +/** Parsing helpers */ void skipUntilWhitespace(ReadBuffer & buf) { @@ -418,14 +435,57 @@ void readStringUntilNextLine(std::string & s, ReadBuffer & buf) } -struct Result +/** Writes the resulting tables to files that can be imported to ClickHouse. + */ +struct ResultWriter { WriteBufferFromFile commits{"commits.tsv"}; WriteBufferFromFile file_changes{"file_changes.tsv"}; WriteBufferFromFile line_changes{"line_changes.tsv"}; + + void appendCommit(const Commit & commit, const CommitDiff & files) + { + /// commits table + { + auto & out = commits; + + commit.writeTextWithoutNewline(out); + writeChar('\n', out); + } + + for (const auto & elem : files) + { + const FileChange & file_change = elem.second.file_change; + + /// file_changes table + { + auto & out = file_changes; + + file_change.writeTextWithoutNewline(out); + writeChar('\t', out); + commit.writeTextWithoutNewline(out); + writeChar('\n', out); + } + + /// line_changes table + for (const auto & line_change : elem.second.line_changes) + { + auto & out = line_changes; + + line_change.writeTextWithoutNewline(out); + writeChar('\t', out); + file_change.writeTextWithoutNewline(out); + writeChar('\t', out); + commit.writeTextWithoutNewline(out); + writeChar('\n', out); + } + } + } }; +/** See description in "main". + */ struct Options { bool skip_commits_without_parents = true; @@ -467,11 +527,23 @@ struct Options }; -/// Rough snapshot of repository calculated by application of diffs. It's used to calculate blame info. +/** Rough snapshot of repository calculated by application of diffs. It's used to calculate blame info. + * Represented by a list of lines. For every line it contains information about commit that modified this line the last time. + * + * Note that there are many cases when this info may become incorrect. + * The first reason is that git history is non-linear but we form this snapshot by application of commit diffs in some order + * that cannot give us correct results even theoretically. + * The second reason is that we don't process merge commits. But merge commits may contain differences for conflict resolution. + * + * We expect that the information will be mostly correct for the purpose of analytics. + * So, it can provide the expected "blame" info for the most of the lines. + */ struct FileBlame { using Lines = std::list; Lines lines; + + /// We walk through this list adding or removing lines. Lines::iterator it; size_t current_idx = 1; @@ -480,6 +552,7 @@ struct FileBlame it = lines.begin(); } + /// This is important when file was copied or renamed. FileBlame & operator=(const FileBlame & rhs) { lines = rhs.lines; @@ -493,6 +566,7 @@ struct FileBlame *this = rhs; } + /// Move iterator to requested line or stop at the end. void walk(uint32_t num) { while (current_idx < num && it != lines.end()) @@ -522,6 +596,7 @@ struct FileBlame { walk(num); + /// If the inserted line is over the end of file, we insert empty lines before it. while (it == lines.end() && current_idx < num) { lines.emplace_back(); @@ -542,334 +617,24 @@ struct FileBlame } }; +/// All files with their blame info. When file is renamed, we also rename it in snapshot. using Snapshot = std::map; -struct FileChangeAndLineChanges + +/** Enrich the line changes data with the history info from the snapshot + * - the author, time and commit of the previous change to every found line (blame). + * And update the snapshot. + */ +void updateSnapshot(Snapshot & snapshot, const Commit & commit, CommitDiff & file_changes) { - FileChangeAndLineChanges(FileChange file_change_) : file_change(file_change_) {} - - FileChange file_change; - LineChanges line_changes; -}; - -using DiffHashes = std::unordered_set; - - -void processCommit( - std::unique_ptr & commit_info, - const Options & options, - size_t commit_num, - size_t total_commits, - std::string hash, - Snapshot & snapshot, - DiffHashes & diff_hashes, - Result & result) -{ - auto & in = commit_info->out; - - Commit commit; - commit.hash = hash; - - time_t commit_time; - readText(commit_time, in); - commit.time = commit_time; - assertChar('\0', in); - readNullTerminated(commit.author, in); - std::string parent_hash; - readNullTerminated(parent_hash, in); - readNullTerminated(commit.message, in); - - if (options.skip_commits_with_messages && re2_st::RE2::PartialMatch(commit.message, *options.skip_commits_with_messages)) - return; - - std::string message_to_print = commit.message; - std::replace_if(message_to_print.begin(), message_to_print.end(), [](char c){ return std::iscntrl(c); }, ' '); - - std::cerr << fmt::format("{}% {} {} {}\n", - commit_num * 100 / total_commits, toString(commit.time), hash, message_to_print); - - if (options.skip_commits_without_parents && commit_num != 0 && parent_hash.empty()) + /// Renames and copies. + for (auto & elem : file_changes) { - std::cerr << "Warning: skipping commit without parents\n"; - return; + auto & file = elem.second.file_change; + if (file.path != file.old_path) + snapshot[file.path] = snapshot[file.old_path]; } - if (!in.eof()) - assertChar('\n', in); - - /// File changes in form - /// :100644 100644 b90fe6bb94 3ffe4c380f M src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp - /// :100644 100644 828dedf6b5 828dedf6b5 R100 dbms/src/Functions/GeoUtils.h dbms/src/Functions/PolygonUtils.h - - std::map file_changes; - - while (checkChar(':', in)) - { - FileChange file_change; - - for (size_t i = 0; i < 4; ++i) - { - skipUntilWhitespace(in); - skipWhitespaceIfAny(in); - } - - char change_type; - readChar(change_type, in); - - int confidence; - switch (change_type) - { - case 'A': - file_change.change_type = FileChangeType::Add; - ++commit.files_added; - break; - case 'D': - file_change.change_type = FileChangeType::Delete; - ++commit.files_deleted; - break; - case 'M': - file_change.change_type = FileChangeType::Modify; - ++commit.files_modified; - break; - case 'R': - file_change.change_type = FileChangeType::Rename; - ++commit.files_renamed; - readText(confidence, in); - break; - case 'C': - file_change.change_type = FileChangeType::Copy; - readText(confidence, in); - break; - case 'T': - file_change.change_type = FileChangeType::Type; - break; - default: - throw Exception(ErrorCodes::INCORRECT_DATA, "Unexpected file change type: {}", change_type); - } - - skipWhitespaceIfAny(in); - - if (change_type == 'R' || change_type == 'C') - { - readText(file_change.old_path, in); - skipWhitespaceIfAny(in); - readText(file_change.path, in); - -// std::cerr << "Move from " << file_change.old_path << " to " << file_change.path << "\n"; - - if (file_change.path != file_change.old_path) - snapshot[file_change.path] = snapshot[file_change.old_path]; - } - else - { - readText(file_change.path, in); - } - - file_change.file_extension = std::filesystem::path(file_change.path).extension(); - - assertChar('\n', in); - - if (!(options.skip_paths && re2_st::RE2::PartialMatch(file_change.path, *options.skip_paths))) - { - file_changes.emplace( - file_change.path, - FileChangeAndLineChanges(file_change)); - } - } - - if (!in.eof()) - { - assertChar('\n', in); - - /// Diffs for every file in form of - /// --- a/src/Storages/StorageReplicatedMergeTree.cpp - /// +++ b/src/Storages/StorageReplicatedMergeTree.cpp - /// @@ -1387,2 +1387 @@ bool StorageReplicatedMergeTree::tryExecuteMerge(const LogEntry & entry) - /// - table_lock, entry.create_time, reserved_space, entry.deduplicate, - /// - entry.force_ttl); - /// + table_lock, entry.create_time, reserved_space, entry.deduplicate); - - std::string old_file_path; - std::string new_file_path; - FileChangeAndLineChanges * file_change_and_line_changes = nullptr; - LineChange line_change; - - while (!in.eof()) - { - if (checkString("@@ ", in)) - { - if (!file_change_and_line_changes) - { - auto file_name = new_file_path.empty() ? old_file_path : new_file_path; - auto it = file_changes.find(file_name); - if (file_changes.end() != it) - file_change_and_line_changes = &it->second; - } - - if (file_change_and_line_changes) - { - uint32_t old_lines = 1; - uint32_t new_lines = 1; - - assertChar('-', in); - readText(line_change.hunk_start_line_number_old, in); - if (checkChar(',', in)) - readText(old_lines, in); - - assertString(" +", in); - readText(line_change.hunk_start_line_number_new, in); - if (checkChar(',', in)) - readText(new_lines, in); - - if (line_change.hunk_start_line_number_new == 0) - line_change.hunk_start_line_number_new = 1; - - assertString(" @@", in); - if (checkChar(' ', in)) - readStringUntilNextLine(line_change.hunk_context, in); - else - assertChar('\n', in); - - line_change.hunk_lines_added = new_lines; - line_change.hunk_lines_deleted = old_lines; - - ++line_change.hunk_num; - line_change.line_number_old = line_change.hunk_start_line_number_old; - line_change.line_number_new = line_change.hunk_start_line_number_new; - - if (old_lines && new_lines) - { - ++commit.hunks_changed; - ++file_change_and_line_changes->file_change.hunks_changed; - } - else if (old_lines) - { - ++commit.hunks_removed; - ++file_change_and_line_changes->file_change.hunks_removed; - } - else if (new_lines) - { - ++commit.hunks_added; - ++file_change_and_line_changes->file_change.hunks_added; - } - } - } - else if (checkChar('-', in)) - { - if (checkString("-- ", in)) - { - if (checkString("a/", in)) - { - readStringUntilNextLine(old_file_path, in); - line_change = LineChange{}; - file_change_and_line_changes = nullptr; - } - else if (checkString("/dev/null", in)) - { - old_file_path.clear(); - assertChar('\n', in); - line_change = LineChange{}; - file_change_and_line_changes = nullptr; - } - else - skipUntilNextLine(in); /// Actually it can be the line in diff. Skip it for simplicity. - } - else - { - if (file_change_and_line_changes) - { - ++commit.lines_deleted; - ++file_change_and_line_changes->file_change.lines_deleted; - - line_change.sign = -1; - readStringUntilNextLine(line_change.line, in); - line_change.setLineInfo(line_change.line); - - file_change_and_line_changes->line_changes.push_back(line_change); - ++line_change.line_number_old; - } - } - } - else if (checkChar('+', in)) - { - if (checkString("++ ", in)) - { - if (checkString("b/", in)) - { - readStringUntilNextLine(new_file_path, in); - line_change = LineChange{}; - file_change_and_line_changes = nullptr; - } - else if (checkString("/dev/null", in)) - { - new_file_path.clear(); - assertChar('\n', in); - line_change = LineChange{}; - file_change_and_line_changes = nullptr; - } - else - skipUntilNextLine(in); /// Actually it can be the line in diff. Skip it for simplicity. - } - else - { - if (file_change_and_line_changes) - { - ++commit.lines_added; - ++file_change_and_line_changes->file_change.lines_added; - - line_change.sign = 1; - readStringUntilNextLine(line_change.line, in); - line_change.setLineInfo(line_change.line); - - file_change_and_line_changes->line_changes.push_back(line_change); - ++line_change.line_number_new; - } - } - } - else - { - skipUntilNextLine(in); - } - } - } - - if (options.diff_size_limit && commit_num != 0 && commit.lines_added + commit.lines_deleted > *options.diff_size_limit) - return; - - /// Calculate hash of diff and skip duplicates - if (options.skip_commits_with_duplicate_diffs) - { - SipHash hasher; - - for (auto & elem : file_changes) - { - hasher.update(elem.second.file_change.change_type); - hasher.update(elem.second.file_change.old_path.size()); - hasher.update(elem.second.file_change.old_path); - hasher.update(elem.second.file_change.path.size()); - hasher.update(elem.second.file_change.path); - - hasher.update(elem.second.line_changes.size()); - for (auto & line_change : elem.second.line_changes) - { - hasher.update(line_change.sign); - hasher.update(line_change.line_number_old); - hasher.update(line_change.line_number_new); - hasher.update(line_change.indent); - hasher.update(line_change.line.size()); - hasher.update(line_change.line); - } - } - - UInt128 hash_of_diff; - hasher.get128(hash_of_diff.low, hash_of_diff.high); - - if (!diff_hashes.insert(hash_of_diff).second) - return; - } - - /// Update snapshot and blame info - for (auto & elem : file_changes) { // std::cerr << elem.first << "\n"; @@ -928,47 +693,379 @@ void processCommit( } } } +} - /// Write the result - /// commits table +/** Deduplication of commits with identical diffs. + */ +using DiffHashes = std::unordered_set; + +UInt128 diffHash(const CommitDiff & file_changes) +{ + SipHash hasher; + + for (auto & elem : file_changes) { - auto & out = result.commits; + hasher.update(elem.second.file_change.change_type); + hasher.update(elem.second.file_change.old_path.size()); + hasher.update(elem.second.file_change.old_path); + hasher.update(elem.second.file_change.path.size()); + hasher.update(elem.second.file_change.path); - commit.writeTextWithoutNewline(out); - writeChar('\n', out); + hasher.update(elem.second.line_changes.size()); + for (auto & line_change : elem.second.line_changes) + { + hasher.update(line_change.sign); + hasher.update(line_change.line_number_old); + hasher.update(line_change.line_number_new); + hasher.update(line_change.indent); + hasher.update(line_change.line.size()); + hasher.update(line_change.line); + } } - for (const auto & elem : file_changes) + UInt128 hash_of_diff; + hasher.get128(hash_of_diff.low, hash_of_diff.high); + + return hash_of_diff; +} + + +/** File changes in form + * :100644 100644 b90fe6bb94 3ffe4c380f M src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp + * :100644 100644 828dedf6b5 828dedf6b5 R100 dbms/src/Functions/GeoUtils.h dbms/src/Functions/PolygonUtils.h + * according to the output of 'git show --raw' + */ +void processFileChanges( + ReadBuffer & in, + const Options & options, + Commit & commit, + CommitDiff & file_changes) +{ + while (checkChar(':', in)) { - const FileChange & file_change = elem.second.file_change; + FileChange file_change; - /// file_changes table + /// We don't care about file mode and content hashes. + for (size_t i = 0; i < 4; ++i) { - auto & out = result.file_changes; - - file_change.writeTextWithoutNewline(out); - writeChar('\t', out); - commit.writeTextWithoutNewline(out); - writeChar('\n', out); + skipUntilWhitespace(in); + skipWhitespaceIfAny(in); } - /// line_changes table - for (const auto & line_change : elem.second.line_changes) - { - auto & out = result.line_changes; + char change_type; + readChar(change_type, in); - line_change.writeTextWithoutNewline(out); - writeChar('\t', out); - file_change.writeTextWithoutNewline(out); - writeChar('\t', out); - commit.writeTextWithoutNewline(out); - writeChar('\n', out); + /// For rename and copy there is a number called "score". We ignore it. + int score; + + switch (change_type) + { + case 'A': + file_change.change_type = FileChangeType::Add; + ++commit.files_added; + break; + case 'D': + file_change.change_type = FileChangeType::Delete; + ++commit.files_deleted; + break; + case 'M': + file_change.change_type = FileChangeType::Modify; + ++commit.files_modified; + break; + case 'R': + file_change.change_type = FileChangeType::Rename; + ++commit.files_renamed; + readText(score, in); + break; + case 'C': + file_change.change_type = FileChangeType::Copy; + readText(score, in); + break; + case 'T': + file_change.change_type = FileChangeType::Type; + break; + default: + throw Exception(ErrorCodes::INCORRECT_DATA, "Unexpected file change type: {}", change_type); + } + + skipWhitespaceIfAny(in); + + if (change_type == 'R' || change_type == 'C') + { + readText(file_change.old_path, in); + skipWhitespaceIfAny(in); + readText(file_change.path, in); + } + else + { + readText(file_change.path, in); + } + + file_change.file_extension = std::filesystem::path(file_change.path).extension(); + /// It gives us extension in form of '.cpp'. There is a reason for it but we remove initial dot for simplicity. + if (!file_change.file_extension.empty() && file_change.file_extension.front() == '.') + file_change.file_extension = file_change.file_extension.substr(1, std::string::npos); + + assertChar('\n', in); + + if (!(options.skip_paths && re2_st::RE2::PartialMatch(file_change.path, *options.skip_paths))) + { + file_changes.emplace( + file_change.path, + FileDiff(file_change)); } } } +/** Process the list of diffs for every file from the result of "git show". + * Caveats: + * - changes in binary files can be ignored; + * - if a line content begins with '+' or '-' it will be skipped + * it means that if you store diffs in repository and "git show" will display diff-of-diff for you, + * it won't be processed correctly; + * - we expect some specific format of the diff; but it may actually depend on git config; + * - non-ASCII file names are not processed correctly (they will not be found and will be ignored). + */ +void processDiffs( + ReadBuffer & in, + std::optional size_limit, + Commit & commit, + CommitDiff & file_changes) +{ + std::string old_file_path; + std::string new_file_path; + FileDiff * file_change_and_line_changes = nullptr; + LineChange line_change; + + /// Diffs for every file in form of + /// --- a/src/Storages/StorageReplicatedMergeTree.cpp + /// +++ b/src/Storages/StorageReplicatedMergeTree.cpp + /// @@ -1387,2 +1387 @@ bool StorageReplicatedMergeTree::tryExecuteMerge(const LogEntry & entry) + /// - table_lock, entry.create_time, reserved_space, entry.deduplicate, + /// - entry.force_ttl); + /// + table_lock, entry.create_time, reserved_space, entry.deduplicate); + + size_t diff_size = 0; + while (!in.eof()) + { + if (checkString("@@ ", in)) + { + if (!file_change_and_line_changes) + { + auto file_name = new_file_path.empty() ? old_file_path : new_file_path; + auto it = file_changes.find(file_name); + if (file_changes.end() != it) + file_change_and_line_changes = &it->second; + } + + if (file_change_and_line_changes) + { + uint32_t old_lines = 1; + uint32_t new_lines = 1; + + assertChar('-', in); + readText(line_change.hunk_start_line_number_old, in); + if (checkChar(',', in)) + readText(old_lines, in); + + assertString(" +", in); + readText(line_change.hunk_start_line_number_new, in); + if (checkChar(',', in)) + readText(new_lines, in); + + /// This is needed to simplify the logic of updating snapshot: + /// When all lines are removed we can treat it as repeated removal of line with number 1. + if (line_change.hunk_start_line_number_new == 0) + line_change.hunk_start_line_number_new = 1; + + assertString(" @@", in); + if (checkChar(' ', in)) + readStringUntilNextLine(line_change.hunk_context, in); + else + assertChar('\n', in); + + line_change.hunk_lines_added = new_lines; + line_change.hunk_lines_deleted = old_lines; + + ++line_change.hunk_num; + line_change.line_number_old = line_change.hunk_start_line_number_old; + line_change.line_number_new = line_change.hunk_start_line_number_new; + + if (old_lines && new_lines) + { + ++commit.hunks_changed; + ++file_change_and_line_changes->file_change.hunks_changed; + } + else if (old_lines) + { + ++commit.hunks_removed; + ++file_change_and_line_changes->file_change.hunks_removed; + } + else if (new_lines) + { + ++commit.hunks_added; + ++file_change_and_line_changes->file_change.hunks_added; + } + } + } + else if (checkChar('-', in)) + { + if (checkString("-- ", in)) + { + if (checkString("a/", in)) + { + readStringUntilNextLine(old_file_path, in); + line_change = LineChange{}; + file_change_and_line_changes = nullptr; + } + else if (checkString("/dev/null", in)) + { + old_file_path.clear(); + assertChar('\n', in); + line_change = LineChange{}; + file_change_and_line_changes = nullptr; + } + else + skipUntilNextLine(in); /// Actually it can be the line in diff. Skip it for simplicity. + } + else + { + ++diff_size; + if (file_change_and_line_changes) + { + ++commit.lines_deleted; + ++file_change_and_line_changes->file_change.lines_deleted; + + line_change.sign = -1; + readStringUntilNextLine(line_change.line, in); + line_change.setLineInfo(line_change.line); + + file_change_and_line_changes->line_changes.push_back(line_change); + ++line_change.line_number_old; + } + } + } + else if (checkChar('+', in)) + { + if (checkString("++ ", in)) + { + if (checkString("b/", in)) + { + readStringUntilNextLine(new_file_path, in); + line_change = LineChange{}; + file_change_and_line_changes = nullptr; + } + else if (checkString("/dev/null", in)) + { + new_file_path.clear(); + assertChar('\n', in); + line_change = LineChange{}; + file_change_and_line_changes = nullptr; + } + else + skipUntilNextLine(in); /// Actually it can be the line in diff. Skip it for simplicity. + } + else + { + ++diff_size; + if (file_change_and_line_changes) + { + ++commit.lines_added; + ++file_change_and_line_changes->file_change.lines_added; + + line_change.sign = 1; + readStringUntilNextLine(line_change.line, in); + line_change.setLineInfo(line_change.line); + + file_change_and_line_changes->line_changes.push_back(line_change); + ++line_change.line_number_new; + } + } + } + else + { + /// Unknown lines are ignored. + skipUntilNextLine(in); + } + + if (size_limit && diff_size > *size_limit) + return; + } +} + + +/** Process the "git show" result for a single commit. Append the result to tables. + */ +void processCommit( + ReadBuffer & in, + const Options & options, + size_t commit_num, + size_t total_commits, + std::string hash, + Snapshot & snapshot, + DiffHashes & diff_hashes, + ResultWriter & result) +{ + Commit commit; + commit.hash = hash; + + time_t commit_time; + readText(commit_time, in); + commit.time = commit_time; + assertChar('\0', in); + readNullTerminated(commit.author, in); + std::string parent_hash; + readNullTerminated(parent_hash, in); + readNullTerminated(commit.message, in); + + if (options.skip_commits_with_messages && re2_st::RE2::PartialMatch(commit.message, *options.skip_commits_with_messages)) + return; + + std::string message_to_print = commit.message; + std::replace_if(message_to_print.begin(), message_to_print.end(), [](char c){ return std::iscntrl(c); }, ' '); + + std::cerr << fmt::format("{}% {} {} {}\n", + commit_num * 100 / total_commits, toString(commit.time), hash, message_to_print); + + if (options.skip_commits_without_parents && commit_num != 0 && parent_hash.empty()) + { + std::cerr << "Warning: skipping commit without parents\n"; + return; + } + + if (!in.eof()) + assertChar('\n', in); + + CommitDiff file_changes; + processFileChanges(in, options, commit, file_changes); + + if (!in.eof()) + { + assertChar('\n', in); + processDiffs(in, commit_num != 0 ? options.diff_size_limit : std::nullopt, commit, file_changes); + } + + /// Skip commits with too large diffs. + if (options.diff_size_limit && commit_num != 0 && commit.lines_added + commit.lines_deleted > *options.diff_size_limit) + return; + + /// Calculate hash of diff and skip duplicates + if (options.skip_commits_with_duplicate_diffs && !diff_hashes.insert(diffHash(file_changes)).second) + return; + + /// Update snapshot and blame info + updateSnapshot(snapshot, commit, file_changes); + + /// Write the result + result.appendCommit(commit, file_changes); +} + + +/** Runs child process and allows to read the result. + * Multiple processes can be run for parallel processing. + */ auto gitShow(const std::string & hash) { std::string command = fmt::format( @@ -979,9 +1076,11 @@ auto gitShow(const std::string & hash) } +/** Obtain the list of commits and process them. + */ void processLog(const Options & options) { - Result result; + ResultWriter result; std::string command = "git log --reverse --no-merges --pretty=%H"; fmt::print("{}\n", command); @@ -1019,7 +1118,7 @@ void processLog(const Options & options) for (size_t i = 0; i < num_commits; ++i) { - processCommit(show_commands[i % num_threads], options, i, num_commits, hashes[i], snapshot, diff_hashes, result); + processCommit(show_commands[i % num_threads]->out, options, i, num_commits, hashes[i], snapshot, diff_hashes, result); if (!options.stop_after_commit.empty() && hashes[i] == options.stop_after_commit) break; From 94d49e4197b443a6bced0ac0d137ad646c1c1946 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Mon, 7 Sep 2020 03:18:15 +0300 Subject: [PATCH 021/121] Minor modifications --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index 6e43853d6ba..2add6813008 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -120,7 +120,9 @@ CREATE TABLE git.line_changes commit_hunks_changed UInt32 ) ENGINE = MergeTree ORDER BY time; -Insert the data with the following commands: +Run the tool. + +Then insert the data with the following commands: clickhouse-client --query "INSERT INTO git.commits FORMAT TSV" < commits.tsv clickhouse-client --query "INSERT INTO git.file_changes FORMAT TSV" < file_changes.tsv From 47ca6211604c6fcb7b2c4e137d739ebff88da975 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Mon, 7 Sep 2020 03:25:06 +0300 Subject: [PATCH 022/121] Minor modifications --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index 2add6813008..875da3ba0ac 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -993,7 +993,12 @@ void processDiffs( } if (size_limit && diff_size > *size_limit) + { + /// Drain to avoid "broken pipe" error in child process. + while (!in.eof()) + in.ignore(in.available()); return; + } } } From 6e0afbecf4fd0ccd04e9dbb82bff6a507545e8d1 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Mon, 7 Sep 2020 04:02:35 +0300 Subject: [PATCH 023/121] Minor modifications --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index 875da3ba0ac..b5488b0d69a 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -138,6 +138,7 @@ namespace DB namespace ErrorCodes { extern const int INCORRECT_DATA; + extern const int CHILD_WAS_NOT_EXITED_NORMALLY; } @@ -994,9 +995,6 @@ void processDiffs( if (size_limit && diff_size > *size_limit) { - /// Drain to avoid "broken pipe" error in child process. - while (!in.eof()) - in.ignore(in.available()); return; } } @@ -1127,6 +1125,19 @@ void processLog(const Options & options) { processCommit(show_commands[i % num_threads]->out, options, i, num_commits, hashes[i], snapshot, diff_hashes, result); + try + { + show_commands[i % num_threads]->wait(); + } + catch (const Exception & e) + { + /// For broken pipe when we stopped reading prematurally. + if (e.code() == ErrorCodes::CHILD_WAS_NOT_EXITED_NORMALLY) + std::cerr << getCurrentExceptionMessage(false) << "\n"; + else + throw; + } + if (!options.stop_after_commit.empty() && hashes[i] == options.stop_after_commit) break; From 69ce9e1f7020df985d7ea6ee450bf0d4b3438a0d Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Mon, 7 Sep 2020 05:36:54 +0300 Subject: [PATCH 024/121] More documentation --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 45 +++++++++++++++++++ 1 file changed, 45 insertions(+) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index b5488b0d69a..d3b6f77d3d7 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -27,6 +27,51 @@ static constexpr auto documentation = R"( +A tool to extract information from Git repository for analytics. + +It dumps the data for the following tables: +- commits - commits with statistics; +- file_changes - files changed in every commit with the info about the change and statistics; +- line_changes - every changed line in every changed file in every commit with full info about the line and the information about previous change of this line. + +The largest and the most important table is "line_changes". + +Allows to answer questions like: +- list files with maximum number of authors; +- show me the oldest lines of code in the repository; +- show me the files with longest history; +- list favorite files for author; +- list largest files with lowest number of authors; +- at what weekday the code has highest chance to stay in repository; +- the distribution of code age across repository; +- files sorted by average code age; +- quickly show file with blame info (rough); +- commits and lines of code distribution by time; by weekday, by author; for specific subdirectories; +- show history for every subdirectory, file, line of file, the number of changes (lines and commits) across time; how the number of contributors was changed across time; +- list files with most modifications; +- list files that were rewritten most number of time or by most of authors; +- what is percentage of code removal by other authors, across authors; +- the matrix of authors that shows what authors tends to rewrite another authors code; +- what is the worst time to write code in sense that the code has highest chance to be rewritten; +- the average time before code will be rewritten and the median (half-life of code decay); +- comments/code percentage change in time / by author / by location; +- who tend to write more tests / cpp code / comments. + +The data is intended for analytical purposes. It can be imprecise by many reasons but it should be good enough for its purpose. + +The data is not intended to provide any conclusions for managers, it is especially counter-indicative for any kinds of "performance review". Instead you can spend multiple days looking at various interesting statistics. + +Run this tool inside your git repository. It will create .tsv files that can be loaded into ClickHouse (or into other DBMS if you dare). + +The tool can process large enough repositories in a reasonable time. +It has been tested on: +- ClickHouse: 31 seconds; 3 million rows; +- LLVM: 8 minues; 62 million rows; +- Linux - 12 minutes; 85 million rows; +- Chromium - 67 minutes; 343 million rows; +(the numbers as of Sep 2020) + + Prepare the database by executing the following queries: DROP DATABASE IF EXISTS git; From 1dc48f66710c5a93e5376320ea7cf3c4a18046d5 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Mon, 7 Sep 2020 05:39:08 +0300 Subject: [PATCH 025/121] Better help --- utils/git-to-clickhouse/git-to-clickhouse.cpp | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index d3b6f77d3d7..6ef82ac3b6b 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -13,6 +13,7 @@ #include +#include #include #include #include @@ -1199,7 +1200,7 @@ try { using namespace DB; - po::options_description desc("Allowed options"); + po::options_description desc("Allowed options", getTerminalWidth()); desc.add_options() ("help,h", "produce help message") ("skip-commits-without-parents", po::value()->default_value(true), From 1400bdbf83c9ebf6e63eeda73966b7e7c0210d80 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Mon, 7 Sep 2020 06:11:35 +0300 Subject: [PATCH 026/121] Fix unit tests --- src/Common/ShellCommand.cpp | 23 +++++++++++++++---- utils/git-to-clickhouse/git-to-clickhouse.cpp | 13 ----------- 2 files changed, 18 insertions(+), 18 deletions(-) diff --git a/src/Common/ShellCommand.cpp b/src/Common/ShellCommand.cpp index 127f95fef06..bbb8801f190 100644 --- a/src/Common/ShellCommand.cpp +++ b/src/Common/ShellCommand.cpp @@ -57,7 +57,16 @@ ShellCommand::~ShellCommand() LOG_WARNING(getLogger(), "Cannot kill shell command pid {} errno '{}'", pid, errnoToString(retcode)); } else if (!wait_called) - tryWait(); + { + try + { + tryWait(); + } + catch (...) + { + tryLogCurrentException(getLogger()); + } + } } void ShellCommand::logCommand(const char * filename, char * const argv[]) @@ -74,7 +83,8 @@ void ShellCommand::logCommand(const char * filename, char * const argv[]) LOG_TRACE(ShellCommand::getLogger(), "Will start shell command '{}' with arguments {}", filename, args.str()); } -std::unique_ptr ShellCommand::executeImpl(const char * filename, char * const argv[], bool pipe_stdin_only, bool terminate_in_destructor) +std::unique_ptr ShellCommand::executeImpl( + const char * filename, char * const argv[], bool pipe_stdin_only, bool terminate_in_destructor) { logCommand(filename, argv); @@ -130,7 +140,8 @@ std::unique_ptr ShellCommand::executeImpl(const char * filename, c _exit(int(ReturnCodes::CANNOT_EXEC)); } - std::unique_ptr res(new ShellCommand(pid, pipe_stdin.fds_rw[1], pipe_stdout.fds_rw[0], pipe_stderr.fds_rw[0], terminate_in_destructor)); + std::unique_ptr res(new ShellCommand( + pid, pipe_stdin.fds_rw[1], pipe_stdout.fds_rw[0], pipe_stderr.fds_rw[0], terminate_in_destructor)); LOG_TRACE(getLogger(), "Started shell command '{}' with pid {}", filename, pid); @@ -143,7 +154,8 @@ std::unique_ptr ShellCommand::executeImpl(const char * filename, c } -std::unique_ptr ShellCommand::execute(const std::string & command, bool pipe_stdin_only, bool terminate_in_destructor) +std::unique_ptr ShellCommand::execute( + const std::string & command, bool pipe_stdin_only, bool terminate_in_destructor) { /// Arguments in non-constant chunks of memory (as required for `execv`). /// Moreover, their copying must be done before calling `vfork`, so after `vfork` do a minimum of things. @@ -157,7 +169,8 @@ std::unique_ptr ShellCommand::execute(const std::string & command, } -std::unique_ptr ShellCommand::executeDirect(const std::string & path, const std::vector & arguments, bool terminate_in_destructor) +std::unique_ptr ShellCommand::executeDirect( + const std::string & path, const std::vector & arguments, bool terminate_in_destructor) { size_t argv_sum_size = path.size() + 1; for (const auto & arg : arguments) diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/utils/git-to-clickhouse/git-to-clickhouse.cpp index 6ef82ac3b6b..a081efa3f47 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/utils/git-to-clickhouse/git-to-clickhouse.cpp @@ -1171,19 +1171,6 @@ void processLog(const Options & options) { processCommit(show_commands[i % num_threads]->out, options, i, num_commits, hashes[i], snapshot, diff_hashes, result); - try - { - show_commands[i % num_threads]->wait(); - } - catch (const Exception & e) - { - /// For broken pipe when we stopped reading prematurally. - if (e.code() == ErrorCodes::CHILD_WAS_NOT_EXITED_NORMALLY) - std::cerr << getCurrentExceptionMessage(false) << "\n"; - else - throw; - } - if (!options.stop_after_commit.empty() && hashes[i] == options.stop_after_commit) break; From d18e7adbc03e4e7d7ee268e8f90a14e73be7b021 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Mon, 7 Sep 2020 06:22:47 +0300 Subject: [PATCH 027/121] Add git-import as a tool --- programs/CMakeLists.txt | 18 ++++++++++++++---- programs/config_tools.h.in | 1 + programs/git-import/CMakeLists.txt | 10 ++++++++++ programs/git-import/clickhouse-git-import.cpp | 2 ++ .../git-import/git-import.cpp | 4 ++-- programs/install/Install.cpp | 1 + programs/main.cpp | 6 ++++++ utils/CMakeLists.txt | 1 - utils/git-to-clickhouse/CMakeLists.txt | 2 -- 9 files changed, 36 insertions(+), 9 deletions(-) create mode 100644 programs/git-import/CMakeLists.txt create mode 100644 programs/git-import/clickhouse-git-import.cpp rename utils/git-to-clickhouse/git-to-clickhouse.cpp => programs/git-import/git-import.cpp (99%) delete mode 100644 utils/git-to-clickhouse/CMakeLists.txt diff --git a/programs/CMakeLists.txt b/programs/CMakeLists.txt index 89220251cda..ae4a72ef62a 100644 --- a/programs/CMakeLists.txt +++ b/programs/CMakeLists.txt @@ -16,6 +16,7 @@ option (ENABLE_CLICKHOUSE_COMPRESSOR "Enable clickhouse-compressor" ${ENABLE_CLI option (ENABLE_CLICKHOUSE_COPIER "Enable clickhouse-copier" ${ENABLE_CLICKHOUSE_ALL}) option (ENABLE_CLICKHOUSE_FORMAT "Enable clickhouse-format" ${ENABLE_CLICKHOUSE_ALL}) option (ENABLE_CLICKHOUSE_OBFUSCATOR "Enable clickhouse-obfuscator" ${ENABLE_CLICKHOUSE_ALL}) +option (ENABLE_CLICKHOUSE_GIT_IMPORT "Enable clickhouse-git-import" ${ENABLE_CLICKHOUSE_ALL}) option (ENABLE_CLICKHOUSE_ODBC_BRIDGE "Enable clickhouse-odbc-bridge" ${ENABLE_CLICKHOUSE_ALL}) if (CLICKHOUSE_SPLIT_BINARY) @@ -91,21 +92,22 @@ add_subdirectory (copier) add_subdirectory (format) add_subdirectory (obfuscator) add_subdirectory (install) +add_subdirectory (git-import) if (ENABLE_CLICKHOUSE_ODBC_BRIDGE) add_subdirectory (odbc-bridge) endif () if (CLICKHOUSE_ONE_SHARED) - add_library(clickhouse-lib SHARED ${CLICKHOUSE_SERVER_SOURCES} ${CLICKHOUSE_CLIENT_SOURCES} ${CLICKHOUSE_LOCAL_SOURCES} ${CLICKHOUSE_BENCHMARK_SOURCES} ${CLICKHOUSE_COPIER_SOURCES} ${CLICKHOUSE_EXTRACT_FROM_CONFIG_SOURCES} ${CLICKHOUSE_COMPRESSOR_SOURCES} ${CLICKHOUSE_FORMAT_SOURCES} ${CLICKHOUSE_OBFUSCATOR_SOURCES} ${CLICKHOUSE_ODBC_BRIDGE_SOURCES}) - target_link_libraries(clickhouse-lib ${CLICKHOUSE_SERVER_LINK} ${CLICKHOUSE_CLIENT_LINK} ${CLICKHOUSE_LOCAL_LINK} ${CLICKHOUSE_BENCHMARK_LINK} ${CLICKHOUSE_COPIER_LINK} ${CLICKHOUSE_EXTRACT_FROM_CONFIG_LINK} ${CLICKHOUSE_COMPRESSOR_LINK} ${CLICKHOUSE_FORMAT_LINK} ${CLICKHOUSE_OBFUSCATOR_LINK} ${CLICKHOUSE_ODBC_BRIDGE_LINK}) - target_include_directories(clickhouse-lib ${CLICKHOUSE_SERVER_INCLUDE} ${CLICKHOUSE_CLIENT_INCLUDE} ${CLICKHOUSE_LOCAL_INCLUDE} ${CLICKHOUSE_BENCHMARK_INCLUDE} ${CLICKHOUSE_COPIER_INCLUDE} ${CLICKHOUSE_EXTRACT_FROM_CONFIG_INCLUDE} ${CLICKHOUSE_COMPRESSOR_INCLUDE} ${CLICKHOUSE_FORMAT_INCLUDE} ${CLICKHOUSE_OBFUSCATOR_INCLUDE} ${CLICKHOUSE_ODBC_BRIDGE_INCLUDE}) + add_library(clickhouse-lib SHARED ${CLICKHOUSE_SERVER_SOURCES} ${CLICKHOUSE_CLIENT_SOURCES} ${CLICKHOUSE_LOCAL_SOURCES} ${CLICKHOUSE_BENCHMARK_SOURCES} ${CLICKHOUSE_COPIER_SOURCES} ${CLICKHOUSE_EXTRACT_FROM_CONFIG_SOURCES} ${CLICKHOUSE_COMPRESSOR_SOURCES} ${CLICKHOUSE_FORMAT_SOURCES} ${CLICKHOUSE_OBFUSCATOR_SOURCES} ${CLICKHOUSE_GIT_IMPORT_SOURCES} ${CLICKHOUSE_ODBC_BRIDGE_SOURCES}) + target_link_libraries(clickhouse-lib ${CLICKHOUSE_SERVER_LINK} ${CLICKHOUSE_CLIENT_LINK} ${CLICKHOUSE_LOCAL_LINK} ${CLICKHOUSE_BENCHMARK_LINK} ${CLICKHOUSE_COPIER_LINK} ${CLICKHOUSE_EXTRACT_FROM_CONFIG_LINK} ${CLICKHOUSE_COMPRESSOR_LINK} ${CLICKHOUSE_FORMAT_LINK} ${CLICKHOUSE_OBFUSCATOR_LINK} ${CLICKHOUSE_GIT_IMPORT_LINK} ${CLICKHOUSE_ODBC_BRIDGE_LINK}) + target_include_directories(clickhouse-lib ${CLICKHOUSE_SERVER_INCLUDE} ${CLICKHOUSE_CLIENT_INCLUDE} ${CLICKHOUSE_LOCAL_INCLUDE} ${CLICKHOUSE_BENCHMARK_INCLUDE} ${CLICKHOUSE_COPIER_INCLUDE} ${CLICKHOUSE_EXTRACT_FROM_CONFIG_INCLUDE} ${CLICKHOUSE_COMPRESSOR_INCLUDE} ${CLICKHOUSE_FORMAT_INCLUDE} ${CLICKHOUSE_OBFUSCATOR_INCLUDE} ${CLICKHOUSE_GIT_IMPORT_INCLUDE} ${CLICKHOUSE_ODBC_BRIDGE_INCLUDE}) set_target_properties(clickhouse-lib PROPERTIES SOVERSION ${VERSION_MAJOR}.${VERSION_MINOR} VERSION ${VERSION_SO} OUTPUT_NAME clickhouse DEBUG_POSTFIX "") install (TARGETS clickhouse-lib LIBRARY DESTINATION ${CMAKE_INSTALL_LIBDIR} COMPONENT clickhouse) endif() if (CLICKHOUSE_SPLIT_BINARY) - set (CLICKHOUSE_ALL_TARGETS clickhouse-server clickhouse-client clickhouse-local clickhouse-benchmark clickhouse-extract-from-config clickhouse-compressor clickhouse-format clickhouse-obfuscator clickhouse-copier) + set (CLICKHOUSE_ALL_TARGETS clickhouse-server clickhouse-client clickhouse-local clickhouse-benchmark clickhouse-extract-from-config clickhouse-compressor clickhouse-format clickhouse-obfuscator clickhouse-git-import clickhouse-copier) if (ENABLE_CLICKHOUSE_ODBC_BRIDGE) list (APPEND CLICKHOUSE_ALL_TARGETS clickhouse-odbc-bridge) @@ -149,6 +151,9 @@ else () if (ENABLE_CLICKHOUSE_OBFUSCATOR) clickhouse_target_link_split_lib(clickhouse obfuscator) endif () + if (ENABLE_CLICKHOUSE_GIT_IMPORT) + clickhouse_target_link_split_lib(clickhouse git-import) + endif () if (ENABLE_CLICKHOUSE_INSTALL) clickhouse_target_link_split_lib(clickhouse install) endif () @@ -199,6 +204,11 @@ else () install (FILES ${CMAKE_CURRENT_BINARY_DIR}/clickhouse-obfuscator DESTINATION ${CMAKE_INSTALL_BINDIR} COMPONENT clickhouse) list(APPEND CLICKHOUSE_BUNDLE clickhouse-obfuscator) endif () + if (ENABLE_CLICKHOUSE_GIT_IMPORT) + add_custom_target (clickhouse-git-import ALL COMMAND ${CMAKE_COMMAND} -E create_symlink clickhouse clickhouse-git-import DEPENDS clickhouse) + install (FILES ${CMAKE_CURRENT_BINARY_DIR}/clickhouse-git-import DESTINATION ${CMAKE_INSTALL_BINDIR} COMPONENT clickhouse) + list(APPEND CLICKHOUSE_BUNDLE clickhouse-git-import) + endif () if(ENABLE_CLICKHOUSE_ODBC_BRIDGE) list(APPEND CLICKHOUSE_BUNDLE clickhouse-odbc-bridge) endif() diff --git a/programs/config_tools.h.in b/programs/config_tools.h.in index 11386aca60e..7cb5a6d883a 100644 --- a/programs/config_tools.h.in +++ b/programs/config_tools.h.in @@ -12,5 +12,6 @@ #cmakedefine01 ENABLE_CLICKHOUSE_COMPRESSOR #cmakedefine01 ENABLE_CLICKHOUSE_FORMAT #cmakedefine01 ENABLE_CLICKHOUSE_OBFUSCATOR +#cmakedefine01 ENABLE_CLICKHOUSE_GIT_IMPORT #cmakedefine01 ENABLE_CLICKHOUSE_INSTALL #cmakedefine01 ENABLE_CLICKHOUSE_ODBC_BRIDGE diff --git a/programs/git-import/CMakeLists.txt b/programs/git-import/CMakeLists.txt new file mode 100644 index 00000000000..279bb35a272 --- /dev/null +++ b/programs/git-import/CMakeLists.txt @@ -0,0 +1,10 @@ +set (CLICKHOUSE_GIT_IMPORT_SOURCES git-import.cpp) + +set (CLICKHOUSE_GIT_IMPORT_LINK + PRIVATE + boost::program_options + dbms +) + +clickhouse_program_add(git-import) + diff --git a/programs/git-import/clickhouse-git-import.cpp b/programs/git-import/clickhouse-git-import.cpp new file mode 100644 index 00000000000..cfa06306604 --- /dev/null +++ b/programs/git-import/clickhouse-git-import.cpp @@ -0,0 +1,2 @@ +int mainEntryClickHouseGitImport(int argc, char ** argv); +int main(int argc_, char ** argv_) { return mainEntryClickHouseGitImport(argc_, argv_); } diff --git a/utils/git-to-clickhouse/git-to-clickhouse.cpp b/programs/git-import/git-import.cpp similarity index 99% rename from utils/git-to-clickhouse/git-to-clickhouse.cpp rename to programs/git-import/git-import.cpp index a081efa3f47..f1ed4d28c6e 100644 --- a/utils/git-to-clickhouse/git-to-clickhouse.cpp +++ b/programs/git-import/git-import.cpp @@ -1182,7 +1182,7 @@ void processLog(const Options & options) } -int main(int argc, char ** argv) +int mainEntryClickHouseGitImport(int argc, char ** argv) try { using namespace DB; @@ -1219,7 +1219,7 @@ try << "Usage: " << argv[0] << '\n' << desc << '\n' << "\nExample:\n" - << "\n./git-to-clickhouse --skip-paths 'generated\\.cpp|^(contrib|docs?|website|libs/(libcityhash|liblz4|libdivide|libvectorclass|libdouble-conversion|libcpuid|libzstd|libfarmhash|libmetrohash|libpoco|libwidechar_width))/' --skip-commits-with-messages '^Merge branch '\n"; + << "\nclickhouse git-import --skip-paths 'generated\\.cpp|^(contrib|docs?|website|libs/(libcityhash|liblz4|libdivide|libvectorclass|libdouble-conversion|libcpuid|libzstd|libfarmhash|libmetrohash|libpoco|libwidechar_width))/' --skip-commits-with-messages '^Merge branch '\n"; return 1; } diff --git a/programs/install/Install.cpp b/programs/install/Install.cpp index 7b7ab149447..bd60fbb63ba 100644 --- a/programs/install/Install.cpp +++ b/programs/install/Install.cpp @@ -205,6 +205,7 @@ int mainEntryClickHouseInstall(int argc, char ** argv) "clickhouse-benchmark", "clickhouse-copier", "clickhouse-obfuscator", + "clickhouse-git-import", "clickhouse-compressor", "clickhouse-format", "clickhouse-extract-from-config" diff --git a/programs/main.cpp b/programs/main.cpp index 3df5f9f683b..b91bd732f21 100644 --- a/programs/main.cpp +++ b/programs/main.cpp @@ -46,6 +46,9 @@ int mainEntryClickHouseClusterCopier(int argc, char ** argv); #if ENABLE_CLICKHOUSE_OBFUSCATOR int mainEntryClickHouseObfuscator(int argc, char ** argv); #endif +#if ENABLE_CLICKHOUSE_GIT_IMPORT +int mainEntryClickHouseGitImport(int argc, char ** argv); +#endif #if ENABLE_CLICKHOUSE_INSTALL int mainEntryClickHouseInstall(int argc, char ** argv); int mainEntryClickHouseStart(int argc, char ** argv); @@ -91,6 +94,9 @@ std::pair clickhouse_applications[] = #if ENABLE_CLICKHOUSE_OBFUSCATOR {"obfuscator", mainEntryClickHouseObfuscator}, #endif +#if ENABLE_CLICKHOUSE_GIT_IMPORT + {"git-import", mainEntryClickHouseGitImport}, +#endif #if ENABLE_CLICKHOUSE_INSTALL {"install", mainEntryClickHouseInstall}, {"start", mainEntryClickHouseStart}, diff --git a/utils/CMakeLists.txt b/utils/CMakeLists.txt index 93490fba565..b4408a298c3 100644 --- a/utils/CMakeLists.txt +++ b/utils/CMakeLists.txt @@ -30,7 +30,6 @@ if (NOT DEFINED ENABLE_UTILS OR ENABLE_UTILS) add_subdirectory (checksum-for-compressed-block) add_subdirectory (db-generator) add_subdirectory (wal-dump) - add_subdirectory (git-to-clickhouse) endif () if (ENABLE_CODE_QUALITY) diff --git a/utils/git-to-clickhouse/CMakeLists.txt b/utils/git-to-clickhouse/CMakeLists.txt deleted file mode 100644 index 0e46b68d471..00000000000 --- a/utils/git-to-clickhouse/CMakeLists.txt +++ /dev/null @@ -1,2 +0,0 @@ -add_executable (git-to-clickhouse git-to-clickhouse.cpp) -target_link_libraries(git-to-clickhouse PRIVATE dbms boost::program_options) From ee54971c3d26ca1219da4909bd30f44bee77fd97 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Mon, 7 Sep 2020 07:11:03 +0300 Subject: [PATCH 028/121] Fix build --- programs/git-import/git-import.cpp | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/programs/git-import/git-import.cpp b/programs/git-import/git-import.cpp index f1ed4d28c6e..d314969a1a8 100644 --- a/programs/git-import/git-import.cpp +++ b/programs/git-import/git-import.cpp @@ -184,7 +184,6 @@ namespace DB namespace ErrorCodes { extern const int INCORRECT_DATA; - extern const int CHILD_WAS_NOT_EXITED_NORMALLY; } @@ -419,7 +418,7 @@ using LineChanges = std::vector; struct FileDiff { - FileDiff(FileChange file_change_) : file_change(file_change_) {} + explicit FileDiff(FileChange file_change_) : file_change(file_change_) {} FileChange file_change; LineChanges line_changes; @@ -546,7 +545,7 @@ struct Options std::optional diff_size_limit; std::string stop_after_commit; - Options(const po::variables_map & options) + explicit Options(const po::variables_map & options) { skip_commits_without_parents = options["skip-commits-without-parents"].as(); skip_commits_with_duplicate_diffs = options["skip-commits-with-duplicate-diffs"].as(); @@ -753,7 +752,7 @@ UInt128 diffHash(const CommitDiff & file_changes) { SipHash hasher; - for (auto & elem : file_changes) + for (const auto & elem : file_changes) { hasher.update(elem.second.file_change.change_type); hasher.update(elem.second.file_change.old_path.size()); @@ -762,7 +761,7 @@ UInt128 diffHash(const CommitDiff & file_changes) hasher.update(elem.second.file_change.path); hasher.update(elem.second.line_changes.size()); - for (auto & line_change : elem.second.line_changes) + for (const auto & line_change : elem.second.line_changes) { hasher.update(line_change.sign); hasher.update(line_change.line_number_old); @@ -1159,6 +1158,8 @@ void processLog(const Options & options) /// Will run multiple processes in parallel size_t num_threads = options.threads; + if (num_threads == 0) + throw Exception("num-threads cannot be zero", ErrorCodes::INCORRECT_DATA); std::vector> show_commands(num_threads); for (size_t i = 0; i < num_commits && i < num_threads; ++i) @@ -1223,7 +1224,7 @@ try return 1; } - processLog(options); + processLog(Options(options)); return 0; } catch (...) From 3942cc615f03ecb8e5b9e7437fdc5c57613c245d Mon Sep 17 00:00:00 2001 From: alexey-milovidov Date: Mon, 7 Sep 2020 10:09:42 +0300 Subject: [PATCH 029/121] Update git-import.cpp --- programs/git-import/git-import.cpp | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/programs/git-import/git-import.cpp b/programs/git-import/git-import.cpp index d314969a1a8..45bc47348e7 100644 --- a/programs/git-import/git-import.cpp +++ b/programs/git-import/git-import.cpp @@ -9,7 +9,7 @@ #include #include -#include +#include #include @@ -539,8 +539,8 @@ struct Options bool skip_commits_without_parents = true; bool skip_commits_with_duplicate_diffs = true; size_t threads = 1; - std::optional skip_paths; - std::optional skip_commits_with_messages; + std::optional skip_paths; + std::optional skip_commits_with_messages; std::unordered_set skip_commits; std::optional diff_size_limit; std::string stop_after_commit; @@ -857,7 +857,7 @@ void processFileChanges( assertChar('\n', in); - if (!(options.skip_paths && re2_st::RE2::PartialMatch(file_change.path, *options.skip_paths))) + if (!(options.skip_paths && re2::RE2::PartialMatch(file_change.path, *options.skip_paths))) { file_changes.emplace( file_change.path, @@ -1070,7 +1070,7 @@ void processCommit( readNullTerminated(parent_hash, in); readNullTerminated(commit.message, in); - if (options.skip_commits_with_messages && re2_st::RE2::PartialMatch(commit.message, *options.skip_commits_with_messages)) + if (options.skip_commits_with_messages && re2::RE2::PartialMatch(commit.message, *options.skip_commits_with_messages)) return; std::string message_to_print = commit.message; From bee629c971d8f5add8fe4f205aa30f8f4e66375f Mon Sep 17 00:00:00 2001 From: Vitaly Baranov Date: Tue, 8 Sep 2020 02:08:42 +0300 Subject: [PATCH 030/121] Use join() instead of detach() for the lists_writing_thread in DiskAccessStorage. --- src/Access/DiskAccessStorage.cpp | 47 ++++++++++++++------------------ src/Access/DiskAccessStorage.h | 5 ++-- 2 files changed, 22 insertions(+), 30 deletions(-) diff --git a/src/Access/DiskAccessStorage.cpp b/src/Access/DiskAccessStorage.cpp index fc80859885d..6162e4aacc2 100644 --- a/src/Access/DiskAccessStorage.cpp +++ b/src/Access/DiskAccessStorage.cpp @@ -426,33 +426,41 @@ bool DiskAccessStorage::writeLists() void DiskAccessStorage::scheduleWriteLists(EntityType type) { if (failed_to_write_lists) - return; + return; /// We don't try to write list files after the first fail. + /// The next restart of the server will invoke rebuilding of the list files. - bool already_scheduled = !types_of_lists_to_write.empty(); types_of_lists_to_write.insert(type); - if (already_scheduled) - return; + if (lists_writing_thread_is_waiting) + return; /// If the lists' writing thread is still waiting we can update `types_of_lists_to_write` easily, + /// without restarting that thread. + + if (lists_writing_thread.joinable()) + lists_writing_thread.join(); /// Create the 'need_rebuild_lists.mark' file. /// This file will be used later to find out if writing lists is successful or not. std::ofstream{getNeedRebuildListsMarkFilePath(directory_path)}; - startListsWritingThread(); + lists_writing_thread = ThreadFromGlobalPool{&DiskAccessStorage::listsWritingThreadFunc, this}; + lists_writing_thread_is_waiting = true; } -void DiskAccessStorage::startListsWritingThread() +void DiskAccessStorage::listsWritingThreadFunc() { - if (lists_writing_thread.joinable()) + std::unique_lock lock{mutex}; + { - if (!lists_writing_thread_exited) - return; - lists_writing_thread.detach(); + /// It's better not to write the lists files too often, that's why we need + /// the following timeout. + const auto timeout = std::chrono::minutes(1); + SCOPE_EXIT({ lists_writing_thread_is_waiting = false; }); + if (lists_writing_thread_should_exit.wait_for(lock, timeout) != std::cv_status::timeout) + return; /// The destructor requires us to exit. } - lists_writing_thread_exited = false; - lists_writing_thread = ThreadFromGlobalPool{&DiskAccessStorage::listsWritingThreadFunc, this}; + writeLists(); } @@ -466,21 +474,6 @@ void DiskAccessStorage::stopListsWritingThread() } -void DiskAccessStorage::listsWritingThreadFunc() -{ - std::unique_lock lock{mutex}; - SCOPE_EXIT({ lists_writing_thread_exited = true; }); - - /// It's better not to write the lists files too often, that's why we need - /// the following timeout. - const auto timeout = std::chrono::minutes(1); - if (lists_writing_thread_should_exit.wait_for(lock, timeout) != std::cv_status::timeout) - return; /// The destructor requires us to exit. - - writeLists(); -} - - /// Reads and parses all the ".sql" files from a specified directory /// and then saves the files "users.list", "roles.list", etc. to the same directory. bool DiskAccessStorage::rebuildLists() diff --git a/src/Access/DiskAccessStorage.h b/src/Access/DiskAccessStorage.h index 11eb1c3b1ad..ed2dc8b1242 100644 --- a/src/Access/DiskAccessStorage.h +++ b/src/Access/DiskAccessStorage.h @@ -42,9 +42,8 @@ private: void scheduleWriteLists(EntityType type); bool rebuildLists(); - void startListsWritingThread(); - void stopListsWritingThread(); void listsWritingThreadFunc(); + void stopListsWritingThread(); void insertNoLock(const UUID & id, const AccessEntityPtr & new_entity, bool replace_if_exists, Notifications & notifications); void removeNoLock(const UUID & id, Notifications & notifications); @@ -74,7 +73,7 @@ private: bool failed_to_write_lists = false; /// Whether writing of the list files has been failed since the recent restart of the server. ThreadFromGlobalPool lists_writing_thread; /// List files are written in a separate thread. std::condition_variable lists_writing_thread_should_exit; /// Signals `lists_writing_thread` to exit. - std::atomic lists_writing_thread_exited = false; + bool lists_writing_thread_is_waiting = false; mutable std::list handlers_by_type[static_cast(EntityType::MAX)]; mutable std::mutex mutex; }; From cce970e40cdf1eba81a1d34c6e692ec883d544e2 Mon Sep 17 00:00:00 2001 From: Vitaly Baranov Date: Tue, 8 Sep 2020 02:09:03 +0300 Subject: [PATCH 031/121] Use join() instead of detach() for loading threads in ExternalLoader. --- src/Interpreters/ExternalLoader.cpp | 29 +++++++++++++++++++++++------ 1 file changed, 23 insertions(+), 6 deletions(-) diff --git a/src/Interpreters/ExternalLoader.cpp b/src/Interpreters/ExternalLoader.cpp index e8df205760a..dcef36de175 100644 --- a/src/Interpreters/ExternalLoader.cpp +++ b/src/Interpreters/ExternalLoader.cpp @@ -893,6 +893,8 @@ private: cancelLoading(info); } + putBackFinishedThreadsToPool(); + /// All loadings have unique loading IDs. size_t loading_id = next_id_counter++; info.loading_id = loading_id; @@ -914,6 +916,21 @@ private: } } + void putBackFinishedThreadsToPool() + { + for (auto loading_id : recently_finished_loadings) + { + auto it = loading_threads.find(loading_id); + if (it != loading_threads.end()) + { + auto thread = std::move(it->second); + loading_threads.erase(it); + thread.join(); /// It's very likely that `thread` has already finished. + } + } + recently_finished_loadings.clear(); + } + static void cancelLoading(Info & info) { if (!info.isLoading()) @@ -1095,12 +1112,11 @@ private: } min_id_to_finish_loading_dependencies.erase(std::this_thread::get_id()); - auto it = loading_threads.find(loading_id); - if (it != loading_threads.end()) - { - it->second.detach(); - loading_threads.erase(it); - } + /// Add `loading_id` to the list of recently finished loadings. + /// This list is used to later put the threads which finished loading back to the thread pool. + /// (We can't put the loading thread back to the thread pool immediately here because at this point + /// the loading thread is about to finish but it's not finished yet right now.) + recently_finished_loadings.push_back(loading_id); } /// Calculate next update time for loaded_object. Can be called without mutex locking, @@ -1158,6 +1174,7 @@ private: bool always_load_everything = false; std::atomic enable_async_loading = false; std::unordered_map loading_threads; + std::vector recently_finished_loadings; std::unordered_map min_id_to_finish_loading_dependencies; size_t next_id_counter = 1; /// should always be > 0 mutable pcg64 rnd_engine{randomSeed()}; From c34eaf5de3380e8b12f0f6e8b578bb13744660bf Mon Sep 17 00:00:00 2001 From: alesapin Date: Wed, 9 Sep 2020 10:08:38 +0300 Subject: [PATCH 032/121] Update ci_config and llvm --- contrib/llvm | 2 +- tests/ci/ci_config.json | 26 +++++++++++++------------- 2 files changed, 14 insertions(+), 14 deletions(-) diff --git a/contrib/llvm b/contrib/llvm index 3d6c7e91676..8f24d507c1c 160000 --- a/contrib/llvm +++ b/contrib/llvm @@ -1 +1 @@ -Subproject commit 3d6c7e916760b395908f28a1c885c8334d4fa98b +Subproject commit 8f24d507c1cfeec66d27f48fe74518fd278e2d25 diff --git a/tests/ci/ci_config.json b/tests/ci/ci_config.json index 44e9df49216..adb736a8df3 100644 --- a/tests/ci/ci_config.json +++ b/tests/ci/ci_config.json @@ -1,7 +1,7 @@ { "build_config": [ { - "compiler": "gcc-9", + "compiler": "gcc-10", "build-type": "", "sanitizer": "", "package-type": "deb", @@ -12,7 +12,7 @@ "with_coverage": false }, { - "compiler": "gcc-9", + "compiler": "gcc-10", "build-type": "", "sanitizer": "", "package-type": "performance", @@ -22,7 +22,7 @@ "with_coverage": false }, { - "compiler": "gcc-9", + "compiler": "gcc-10", "build-type": "", "sanitizer": "", "package-type": "binary", @@ -92,7 +92,7 @@ "with_coverage": false }, { - "compiler": "gcc-9", + "compiler": "gcc-10", "build-type": "", "sanitizer": "", "package-type": "deb", @@ -227,7 +227,7 @@ }, "Functional stateful tests (release)": { "required_build_properties": { - "compiler": "gcc-9", + "compiler": "gcc-10", "package_type": "deb", "build_type": "relwithdebuginfo", "sanitizer": "none", @@ -239,7 +239,7 @@ }, "Functional stateful tests (release, DatabaseAtomic)": { "required_build_properties": { - "compiler": "gcc-9", + "compiler": "gcc-10", "package_type": "deb", "build_type": "relwithdebuginfo", "sanitizer": "none", @@ -311,7 +311,7 @@ }, "Functional stateless tests (release)": { "required_build_properties": { - "compiler": "gcc-9", + "compiler": "gcc-10", "package_type": "deb", "build_type": "relwithdebuginfo", "sanitizer": "none", @@ -323,7 +323,7 @@ }, "Functional stateless tests (unbundled)": { "required_build_properties": { - "compiler": "gcc-9", + "compiler": "gcc-10", "package_type": "deb", "build_type": "relwithdebuginfo", "sanitizer": "none", @@ -335,7 +335,7 @@ }, "Functional stateless tests (release, polymorphic parts enabled)": { "required_build_properties": { - "compiler": "gcc-9", + "compiler": "gcc-10", "package_type": "deb", "build_type": "relwithdebuginfo", "sanitizer": "none", @@ -347,7 +347,7 @@ }, "Functional stateless tests (release, DatabaseAtomic)": { "required_build_properties": { - "compiler": "gcc-9", + "compiler": "gcc-10", "package_type": "deb", "build_type": "relwithdebuginfo", "sanitizer": "none", @@ -443,7 +443,7 @@ }, "Compatibility check": { "required_build_properties": { - "compiler": "gcc-9", + "compiler": "gcc-10", "package_type": "deb", "build_type": "relwithdebuginfo", "sanitizer": "none", @@ -467,7 +467,7 @@ }, "Testflows check": { "required_build_properties": { - "compiler": "gcc-9", + "compiler": "gcc-10", "package_type": "deb", "build_type": "relwithdebuginfo", "sanitizer": "none", @@ -479,7 +479,7 @@ }, "Unit tests release gcc": { "required_build_properties": { - "compiler": "gcc-9", + "compiler": "gcc-10", "package_type": "binary", "build_type": "relwithdebuginfo", "sanitizer": "none", From 4ba8f8960bd4e86a57dafba6a0aa1574b66d97db Mon Sep 17 00:00:00 2001 From: alesapin Date: Wed, 9 Sep 2020 12:53:24 +0300 Subject: [PATCH 033/121] Increase frame-larger-than --- cmake/warnings.cmake | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/cmake/warnings.cmake b/cmake/warnings.cmake index 2f78dc34079..aec3e46ffa6 100644 --- a/cmake/warnings.cmake +++ b/cmake/warnings.cmake @@ -23,7 +23,7 @@ option (WEVERYTHING "Enables -Weverything option with some exceptions. This is i # Control maximum size of stack frames. It can be important if the code is run in fibers with small stack size. # Only in release build because debug has too large stack frames. if ((NOT CMAKE_BUILD_TYPE_UC STREQUAL "DEBUG") AND (NOT SANITIZE)) - add_warning(frame-larger-than=16384) + add_warning(frame-larger-than=32768) endif () if (COMPILER_CLANG) From f528cd9f97b4f7c54a6c22406f09983d055ce642 Mon Sep 17 00:00:00 2001 From: alesapin Date: Wed, 9 Sep 2020 13:01:12 +0300 Subject: [PATCH 034/121] Forward compiler version to unbundled build --- docker/packager/packager | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docker/packager/packager b/docker/packager/packager index 5874bedd17a..909f20acd6d 100755 --- a/docker/packager/packager +++ b/docker/packager/packager @@ -93,7 +93,7 @@ def parse_env_variables(build_type, compiler, sanitizer, package_type, image_typ cxx = cc.replace('gcc', 'g++').replace('clang', 'clang++') - if image_type == "deb": + if image_type == "deb" or image_type == "unbundled": result.append("DEB_CC={}".format(cc)) result.append("DEB_CXX={}".format(cxx)) elif image_type == "binary": From ca6b634eb0466361da6f3526a6611ab0ccd8bfc1 Mon Sep 17 00:00:00 2001 From: alesapin Date: Wed, 9 Sep 2020 13:51:01 +0300 Subject: [PATCH 035/121] Install gcc-10 from proposed repo --- docker/packager/binary/Dockerfile | 13 +++++++++++-- docker/packager/deb/Dockerfile | 12 ++++++++++-- 2 files changed, 21 insertions(+), 4 deletions(-) diff --git a/docker/packager/binary/Dockerfile b/docker/packager/binary/Dockerfile index 45c35c2e0f3..b911b59a41d 100644 --- a/docker/packager/binary/Dockerfile +++ b/docker/packager/binary/Dockerfile @@ -32,8 +32,6 @@ RUN apt-get update \ curl \ gcc-9 \ g++-9 \ - gcc-10 \ - g++-10 \ llvm-${LLVM_VERSION} \ clang-${LLVM_VERSION} \ lld-${LLVM_VERSION} \ @@ -93,5 +91,16 @@ RUN wget -nv "https://developer.arm.com/-/media/Files/downloads/gnu-a/8.3-2019.0 # Download toolchain for FreeBSD 11.3 RUN wget -nv https://clickhouse-datasets.s3.yandex.net/toolchains/toolchains/freebsd-11.3-toolchain.tar.xz +# NOTE: For some reason we have outdated version of gcc-10 in ubuntu 20.04 stable. +# Current workaround is to use latest version proposed repo. Remove as soon as +# gcc-10.2 appear in stable repo. +RUN echo 'deb http://archive.ubuntu.com/ubuntu/ focal-proposed restricted main multiverse universe' > /etc/apt/sources.list.d/proposed-repositories.list + +RUN apt-get update \ + && apt-get install gcc-10 g++10 --yes + +RUN rm /etc/apt/sources.list.d/proposed-repositories.list + + COPY build.sh / CMD ["/bin/bash", "/build.sh"] diff --git a/docker/packager/deb/Dockerfile b/docker/packager/deb/Dockerfile index 87f4582f8e2..30334504c55 100644 --- a/docker/packager/deb/Dockerfile +++ b/docker/packager/deb/Dockerfile @@ -42,8 +42,6 @@ RUN export CODENAME="$(lsb_release --codename --short | tr 'A-Z' 'a-z')" \ # Libraries from OS are only needed to test the "unbundled" build (this is not used in production). RUN apt-get update \ && apt-get install \ - gcc-10 \ - g++-10 \ gcc-9 \ g++-9 \ clang-11 \ @@ -75,6 +73,16 @@ RUN apt-get update \ pigz \ --yes --no-install-recommends +# NOTE: For some reason we have outdated version of gcc-10 in ubuntu 20.04 stable. +# Current workaround is to use latest version proposed repo. Remove as soon as +# gcc-10.2 appear in stable repo. +RUN echo 'deb http://archive.ubuntu.com/ubuntu/ focal-proposed restricted main multiverse universe' > /etc/apt/sources.list.d/proposed-repositories.list + +RUN apt-get update \ + && apt-get install gcc-10 g++10 --yes --no-install-recommends + +RUN rm /etc/apt/sources.list.d/proposed-repositories.list + # This symlink required by gcc to find lld compiler RUN ln -s /usr/bin/lld-${LLVM_VERSION} /usr/bin/ld.lld From c535d752438c9616dab8fac79bf8594acb44665a Mon Sep 17 00:00:00 2001 From: alesapin Date: Wed, 9 Sep 2020 14:47:34 +0300 Subject: [PATCH 036/121] Add update --- docker/packager/binary/Dockerfile | 2 +- docker/packager/deb/Dockerfile | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/docker/packager/binary/Dockerfile b/docker/packager/binary/Dockerfile index b911b59a41d..893e9191b1e 100644 --- a/docker/packager/binary/Dockerfile +++ b/docker/packager/binary/Dockerfile @@ -99,7 +99,7 @@ RUN echo 'deb http://archive.ubuntu.com/ubuntu/ focal-proposed restricted main m RUN apt-get update \ && apt-get install gcc-10 g++10 --yes -RUN rm /etc/apt/sources.list.d/proposed-repositories.list +RUN rm /etc/apt/sources.list.d/proposed-repositories.list && apt-get update COPY build.sh / diff --git a/docker/packager/deb/Dockerfile b/docker/packager/deb/Dockerfile index 30334504c55..4b7c2ae53a4 100644 --- a/docker/packager/deb/Dockerfile +++ b/docker/packager/deb/Dockerfile @@ -81,7 +81,7 @@ RUN echo 'deb http://archive.ubuntu.com/ubuntu/ focal-proposed restricted main m RUN apt-get update \ && apt-get install gcc-10 g++10 --yes --no-install-recommends -RUN rm /etc/apt/sources.list.d/proposed-repositories.list +RUN rm /etc/apt/sources.list.d/proposed-repositories.list && apt-get update # This symlink required by gcc to find lld compiler RUN ln -s /usr/bin/lld-${LLVM_VERSION} /usr/bin/ld.lld From 956138635de536560d0843025720d7ce7b947cf3 Mon Sep 17 00:00:00 2001 From: alesapin Date: Wed, 9 Sep 2020 15:59:26 +0300 Subject: [PATCH 037/121] Fix compiler name --- docker/packager/binary/Dockerfile | 2 +- docker/packager/deb/Dockerfile | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/docker/packager/binary/Dockerfile b/docker/packager/binary/Dockerfile index 893e9191b1e..03bb3b5aefa 100644 --- a/docker/packager/binary/Dockerfile +++ b/docker/packager/binary/Dockerfile @@ -97,7 +97,7 @@ RUN wget -nv https://clickhouse-datasets.s3.yandex.net/toolchains/toolchains/fre RUN echo 'deb http://archive.ubuntu.com/ubuntu/ focal-proposed restricted main multiverse universe' > /etc/apt/sources.list.d/proposed-repositories.list RUN apt-get update \ - && apt-get install gcc-10 g++10 --yes + && apt-get install gcc-10 g++-10 --yes RUN rm /etc/apt/sources.list.d/proposed-repositories.list && apt-get update diff --git a/docker/packager/deb/Dockerfile b/docker/packager/deb/Dockerfile index 4b7c2ae53a4..a3c87f13fe4 100644 --- a/docker/packager/deb/Dockerfile +++ b/docker/packager/deb/Dockerfile @@ -79,7 +79,7 @@ RUN apt-get update \ RUN echo 'deb http://archive.ubuntu.com/ubuntu/ focal-proposed restricted main multiverse universe' > /etc/apt/sources.list.d/proposed-repositories.list RUN apt-get update \ - && apt-get install gcc-10 g++10 --yes --no-install-recommends + && apt-get install gcc-10 g++-10 --yes --no-install-recommends RUN rm /etc/apt/sources.list.d/proposed-repositories.list && apt-get update From b68782d285e5ea76f7318b55bf41cf337dfa71fc Mon Sep 17 00:00:00 2001 From: Alexander Tokmakov Date: Wed, 9 Sep 2020 16:32:50 +0300 Subject: [PATCH 038/121] enable more tests with Atomic database --- docker/test/stress/stress | 2 +- programs/client/Client.cpp | 28 +++++++- src/Interpreters/DatabaseCatalog.cpp | 5 +- src/Interpreters/InterpreterCreateQuery.cpp | 1 + .../MergeTree/MergeTreeWriteAheadLog.cpp | 1 + src/Storages/StorageReplicatedMergeTree.cpp | 16 +++-- src/Storages/System/StorageSystemTables.cpp | 6 ++ .../queries/0_stateless/00116_storage_set.sql | 2 +- .../00180_attach_materialized_view.sql | 2 +- ...per_deduplication_and_unexpected_parts.sql | 2 +- .../00281_compile_sizeof_packed.re | 0 .../0_stateless/00311_array_primary_key.sql | 2 +- .../00423_storage_log_single_thread.sql | 6 +- .../00816_long_concurrent_alter_column.sh | 27 +++++--- .../01190_full_attach_syntax.reference | 13 ++++ .../0_stateless/01190_full_attach_syntax.sql | 66 +++++++++++++++++++ .../01305_replica_create_drop_zookeeper.sh | 20 ++++-- .../00065_loyalty_with_storage_join.sql | 2 +- tests/queries/skip_list.json | 33 +--------- 19 files changed, 172 insertions(+), 62 deletions(-) delete mode 100644 tests/queries/0_stateless/00281_compile_sizeof_packed.re create mode 100644 tests/queries/0_stateless/01190_full_attach_syntax.reference create mode 100644 tests/queries/0_stateless/01190_full_attach_syntax.sql diff --git a/docker/test/stress/stress b/docker/test/stress/stress index e8675da1546..60db5ec465c 100755 --- a/docker/test/stress/stress +++ b/docker/test/stress/stress @@ -28,7 +28,7 @@ def get_options(i): options = "" if 0 < i: options += " --order=random" - if i == 1: + if i % 2 == 1: options += " --atomic-db-engine" return options diff --git a/programs/client/Client.cpp b/programs/client/Client.cpp index c9701950dc5..83e4062b1f3 100644 --- a/programs/client/Client.cpp +++ b/programs/client/Client.cpp @@ -919,7 +919,33 @@ private: while (begin < end) { const char * pos = begin; - ASTPtr orig_ast = parseQuery(pos, end, true); + + ASTPtr orig_ast; + try + { + orig_ast = parseQuery(pos, end, true); + } + catch (Exception & e) + { + if (!test_mode) + throw; + + /// Try find test hint for syntax error + const char * end_of_line = find_first_symbols<'\n'>(begin, end); + TestHint hint(true, String(begin, end_of_line - begin)); + if (hint.serverError()) /// Syntax errors are considered as client errors + throw; + if (hint.clientError() != e.code()) + { + if (hint.clientError()) + e.addMessage("\nExpected clinet error: " + std::to_string(hint.clientError())); + throw; + } + + /// It's expected syntax error, skip the line + begin = end_of_line; + continue; + } if (!orig_ast) { diff --git a/src/Interpreters/DatabaseCatalog.cpp b/src/Interpreters/DatabaseCatalog.cpp index 6153f6b52fb..049341918b9 100644 --- a/src/Interpreters/DatabaseCatalog.cpp +++ b/src/Interpreters/DatabaseCatalog.cpp @@ -657,7 +657,10 @@ void DatabaseCatalog::enqueueDroppedTableCleanup(StorageID table_id, StoragePtr /// Table was removed from database. Enqueue removal of its data from disk. time_t drop_time; if (table) + { drop_time = std::chrono::system_clock::to_time_t(std::chrono::system_clock::now()); + table->is_dropped = true; + } else { /// Try load table from metadata to drop it correctly (e.g. remove metadata from zk or remove data from all volumes) @@ -674,6 +677,7 @@ void DatabaseCatalog::enqueueDroppedTableCleanup(StorageID table_id, StoragePtr try { table = createTableFromAST(*create, table_id.getDatabaseName(), data_path, *global_context, false).second; + table->is_dropped = true; } catch (...) { @@ -763,7 +767,6 @@ void DatabaseCatalog::dropTableFinally(const TableMarkedAsDropped & table) const if (table.table) { table.table->drop(); - table.table->is_dropped = true; } /// Even if table is not loaded, try remove its data from disk. diff --git a/src/Interpreters/InterpreterCreateQuery.cpp b/src/Interpreters/InterpreterCreateQuery.cpp index 06973ab029b..d7230940bb2 100644 --- a/src/Interpreters/InterpreterCreateQuery.cpp +++ b/src/Interpreters/InterpreterCreateQuery.cpp @@ -673,6 +673,7 @@ BlockIO InterpreterCreateQuery::createTable(ASTCreateQuery & create) create.attach_short_syntax = true; create.if_not_exists = if_not_exists; } + /// TODO maybe assert table structure if create.attach_short_syntax is false? if (!create.temporary && create.database.empty()) create.database = current_database; diff --git a/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp b/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp index 53ef72f3208..3fa3a7e3e40 100644 --- a/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp +++ b/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp @@ -66,6 +66,7 @@ void MergeTreeWriteAheadLog::dropPart(const String & part_name) writeIntBinary(static_cast(0), *out); writeIntBinary(static_cast(ActionType::DROP_PART), *out); writeStringBinary(part_name, *out); + out->next(); } void MergeTreeWriteAheadLog::rotate(const std::lock_guard &) diff --git a/src/Storages/StorageReplicatedMergeTree.cpp b/src/Storages/StorageReplicatedMergeTree.cpp index 6058632d220..6458fe127da 100644 --- a/src/Storages/StorageReplicatedMergeTree.cpp +++ b/src/Storages/StorageReplicatedMergeTree.cpp @@ -4260,9 +4260,13 @@ bool StorageReplicatedMergeTree::waitForReplicaToProcessLogEntry( * To do this, check its node `log_pointer` - the maximum number of the element taken from `log` + 1. */ - const auto & check_replica_become_inactive = [this, &replica]() + bool waiting_itself = replica == replica_name; + + const auto & stop_waiting = [&]() { - return !getZooKeeper()->exists(zookeeper_path + "/replicas/" + replica + "/is_active"); + bool stop_waiting_itself = waiting_itself && is_dropped; + bool stop_waiting_non_active = !wait_for_non_active && !getZooKeeper()->exists(zookeeper_path + "/replicas/" + replica + "/is_active"); + return stop_waiting_itself || stop_waiting_non_active; }; constexpr auto event_wait_timeout_ms = 1000; @@ -4277,7 +4281,7 @@ bool StorageReplicatedMergeTree::waitForReplicaToProcessLogEntry( LOG_DEBUG(log, "Waiting for {} to pull {} to queue", replica, log_node_name); /// Let's wait until entry gets into the replica queue. - while (wait_for_non_active || !check_replica_become_inactive()) + while (!stop_waiting()) { zkutil::EventPtr event = std::make_shared(); @@ -4325,7 +4329,7 @@ bool StorageReplicatedMergeTree::waitForReplicaToProcessLogEntry( LOG_DEBUG(log, "Waiting for {} to pull {} to queue", replica, log_node_name); /// Let's wait until the entry gets into the replica queue. - while (wait_for_non_active || !check_replica_become_inactive()) + while (!stop_waiting()) { zkutil::EventPtr event = std::make_shared(); @@ -4378,10 +4382,8 @@ bool StorageReplicatedMergeTree::waitForReplicaToProcessLogEntry( /// Third - wait until the entry disappears from the replica queue or replica become inactive. String path_to_wait_on = zookeeper_path + "/replicas/" + replica + "/queue/" + queue_entry_to_wait_for; - if (wait_for_non_active) - return getZooKeeper()->waitForDisappear(path_to_wait_on); - return getZooKeeper()->waitForDisappear(path_to_wait_on, check_replica_become_inactive); + return getZooKeeper()->waitForDisappear(path_to_wait_on, stop_waiting); } diff --git a/src/Storages/System/StorageSystemTables.cpp b/src/Storages/System/StorageSystemTables.cpp index 5b7dad836e9..0ad961ad7d8 100644 --- a/src/Storages/System/StorageSystemTables.cpp +++ b/src/Storages/System/StorageSystemTables.cpp @@ -344,6 +344,12 @@ protected: { ASTPtr ast = database->tryGetCreateTableQuery(table_name, context); + if (ast && !context.getSettingsRef().show_table_uuid_in_table_create_query_if_not_nil) + { + auto & create = ast->as(); + create.uuid = UUIDHelpers::Nil; + } + if (columns_mask[src_index++]) res_columns[res_index++]->insert(ast ? queryToString(ast) : ""); diff --git a/tests/queries/0_stateless/00116_storage_set.sql b/tests/queries/0_stateless/00116_storage_set.sql index aa93a0620d0..0eeed7e859a 100644 --- a/tests/queries/0_stateless/00116_storage_set.sql +++ b/tests/queries/0_stateless/00116_storage_set.sql @@ -19,7 +19,7 @@ INSERT INTO set2 VALUES ('abc'), ('World'); SELECT arrayJoin(['Hello', 'test', 'World', 'world', 'abc', 'xyz']) AS s WHERE s IN set2; DETACH TABLE set2; -ATTACH TABLE set2 (x String) ENGINE = Set; +ATTACH TABLE set2; SELECT arrayJoin(['Hello', 'test', 'World', 'world', 'abc', 'xyz']) AS s WHERE s IN set2; diff --git a/tests/queries/0_stateless/00180_attach_materialized_view.sql b/tests/queries/0_stateless/00180_attach_materialized_view.sql index 089e4926bcf..d674c0bd277 100644 --- a/tests/queries/0_stateless/00180_attach_materialized_view.sql +++ b/tests/queries/0_stateless/00180_attach_materialized_view.sql @@ -6,7 +6,7 @@ CREATE TABLE t_00180 (x UInt8) ENGINE = Null; CREATE MATERIALIZED VIEW mv_00180 ENGINE = Null AS SELECT * FROM t_00180; DETACH TABLE mv_00180; -ATTACH MATERIALIZED VIEW mv_00180 ENGINE = Null AS SELECT * FROM t_00180; +ATTACH TABLE mv_00180; DROP TABLE t_00180; DROP TABLE mv_00180; diff --git a/tests/queries/0_stateless/00226_zookeeper_deduplication_and_unexpected_parts.sql b/tests/queries/0_stateless/00226_zookeeper_deduplication_and_unexpected_parts.sql index 623218af167..c14ce53d4a3 100644 --- a/tests/queries/0_stateless/00226_zookeeper_deduplication_and_unexpected_parts.sql +++ b/tests/queries/0_stateless/00226_zookeeper_deduplication_and_unexpected_parts.sql @@ -21,7 +21,7 @@ INSERT INTO deduplication (x) VALUES (1); SELECT * FROM deduplication; DETACH TABLE deduplication; -ATTACH TABLE deduplication (d Date DEFAULT '2015-01-01', x Int8) ENGINE = ReplicatedMergeTree('/clickhouse/tables/test_00226/deduplication', 'r1', d, x, 1); +ATTACH TABLE deduplication; SELECT * FROM deduplication; diff --git a/tests/queries/0_stateless/00281_compile_sizeof_packed.re b/tests/queries/0_stateless/00281_compile_sizeof_packed.re deleted file mode 100644 index e69de29bb2d..00000000000 diff --git a/tests/queries/0_stateless/00311_array_primary_key.sql b/tests/queries/0_stateless/00311_array_primary_key.sql index 0ea368609da..0e066c64f89 100644 --- a/tests/queries/0_stateless/00311_array_primary_key.sql +++ b/tests/queries/0_stateless/00311_array_primary_key.sql @@ -11,7 +11,7 @@ INSERT INTO array_pk VALUES ([5, 6], 'ghi', 6); SELECT * FROM array_pk ORDER BY n; DETACH TABLE array_pk; -ATTACH TABLE array_pk (key Array(UInt8), s String, n UInt64, d Date MATERIALIZED '2000-01-01') ENGINE = MergeTree(d, (key, s, n), 1); +ATTACH TABLE array_pk; SELECT * FROM array_pk ORDER BY n; diff --git a/tests/queries/0_stateless/00423_storage_log_single_thread.sql b/tests/queries/0_stateless/00423_storage_log_single_thread.sql index 7d5e14c9ee5..8eff9323564 100644 --- a/tests/queries/0_stateless/00423_storage_log_single_thread.sql +++ b/tests/queries/0_stateless/00423_storage_log_single_thread.sql @@ -5,7 +5,7 @@ SELECT * FROM log LIMIT 1; SELECT * FROM log; DETACH TABLE log; -ATTACH TABLE log (s String) ENGINE = Log; +ATTACH TABLE log; SELECT * FROM log; SELECT * FROM log LIMIT 1; @@ -15,13 +15,13 @@ INSERT INTO log VALUES ('Hello'), ('World'); SELECT * FROM log LIMIT 1; DETACH TABLE log; -ATTACH TABLE log (s String) ENGINE = Log; +ATTACH TABLE log; SELECT * FROM log LIMIT 1; SELECT * FROM log; DETACH TABLE log; -ATTACH TABLE log (s String) ENGINE = Log; +ATTACH TABLE log; SELECT * FROM log; SELECT * FROM log LIMIT 1; diff --git a/tests/queries/0_stateless/00816_long_concurrent_alter_column.sh b/tests/queries/0_stateless/00816_long_concurrent_alter_column.sh index 93421e003f6..8fdd6654bae 100755 --- a/tests/queries/0_stateless/00816_long_concurrent_alter_column.sh +++ b/tests/queries/0_stateless/00816_long_concurrent_alter_column.sh @@ -11,34 +11,34 @@ echo "CREATE TABLE concurrent_alter_column (ts DATETIME) ENGINE = MergeTree PART function thread1() { while true; do - for i in {1..500}; do echo "ALTER TABLE concurrent_alter_column ADD COLUMN c$i DOUBLE;"; done | ${CLICKHOUSE_CLIENT} -n --query_id=alter1 + for i in {1..500}; do echo "ALTER TABLE concurrent_alter_column ADD COLUMN c$i DOUBLE;"; done | ${CLICKHOUSE_CLIENT} -n --query_id=alter_00816_1 done } function thread2() { while true; do - echo "ALTER TABLE concurrent_alter_column ADD COLUMN d DOUBLE" | ${CLICKHOUSE_CLIENT} --query_id=alter2; + echo "ALTER TABLE concurrent_alter_column ADD COLUMN d DOUBLE" | ${CLICKHOUSE_CLIENT} --query_id=alter_00816_2; sleep "$(echo 0.0$RANDOM)"; - echo "ALTER TABLE concurrent_alter_column DROP COLUMN d" | ${CLICKHOUSE_CLIENT} --query_id=alter2; + echo "ALTER TABLE concurrent_alter_column DROP COLUMN d" | ${CLICKHOUSE_CLIENT} --query_id=alter_00816_2; done } function thread3() { while true; do - echo "ALTER TABLE concurrent_alter_column ADD COLUMN e DOUBLE" | ${CLICKHOUSE_CLIENT} --query_id=alter3; + echo "ALTER TABLE concurrent_alter_column ADD COLUMN e DOUBLE" | ${CLICKHOUSE_CLIENT} --query_id=alter_00816_3; sleep "$(echo 0.0$RANDOM)"; - echo "ALTER TABLE concurrent_alter_column DROP COLUMN e" | ${CLICKHOUSE_CLIENT} --query_id=alter3; + echo "ALTER TABLE concurrent_alter_column DROP COLUMN e" | ${CLICKHOUSE_CLIENT} --query_id=alter_00816_3; done } function thread4() { while true; do - echo "ALTER TABLE concurrent_alter_column ADD COLUMN f DOUBLE" | ${CLICKHOUSE_CLIENT} --query_id=alter4; + echo "ALTER TABLE concurrent_alter_column ADD COLUMN f DOUBLE" | ${CLICKHOUSE_CLIENT} --query_id=alter_00816_4; sleep "$(echo 0.0$RANDOM)"; - echo "ALTER TABLE concurrent_alter_column DROP COLUMN f" | ${CLICKHOUSE_CLIENT} --query_id=alter4; + echo "ALTER TABLE concurrent_alter_column DROP COLUMN f" | ${CLICKHOUSE_CLIENT} --query_id=alter_00816_4; done } @@ -57,9 +57,18 @@ timeout $TIMEOUT bash -c thread4 2> /dev/null & wait -echo "DROP TABLE concurrent_alter_column" | ${CLICKHOUSE_CLIENT} +echo "DROP TABLE concurrent_alter_column NO DELAY" | ${CLICKHOUSE_CLIENT} # NO DELAY has effect only for Atomic database + +db_engine=`$CLICKHOUSE_CLIENT -q "SELECT engine FROM system.databases WHERE name=currentDatabase()"` +if [[ $db_engine == "Atomic" ]]; then + # DROP is non-blocking, so wait for alters + while true; do + $CLICKHOUSE_CLIENT -q "SELECT c = 0 FROM (SELECT count() as c FROM system.processes WHERE query_id LIKE 'alter_00816_%')" | grep 1 > /dev/null && break; + sleep 1; + done +fi # Check for deadlocks -echo "SELECT * FROM system.processes WHERE query_id LIKE 'alter%'" | ${CLICKHOUSE_CLIENT} +echo "SELECT * FROM system.processes WHERE query_id LIKE 'alter_00816_%'" | ${CLICKHOUSE_CLIENT} echo 'did not crash' diff --git a/tests/queries/0_stateless/01190_full_attach_syntax.reference b/tests/queries/0_stateless/01190_full_attach_syntax.reference new file mode 100644 index 00000000000..619861849c8 --- /dev/null +++ b/tests/queries/0_stateless/01190_full_attach_syntax.reference @@ -0,0 +1,13 @@ +CREATE DICTIONARY test_01190.dict\n(\n `key` UInt64 DEFAULT 0,\n `col` UInt8 DEFAULT 1\n)\nPRIMARY KEY key\nSOURCE(CLICKHOUSE(HOST \'localhost\' PORT 9000 USER \'default\' TABLE \'table_for_dict\' PASSWORD \'\' DB \'test_01190\'))\nLIFETIME(MIN 1 MAX 10)\nLAYOUT(FLAT()) +CREATE DICTIONARY test_01190.dict\n(\n `key` UInt64 DEFAULT 0,\n `col` UInt8 DEFAULT 1\n)\nPRIMARY KEY key\nSOURCE(CLICKHOUSE(HOST \'localhost\' PORT 9000 USER \'default\' TABLE \'table_for_dict\' PASSWORD \'\' DB \'test_01190\'))\nLIFETIME(MIN 1 MAX 10)\nLAYOUT(FLAT()) +CREATE TABLE default.log\n(\n `s` String\n)\nENGINE = Log +CREATE TABLE default.log\n(\n `s` String\n)\nENGINE = Log() +test +CREATE TABLE default.mt\n(\n `key` Array(UInt8),\n `s` String,\n `n` UInt64,\n `d` Date MATERIALIZED \'2000-01-01\'\n)\nENGINE = MergeTree(d, (key, s, n), 1) +[1,2] Hello 2 +CREATE TABLE default.mt\n(\n `key` Array(UInt8),\n `s` String,\n `n` UInt64,\n `d` Date\n)\nENGINE = MergeTree(d, (key, s, n), 1) +CREATE MATERIALIZED VIEW default.mv\n(\n `s` String\n)\nENGINE = Null AS\nSELECT *\nFROM default.log +CREATE MATERIALIZED VIEW default.mv\n(\n `s` String\n)\nENGINE = Null AS\nSELECT *\nFROM default.log +CREATE MATERIALIZED VIEW default.mv\n(\n `key` Array(UInt8),\n `s` String,\n `n` UInt64,\n `d` Date\n)\nENGINE = Null AS\nSELECT *\nFROM default.mt +CREATE LIVE VIEW default.lv\n(\n `1` UInt8\n) AS\nSELECT 1 +CREATE LIVE VIEW default.lv\n(\n `1` UInt8\n) AS\nSELECT 1 diff --git a/tests/queries/0_stateless/01190_full_attach_syntax.sql b/tests/queries/0_stateless/01190_full_attach_syntax.sql new file mode 100644 index 00000000000..3a91eccc8cd --- /dev/null +++ b/tests/queries/0_stateless/01190_full_attach_syntax.sql @@ -0,0 +1,66 @@ +DROP DATABASE IF EXISTS test_01190; +CREATE DATABASE test_01190; + +CREATE TABLE test_01190.table_for_dict (key UInt64, col UInt8) ENGINE = Memory; + +CREATE DICTIONARY test_01190.dict (key UInt64 DEFAULT 0, col UInt8 DEFAULT 1) PRIMARY KEY key SOURCE(CLICKHOUSE(HOST 'localhost' PORT 9000 USER 'default' TABLE 'table_for_dict' PASSWORD '' DB 'test_01190')) LIFETIME(MIN 1 MAX 10) LAYOUT(FLAT()); + +SHOW CREATE DICTIONARY test_01190.dict; + +DETACH DICTIONARY test_01190.dict; +ATTACH TABLE test_01190.dict; -- { serverError 80 } +-- Full ATTACH syntax is not allowed for dictionaries +ATTACH DICTIONARY test_01190.dict (key UInt64 DEFAULT 0, col UInt8 DEFAULT 42) PRIMARY KEY key SOURCE(CLICKHOUSE(HOST 'localhost' PORT 9000 USER 'default' TABLE 'table_for_dict' PASSWORD '' DB 'test_01190')) LIFETIME(MIN 1 MAX 100) LAYOUT(FLAT()); -- { clientError 62 } +ATTACH DICTIONARY test_01190.dict; +SHOW CREATE DICTIONARY test_01190.dict; + +DROP DATABASE test_01190; + + +DROP TABLE IF EXISTS log; +DROP TABLE IF EXISTS mt; +DROP TABLE IF EXISTS mv; +DROP TABLE IF EXISTS lv; + +CREATE TABLE log ENGINE = Log AS SELECT 'test' AS s; +SHOW CREATE log; +DETACH TABLE log; +ATTACH DICTIONARY log; -- { serverError 487 } +ATTACH TABLE log (s String) ENGINE = Log(); +SHOW CREATE log; +SELECT * FROM log; + +DROP TABLE IF EXISTS mt; +CREATE TABLE mt (key Array(UInt8), s String, n UInt64, d Date MATERIALIZED '2000-01-01') ENGINE = MergeTree(d, (key, s, n), 1); +INSERT INTO mt VALUES ([1, 2], 'Hello', 2); +DETACH TABLE mt; +ATTACH TABLE mt (key Array(UInt8), s String, n UInt64, d Date MATERIALIZED '2000-01-01') ENGINE = MergeTree ORDER BY (key, s, n) PARTITION BY toYYYYMM(d); -- { serverError 342 } +ATTACH TABLE mt (key Array(UInt8), s String, n UInt64, d Date MATERIALIZED '2000-01-01') ENGINE = MergeTree(d, (key, s, n), 1); +SHOW CREATE mt; +SELECT * FROM mt; +DETACH TABLE mt; +ATTACH TABLE mt (key Array(UInt8), s String, n UInt64, d Date) ENGINE = MergeTree(d, (key, s, n), 1); -- It works (with Ordinary database), but probably it shouldn't +SHOW CREATE mt; + +CREATE MATERIALIZED VIEW mv ENGINE = Null AS SELECT * FROM log; +SHOW CREATE mv; +DETACH VIEW mv; +ATTACH MATERIALIZED VIEW mv ENGINE = Null AS SELECT * FROM log; +SHOW CREATE mv; +DETACH VIEW mv; +ATTACH MATERIALIZED VIEW mv ENGINE = Null AS SELECT * FROM mt; -- It works (with Ordinary database), but probably it shouldn't +SHOW CREATE mv; + +SET allow_experimental_live_view = 1; +CREATE LIVE VIEW lv AS SELECT 1; +SHOW CREATE lv; +DETACH VIEW lv; +ATTACH LIVE VIEW lv AS SELECT 1; +SHOW CREATE lv; + +DROP TABLE log; +DROP TABLE mt; +DROP TABLE mv; +DROP TABLE lv; + + diff --git a/tests/queries/0_stateless/01305_replica_create_drop_zookeeper.sh b/tests/queries/0_stateless/01305_replica_create_drop_zookeeper.sh index 0a47c6df46c..1313830d589 100755 --- a/tests/queries/0_stateless/01305_replica_create_drop_zookeeper.sh +++ b/tests/queries/0_stateless/01305_replica_create_drop_zookeeper.sh @@ -7,11 +7,21 @@ set -e function thread() { - while true; do - $CLICKHOUSE_CLIENT -n -q "DROP TABLE IF EXISTS test_table_$1; - CREATE TABLE test_table_$1 (a UInt8) ENGINE = ReplicatedMergeTree('/clickhouse/tables/test_01305/alter_table', 'r_$1') ORDER BY tuple();" 2>&1 | - grep -vP '(^$)|(^Received exception from server)|(^\d+\. )|because the last replica of the table was dropped right now|is already started to be removing by another replica right now|is already finished removing by another replica right now|Removing leftovers from table|Another replica was suddenly created|was successfully removed from ZooKeeper|was created by another server at the same moment|was suddenly removed|some other replicas were created at the same time' - done + db_engine=`$CLICKHOUSE_CLIENT -q "SELECT engine FROM system.databases WHERE name=currentDatabase()"` + if [[ $db_engine == "Atomic" ]]; then + # Ignore "Replica already exists" exception + while true; do + $CLICKHOUSE_CLIENT -n -q "DROP TABLE IF EXISTS test_table_$1 NO DELAY; + CREATE TABLE test_table_$1 (a UInt8) ENGINE = ReplicatedMergeTree('/clickhouse/tables/test_01305/alter_table', 'r_$1') ORDER BY tuple();" 2>&1 | + grep -vP '(^$)|(^Received exception from server)|(^\d+\. )|because the last replica of the table was dropped right now|is already started to be removing by another replica right now|is already finished removing by another replica right now|Removing leftovers from table|Another replica was suddenly created|was successfully removed from ZooKeeper|was created by another server at the same moment|was suddenly removed|some other replicas were created at the same time|already exists' + done + else + while true; do + $CLICKHOUSE_CLIENT -n -q "DROP TABLE IF EXISTS test_table_$1; + CREATE TABLE test_table_$1 (a UInt8) ENGINE = ReplicatedMergeTree('/clickhouse/tables/test_01305/alter_table', 'r_$1') ORDER BY tuple();" 2>&1 | + grep -vP '(^$)|(^Received exception from server)|(^\d+\. )|because the last replica of the table was dropped right now|is already started to be removing by another replica right now|is already finished removing by another replica right now|Removing leftovers from table|Another replica was suddenly created|was successfully removed from ZooKeeper|was created by another server at the same moment|was suddenly removed|some other replicas were created at the same time' + done + fi } diff --git a/tests/queries/1_stateful/00065_loyalty_with_storage_join.sql b/tests/queries/1_stateful/00065_loyalty_with_storage_join.sql index 15a2a75cf58..515a2410583 100644 --- a/tests/queries/1_stateful/00065_loyalty_with_storage_join.sql +++ b/tests/queries/1_stateful/00065_loyalty_with_storage_join.sql @@ -22,7 +22,7 @@ GROUP BY loyalty ORDER BY loyalty ASC; DETACH TABLE join; -ATTACH TABLE join (UserID UInt64, loyalty Int8) ENGINE = Join(SEMI, LEFT, UserID); +ATTACH TABLE join; SELECT loyalty, diff --git a/tests/queries/skip_list.json b/tests/queries/skip_list.json index adfc5f0e582..efd622402b2 100644 --- a/tests/queries/skip_list.json +++ b/tests/queries/skip_list.json @@ -3,10 +3,8 @@ */ { "thread-sanitizer": [ - "00281", "00877", "00985", - "avx2", "query_profiler", "memory_profiler", /// 01083 and 00505 and 00505 are critical and temproray disabled @@ -21,9 +19,7 @@ "01193_metadata_loading" ], "address-sanitizer": [ - "00281", "00877", - "avx2", "query_profiler", "memory_profiler", "odbc_roundtrip", @@ -31,9 +27,7 @@ "01193_metadata_loading" ], "ub-sanitizer": [ - "00281", "capnproto", - "avx2", "query_profiler", "memory_profiler", "01103_check_cpu_instructions_at_startup", @@ -41,9 +35,7 @@ "01193_metadata_loading" ], "memory-sanitizer": [ - "00281", "capnproto", - "avx2", "query_profiler", "memory_profiler", "01103_check_cpu_instructions_at_startup", @@ -53,8 +45,6 @@ "01193_metadata_loading" ], "debug-build": [ - "00281", - "avx2", "query_profiler", "memory_profiler", "00899_long_attach", @@ -70,12 +60,10 @@ ], "unbundled-build": [ "00429", - "00428", "00877", "pocopatch", "parquet", "xxhash", - "avx2", "_h3", "query_profiler", "memory_profiler", @@ -98,33 +86,19 @@ "01455_time_zones" ], "release-build": [ - "avx2" ], "database-atomic": [ - "00065_loyalty_with_storage_join", - "avx", /// Inner tables of materialized views have different names "00738_lock_for_inner_table", - "00699_materialized_view_mutations", "00609_mv_index_in_in", "00510_materizlized_view_and_deduplication_zookeeper", - /// Create queries contain UUID + /// Different database engine "00604_show_create_database", - "00080_show_tables_and_system_tables", - "01272_suspicious_codecs", /// UUID must be specified in ATTACH TABLE - "01249_bad_arguments_for_bloom_filter", - "00423_storage_log_single_thread", - "00311_array_primary_key", - "00226_zookeeper_deduplication_and_unexpected_parts", - "00180_attach_materialized_view", - "00116_storage_set", + "01190_full_attach_syntax", /// Assumes blocking DROP - "00816_long_concurrent_alter_column", - "00992_system_parts_race_condition_zookeeper", /// FIXME "01320_create_sync_race_condition", - "01305_replica_create_drop_zookeeper", - "01130_in_memory_parts_partitons", + /// Internal distionary name is different "01225_show_create_table_from_dictionary", "01224_no_superfluous_dict_reload" ], @@ -132,7 +106,6 @@ /// These tests fail with compact parts, because they /// check some implementation defined things /// like checksums, computed granularity, ProfileEvents, etc. - "avx", "01045_order_by_pk_special_storages", "01042_check_query_and_last_granule_size", "00961_checksums_in_system_parts_columns_table", From 50dee3f4493d7ffb2c75d195cc39862f8f8d8a86 Mon Sep 17 00:00:00 2001 From: alesapin Date: Wed, 9 Sep 2020 17:43:17 +0300 Subject: [PATCH 039/121] Remove false-positive warning --- src/Storages/MergeTree/MergeTreePartition.cpp | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/src/Storages/MergeTree/MergeTreePartition.cpp b/src/Storages/MergeTree/MergeTreePartition.cpp index 4a846f63b7c..8ef3e458871 100644 --- a/src/Storages/MergeTree/MergeTreePartition.cpp +++ b/src/Storages/MergeTree/MergeTreePartition.cpp @@ -29,6 +29,9 @@ String MergeTreePartition::getID(const MergeTreeData & storage) const return getID(storage.getInMemoryMetadataPtr()->getPartitionKey().sample_block); } +#pragma GCC diagnostic push +#pragma GCC diagnostic ignored "-Wstringop-overflow" + /// NOTE: This ID is used to create part names which are then persisted in ZK and as directory names on the file system. /// So if you want to change this method, be sure to guarantee compatibility with existing table data. String MergeTreePartition::getID(const Block & partition_key_sample) const @@ -87,6 +90,8 @@ String MergeTreePartition::getID(const Block & partition_key_sample) const return result; } +#pragma GCC diagnostic pop + void MergeTreePartition::serializeText(const MergeTreeData & storage, WriteBuffer & out, const FormatSettings & format_settings) const { auto metadata_snapshot = storage.getInMemoryMetadataPtr(); From 7f4106687cb14491246f218654ed8a0a3b751b29 Mon Sep 17 00:00:00 2001 From: Alexander Tokmakov Date: Wed, 9 Sep 2020 19:23:31 +0300 Subject: [PATCH 040/121] fix --- tests/queries/0_stateless/01114_database_atomic.reference | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/queries/0_stateless/01114_database_atomic.reference b/tests/queries/0_stateless/01114_database_atomic.reference index 7980819f9af..a79784230a6 100644 --- a/tests/queries/0_stateless/01114_database_atomic.reference +++ b/tests/queries/0_stateless/01114_database_atomic.reference @@ -7,7 +7,7 @@ test_01114_3 Ordinary test_01114_3 test_01114_3 1 20 100 CREATE TABLE test_01114_2.mt UUID \'00001114-0000-4000-8000-000000000002\'\n(\n `n` UInt64\n)\nENGINE = MergeTree()\nPARTITION BY n % 5\nORDER BY tuple()\nSETTINGS index_granularity = 8192 -mt 00001114-0000-4000-8000-000000000002 CREATE TABLE test_01114_2.mt UUID \'00001114-0000-4000-8000-000000000002\' (`n` UInt64) ENGINE = MergeTree() PARTITION BY n % 5 ORDER BY tuple() SETTINGS index_granularity = 8192 +mt 00001114-0000-4000-8000-000000000002 CREATE TABLE test_01114_2.mt (`n` UInt64) ENGINE = MergeTree() PARTITION BY n % 5 ORDER BY tuple() SETTINGS index_granularity = 8192 20 CREATE TABLE test_01114_1.mt UUID \'00001114-0000-4000-8000-000000000001\'\n(\n `n` UInt64\n)\nENGINE = MergeTree()\nPARTITION BY n % 5\nORDER BY tuple()\nSETTINGS index_granularity = 8192 CREATE TABLE test_01114_2.mt UUID \'00001114-0000-4000-8000-000000000002\'\n(\n `n` UInt64\n)\nENGINE = MergeTree()\nPARTITION BY n % 5\nORDER BY tuple()\nSETTINGS index_granularity = 8192 From 62428845a0fdcaaa19ecc5fd33f3ecd849104cf5 Mon Sep 17 00:00:00 2001 From: alesapin Date: Wed, 9 Sep 2020 19:47:06 +0300 Subject: [PATCH 041/121] Bug in mutation --- src/Columns/ColumnVector.h | 7 ++++--- .../0_stateless/01475_mutation_with_if.reference | 1 + .../0_stateless/01475_mutation_with_if.sql | 16 ++++++++++++++++ 3 files changed, 21 insertions(+), 3 deletions(-) create mode 100644 tests/queries/0_stateless/01475_mutation_with_if.reference create mode 100644 tests/queries/0_stateless/01475_mutation_with_if.sql diff --git a/src/Columns/ColumnVector.h b/src/Columns/ColumnVector.h index 1090de556a0..55ab67d6214 100644 --- a/src/Columns/ColumnVector.h +++ b/src/Columns/ColumnVector.h @@ -7,6 +7,7 @@ #include #include #include +#include namespace DB @@ -130,7 +131,7 @@ public: void insertFrom(const IColumn & src, size_t n) override { - data.push_back(static_cast(src).getData()[n]); + data.push_back(assert_cast(src).getData()[n]); } void insertData(const char * pos, size_t) override @@ -205,14 +206,14 @@ public: /// This method implemented in header because it could be possibly devirtualized. int compareAt(size_t n, size_t m, const IColumn & rhs_, int nan_direction_hint) const override { - return CompareHelper::compare(data[n], static_cast(rhs_).data[m], nan_direction_hint); + return CompareHelper::compare(data[n], assert_cast(rhs_).data[m], nan_direction_hint); } void compareColumn(const IColumn & rhs, size_t rhs_row_num, PaddedPODArray * row_indexes, PaddedPODArray & compare_results, int direction, int nan_direction_hint) const override { - return this->template doCompareColumn(static_cast(rhs), rhs_row_num, row_indexes, + return this->template doCompareColumn(assert_cast(rhs), rhs_row_num, row_indexes, compare_results, direction, nan_direction_hint); } diff --git a/tests/queries/0_stateless/01475_mutation_with_if.reference b/tests/queries/0_stateless/01475_mutation_with_if.reference new file mode 100644 index 00000000000..2874a18147f --- /dev/null +++ b/tests/queries/0_stateless/01475_mutation_with_if.reference @@ -0,0 +1 @@ +1 150 diff --git a/tests/queries/0_stateless/01475_mutation_with_if.sql b/tests/queries/0_stateless/01475_mutation_with_if.sql new file mode 100644 index 00000000000..6f0ef8924be --- /dev/null +++ b/tests/queries/0_stateless/01475_mutation_with_if.sql @@ -0,0 +1,16 @@ +DROP TABLE IF EXISTS mutation_table; +CREATE TABLE mutation_table ( + id int, + price Nullable(Int32) +) +ENGINE = MergeTree() +PARTITION BY id +ORDER BY id; + +INSERT INTO mutation_table (id, price) VALUES (1, 100); + +ALTER TABLE mutation_table UPDATE price = 150 WHERE id = 1 SETTINGS mutations_sync = 2; + +SELECT * FROM mutation_table; + +DROP TABLE IF EXISTS mutation_table; From 2a9ab482792cdadf0d4e2365c3d11494a3e38230 Mon Sep 17 00:00:00 2001 From: Vitaly Baranov Date: Tue, 8 Sep 2020 02:08:17 +0300 Subject: [PATCH 042/121] Use join() instead of detach() for the no_users_thread in StorageLiveView. --- src/Interpreters/Context.cpp | 8 + src/Interpreters/Context.h | 8 +- src/Interpreters/InterpreterDropQuery.h | 1 + .../LiveView/LiveViewBlockInputStream.h | 15 +- .../LiveView/LiveViewEventsBlockInputStream.h | 14 +- src/Storages/LiveView/StorageLiveView.cpp | 144 +---------------- src/Storages/LiveView/StorageLiveView.h | 23 ++- .../LiveView/TemporaryLiveViewCleaner.cpp | 148 ++++++++++++++++++ .../LiveView/TemporaryLiveViewCleaner.h | 51 ++++++ src/Storages/ya.make | 1 + 10 files changed, 233 insertions(+), 180 deletions(-) create mode 100644 src/Storages/LiveView/TemporaryLiveViewCleaner.cpp create mode 100644 src/Storages/LiveView/TemporaryLiveViewCleaner.h diff --git a/src/Interpreters/Context.cpp b/src/Interpreters/Context.cpp index 70cf41a679c..3c4c095cc26 100644 --- a/src/Interpreters/Context.cpp +++ b/src/Interpreters/Context.cpp @@ -23,6 +23,7 @@ #include #include #include +#include #include #include #include @@ -423,6 +424,7 @@ struct ContextShared if (system_logs) system_logs->shutdown(); + TemporaryLiveViewCleaner::shutdown(); DatabaseCatalog::shutdown(); /// Preemptive destruction is important, because these objects may have a refcount to ContextShared (cyclic reference). @@ -479,6 +481,12 @@ Context Context::createGlobal(ContextShared * shared) return res; } +void Context::initGlobal() +{ + DatabaseCatalog::init(this); + TemporaryLiveViewCleaner::init(*this); +} + SharedContextHolder Context::createShared() { return SharedContextHolder(std::make_unique()); diff --git a/src/Interpreters/Context.h b/src/Interpreters/Context.h index c8d13baa9ae..743c92d56b5 100644 --- a/src/Interpreters/Context.h +++ b/src/Interpreters/Context.h @@ -445,11 +445,7 @@ public: void makeQueryContext() { query_context = this; } void makeSessionContext() { session_context = this; } - void makeGlobalContext() - { - global_context = this; - DatabaseCatalog::init(this); - } + void makeGlobalContext() { initGlobal(); global_context = this; } const Settings & getSettingsRef() const { return settings; } @@ -622,6 +618,8 @@ public: private: std::unique_lock getLock() const; + void initGlobal(); + /// Compute and set actual user settings, client_info.current_user should be set void calculateAccessRights(); diff --git a/src/Interpreters/InterpreterDropQuery.h b/src/Interpreters/InterpreterDropQuery.h index 80bd6c6531a..b54736b5c21 100644 --- a/src/Interpreters/InterpreterDropQuery.h +++ b/src/Interpreters/InterpreterDropQuery.h @@ -10,6 +10,7 @@ namespace DB { class Context; using DatabaseAndTable = std::pair; +class AccessRightsElements; /** Allow to either drop table with all its data (DROP), * or remove information about table (just forget) from server (DETACH), diff --git a/src/Storages/LiveView/LiveViewBlockInputStream.h b/src/Storages/LiveView/LiveViewBlockInputStream.h index 7cab2cb41ed..737e76754c5 100644 --- a/src/Storages/LiveView/LiveViewBlockInputStream.h +++ b/src/Storages/LiveView/LiveViewBlockInputStream.h @@ -16,27 +16,17 @@ class LiveViewBlockInputStream : public IBlockInputStream using NonBlockingResult = std::pair; public: - ~LiveViewBlockInputStream() override - { - /// Start storage no users thread - /// if we are the last active user - if (!storage->is_dropped && blocks_ptr.use_count() < 3) - storage->startNoUsersThread(temporary_live_view_timeout_sec); - } - LiveViewBlockInputStream(std::shared_ptr storage_, std::shared_ptr blocks_ptr_, std::shared_ptr blocks_metadata_ptr_, std::shared_ptr active_ptr_, const bool has_limit_, const UInt64 limit_, - const UInt64 heartbeat_interval_sec_, - const UInt64 temporary_live_view_timeout_sec_) + const UInt64 heartbeat_interval_sec_) : storage(std::move(storage_)), blocks_ptr(std::move(blocks_ptr_)), blocks_metadata_ptr(std::move(blocks_metadata_ptr_)), active_ptr(std::move(active_ptr_)), has_limit(has_limit_), limit(limit_), - heartbeat_interval_usec(heartbeat_interval_sec_ * 1000000), - temporary_live_view_timeout_sec(temporary_live_view_timeout_sec_) + heartbeat_interval_usec(heartbeat_interval_sec_ * 1000000) { /// grab active pointer active = active_ptr.lock(); @@ -205,7 +195,6 @@ private: Int64 num_updates = -1; bool end_of_blocks = false; UInt64 heartbeat_interval_usec; - UInt64 temporary_live_view_timeout_sec; UInt64 last_event_timestamp_usec = 0; }; diff --git a/src/Storages/LiveView/LiveViewEventsBlockInputStream.h b/src/Storages/LiveView/LiveViewEventsBlockInputStream.h index ac5e7e3d6fd..4060b17c1ed 100644 --- a/src/Storages/LiveView/LiveViewEventsBlockInputStream.h +++ b/src/Storages/LiveView/LiveViewEventsBlockInputStream.h @@ -34,13 +34,6 @@ class LiveViewEventsBlockInputStream : public IBlockInputStream using NonBlockingResult = std::pair; public: - ~LiveViewEventsBlockInputStream() override - { - /// Start storage no users thread - /// if we are the last active user - if (!storage->is_dropped && blocks_ptr.use_count() < 3) - storage->startNoUsersThread(temporary_live_view_timeout_sec); - } /// length default -2 because we want LIMIT to specify number of updates so that LIMIT 1 waits for 1 update /// and LIMIT 0 just returns data without waiting for any updates LiveViewEventsBlockInputStream(std::shared_ptr storage_, @@ -48,14 +41,12 @@ public: std::shared_ptr blocks_metadata_ptr_, std::shared_ptr active_ptr_, const bool has_limit_, const UInt64 limit_, - const UInt64 heartbeat_interval_sec_, - const UInt64 temporary_live_view_timeout_sec_) + const UInt64 heartbeat_interval_sec_) : storage(std::move(storage_)), blocks_ptr(std::move(blocks_ptr_)), blocks_metadata_ptr(std::move(blocks_metadata_ptr_)), active_ptr(std::move(active_ptr_)), has_limit(has_limit_), limit(limit_), - heartbeat_interval_usec(heartbeat_interval_sec_ * 1000000), - temporary_live_view_timeout_sec(temporary_live_view_timeout_sec_) + heartbeat_interval_usec(heartbeat_interval_sec_ * 1000000) { /// grab active pointer active = active_ptr.lock(); @@ -236,7 +227,6 @@ private: Int64 num_updates = -1; bool end_of_blocks = false; UInt64 heartbeat_interval_usec; - UInt64 temporary_live_view_timeout_sec; UInt64 last_event_timestamp_usec = 0; Poco::Timestamp timestamp; }; diff --git a/src/Storages/LiveView/StorageLiveView.cpp b/src/Storages/LiveView/StorageLiveView.cpp index 54ac5bcc791..b16c02eec6b 100644 --- a/src/Storages/LiveView/StorageLiveView.cpp +++ b/src/Storages/LiveView/StorageLiveView.cpp @@ -12,10 +12,8 @@ limitations under the License. */ #include #include #include -#include #include #include -#include #include #include #include @@ -31,6 +29,7 @@ limitations under the License. */ #include #include #include +#include #include #include @@ -276,7 +275,7 @@ StorageLiveView::StorageLiveView( if (query.live_view_timeout) { is_temporary = true; - temporary_live_view_timeout = *query.live_view_timeout; + temporary_live_view_timeout = std::chrono::seconds{*query.live_view_timeout}; } blocks_ptr = std::make_shared(); @@ -384,128 +383,21 @@ void StorageLiveView::checkTableCanBeDropped() const } } -void StorageLiveView::noUsersThread(std::shared_ptr storage, const UInt64 & timeout) -{ - bool drop_table = false; - - if (storage->shutdown_called) - return; - - auto table_id = storage->getStorageID(); - { - while (true) - { - std::unique_lock lock(storage->no_users_thread_wakeup_mutex); - if (!storage->no_users_thread_condition.wait_for(lock, std::chrono::seconds(timeout), [&] { return storage->no_users_thread_wakeup; })) - { - storage->no_users_thread_wakeup = false; - if (storage->shutdown_called) - return; - if (storage->hasUsers()) - return; - if (!DatabaseCatalog::instance().getDependencies(table_id).empty()) - continue; - drop_table = true; - } - break; - } - } - - if (drop_table) - { - if (DatabaseCatalog::instance().tryGetTable(table_id, storage->global_context)) - { - try - { - /// We create and execute `drop` query for this table - auto drop_query = std::make_shared(); - drop_query->database = table_id.database_name; - drop_query->table = table_id.table_name; - drop_query->kind = ASTDropQuery::Kind::Drop; - ASTPtr ast_drop_query = drop_query; - InterpreterDropQuery drop_interpreter(ast_drop_query, storage->global_context); - drop_interpreter.execute(); - } - catch (...) - { - tryLogCurrentException(__PRETTY_FUNCTION__); - } - } - } -} - -void StorageLiveView::startNoUsersThread(const UInt64 & timeout) -{ - bool expected = false; - if (!start_no_users_thread_called.compare_exchange_strong(expected, true)) - return; - - if (is_temporary) - { - std::lock_guard no_users_thread_lock(no_users_thread_mutex); - - if (shutdown_called) - return; - - if (no_users_thread.joinable()) - { - { - std::lock_guard lock(no_users_thread_wakeup_mutex); - no_users_thread_wakeup = true; - no_users_thread_condition.notify_one(); - } - no_users_thread.join(); - } - { - std::lock_guard lock(no_users_thread_wakeup_mutex); - no_users_thread_wakeup = false; - } - if (!is_dropped) - no_users_thread = std::thread(&StorageLiveView::noUsersThread, - std::static_pointer_cast(shared_from_this()), timeout); - } - - start_no_users_thread_called = false; -} - void StorageLiveView::startup() { - startNoUsersThread(temporary_live_view_timeout); + if (is_temporary) + TemporaryLiveViewCleaner::instance().addView(std::static_pointer_cast(shared_from_this())); } void StorageLiveView::shutdown() { + shutdown_called = true; DatabaseCatalog::instance().removeDependency(select_table_id, getStorageID()); - bool expected = false; - if (!shutdown_called.compare_exchange_strong(expected, true)) - return; - - /// WATCH queries should be stopped after setting shutdown_called to true. - /// Otherwise livelock is possible for LiveView table in Atomic database: - /// WATCH query will wait for table to be dropped and DatabaseCatalog will wait for queries to finish - - { - std::lock_guard no_users_thread_lock(no_users_thread_mutex); - if (no_users_thread.joinable()) - { - { - std::lock_guard lock(no_users_thread_wakeup_mutex); - no_users_thread_wakeup = true; - no_users_thread_condition.notify_one(); - } - } - } } StorageLiveView::~StorageLiveView() { shutdown(); - - { - std::lock_guard lock(no_users_thread_mutex); - if (no_users_thread.joinable()) - no_users_thread.detach(); - } } void StorageLiveView::drop() @@ -572,18 +464,7 @@ BlockInputStreams StorageLiveView::watch( auto reader = std::make_shared( std::static_pointer_cast(shared_from_this()), blocks_ptr, blocks_metadata_ptr, active_ptr, has_limit, limit, - context.getSettingsRef().live_view_heartbeat_interval.totalSeconds(), - temporary_live_view_timeout); - - { - std::lock_guard no_users_thread_lock(no_users_thread_mutex); - if (no_users_thread.joinable()) - { - std::lock_guard lock(no_users_thread_wakeup_mutex); - no_users_thread_wakeup = true; - no_users_thread_condition.notify_one(); - } - } + context.getSettingsRef().live_view_heartbeat_interval.totalSeconds()); { std::lock_guard lock(mutex); @@ -603,18 +484,7 @@ BlockInputStreams StorageLiveView::watch( auto reader = std::make_shared( std::static_pointer_cast(shared_from_this()), blocks_ptr, blocks_metadata_ptr, active_ptr, has_limit, limit, - context.getSettingsRef().live_view_heartbeat_interval.totalSeconds(), - temporary_live_view_timeout); - - { - std::lock_guard no_users_thread_lock(no_users_thread_mutex); - if (no_users_thread.joinable()) - { - std::lock_guard lock(no_users_thread_wakeup_mutex); - no_users_thread_wakeup = true; - no_users_thread_condition.notify_one(); - } - } + context.getSettingsRef().live_view_heartbeat_interval.totalSeconds()); { std::lock_guard lock(mutex); diff --git a/src/Storages/LiveView/StorageLiveView.h b/src/Storages/LiveView/StorageLiveView.h index 43afd169a92..32e18ef6092 100644 --- a/src/Storages/LiveView/StorageLiveView.h +++ b/src/Storages/LiveView/StorageLiveView.h @@ -38,6 +38,10 @@ using ASTPtr = std::shared_ptr; using BlocksMetadataPtr = std::shared_ptr; using MergeableBlocksPtr = std::shared_ptr; +class Pipe; +using Pipes = std::vector; + + class StorageLiveView final : public ext::shared_ptr_helper, public IStorage { friend struct ext::shared_ptr_helper; @@ -70,7 +74,9 @@ public: NamesAndTypesList getVirtuals() const override; - bool isTemporary() { return is_temporary; } + bool isTemporary() const { return is_temporary; } + std::chrono::seconds getTimeout() const { return temporary_live_view_timeout; } + /// Check if we have any readers /// must be called with mutex locked @@ -85,11 +91,7 @@ public: { return active_ptr.use_count() > 1; } - /// No users thread mutex, predicate and wake up condition - void startNoUsersThread(const UInt64 & timeout); - std::mutex no_users_thread_wakeup_mutex; - bool no_users_thread_wakeup = false; - std::condition_variable no_users_thread_condition; + /// Get blocks hash /// must be called with mutex locked String getBlocksHashKey() @@ -175,6 +177,8 @@ private: std::unique_ptr live_view_context; bool is_temporary = false; + std::chrono::seconds temporary_live_view_timeout; + /// Mutex to protect access to sample block and inner_blocks_query mutable std::mutex sample_block_lock; mutable Block sample_block; @@ -193,14 +197,7 @@ private: std::shared_ptr blocks_metadata_ptr; MergeableBlocksPtr mergeable_blocks; - /// Background thread for temporary tables - /// which drops this table if there are no users - static void noUsersThread(std::shared_ptr storage, const UInt64 & timeout); - std::mutex no_users_thread_mutex; - std::thread no_users_thread; std::atomic shutdown_called = false; - std::atomic start_no_users_thread_called = false; - UInt64 temporary_live_view_timeout; StorageLiveView( const StorageID & table_id_, diff --git a/src/Storages/LiveView/TemporaryLiveViewCleaner.cpp b/src/Storages/LiveView/TemporaryLiveViewCleaner.cpp new file mode 100644 index 00000000000..0f7c1039d72 --- /dev/null +++ b/src/Storages/LiveView/TemporaryLiveViewCleaner.cpp @@ -0,0 +1,148 @@ +#include +#include +#include +#include +#include + + +namespace DB +{ +namespace ErrorCodes +{ + extern const int LOGICAL_ERROR; +} + + +namespace +{ + void executeDropQuery(const StorageID & storage_id, Context & context) + { + if (!DatabaseCatalog::instance().isTableExist(storage_id, context)) + return; + try + { + /// We create and execute `drop` query for this table + auto drop_query = std::make_shared(); + drop_query->database = storage_id.database_name; + drop_query->table = storage_id.table_name; + drop_query->kind = ASTDropQuery::Kind::Drop; + ASTPtr ast_drop_query = drop_query; + InterpreterDropQuery drop_interpreter(ast_drop_query, context); + drop_interpreter.execute(); + } + catch (...) + { + tryLogCurrentException(__PRETTY_FUNCTION__); + } + } +} + + +std::unique_ptr TemporaryLiveViewCleaner::the_instance; + + +void TemporaryLiveViewCleaner::init(Context & global_context_) +{ + if (the_instance) + throw Exception("TemporaryLiveViewCleaner already initialized", ErrorCodes::LOGICAL_ERROR); + the_instance.reset(new TemporaryLiveViewCleaner(global_context_)); +} + + +void TemporaryLiveViewCleaner::shutdown() +{ + the_instance.reset(); +} + + +TemporaryLiveViewCleaner::TemporaryLiveViewCleaner(Context & global_context_) + : global_context(global_context_) +{ +} + + +TemporaryLiveViewCleaner::~TemporaryLiveViewCleaner() +{ + stopBackgroundThread(); +} + + +void TemporaryLiveViewCleaner::addView(const std::shared_ptr & view) +{ + if (!view->isTemporary()) + return; + + auto current_time = std::chrono::system_clock::now(); + auto time_of_next_check = current_time + view->getTimeout(); + + std::lock_guard lock{mutex}; + + /// Keep the vector `views` sorted by time of next check. + StorageAndTimeOfCheck storage_and_time_of_check{view, time_of_next_check}; + views.insert(std::upper_bound(views.begin(), views.end(), storage_and_time_of_check), storage_and_time_of_check); + + if (!background_thread.joinable()) + background_thread = ThreadFromGlobalPool{&TemporaryLiveViewCleaner::backgroundThreadFunc, this}; + + background_thread_wake_up.notify_one(); +} + + +void TemporaryLiveViewCleaner::backgroundThreadFunc() +{ + std::unique_lock lock{mutex}; + while (!background_thread_should_exit && !views.empty()) + { + background_thread_wake_up.wait_until(lock, views.front().time_of_check); + if (background_thread_should_exit) + return; + + auto current_time = std::chrono::system_clock::now(); + std::vector storages_to_drop; + + auto it = views.begin(); + while (it != views.end()) + { + std::shared_ptr storage = it->storage.lock(); + auto & time_of_check = it->time_of_check; + if (!storage) + { + /// Storage has been already removed. + it = views.erase(it); + continue; + } + + ++it; + + if (current_time < time_of_check) + break; /// It's not the time to check it yet. + + time_of_check = current_time + storage->getTimeout(); + + auto storage_id = storage->getStorageID(); + if (storage->hasUsers() || !DatabaseCatalog::instance().getDependencies(storage_id).empty()) + continue; + + storages_to_drop.emplace_back(storage_id); + } + + lock.unlock(); + for (const auto & storage_id : storages_to_drop) + executeDropQuery(storage_id, global_context); + lock.lock(); + } +} + + +void TemporaryLiveViewCleaner::stopBackgroundThread() +{ + std::lock_guard lock{mutex}; + if (background_thread.joinable()) + { + background_thread_should_exit = true; + background_thread_wake_up.notify_one(); + background_thread.join(); + } +} + +} diff --git a/src/Storages/LiveView/TemporaryLiveViewCleaner.h b/src/Storages/LiveView/TemporaryLiveViewCleaner.h new file mode 100644 index 00000000000..57c12bd1c07 --- /dev/null +++ b/src/Storages/LiveView/TemporaryLiveViewCleaner.h @@ -0,0 +1,51 @@ +#pragma once + +#include +#include + + +namespace DB +{ +class StorageLiveView; +struct StorageID; + +/// This class removes temporary live views in the background thread when it's possible. +/// There should only a single instance of this class. +class TemporaryLiveViewCleaner +{ +public: + static TemporaryLiveViewCleaner & instance() { return *the_instance; } + + /// Drops a specified live view after a while if it's temporary. + void addView(const std::shared_ptr & view); + + /// Should be called once. + static void init(Context & global_context_); + static void shutdown(); + +private: + friend std::unique_ptr::deleter_type; + + TemporaryLiveViewCleaner(Context & global_context_); + ~TemporaryLiveViewCleaner(); + + void backgroundThreadFunc(); + void stopBackgroundThread(); + + struct StorageAndTimeOfCheck + { + std::weak_ptr storage; + std::chrono::system_clock::time_point time_of_check; + bool operator <(const StorageAndTimeOfCheck & other) const { return time_of_check < other.time_of_check; } + }; + + static std::unique_ptr the_instance; + Context & global_context; + std::mutex mutex; + std::vector views; + ThreadFromGlobalPool background_thread; + std::atomic background_thread_should_exit = false; + std::condition_variable background_thread_wake_up; +}; + +} diff --git a/src/Storages/ya.make b/src/Storages/ya.make index 1ddb8c77072..fed961ed2bb 100644 --- a/src/Storages/ya.make +++ b/src/Storages/ya.make @@ -20,6 +20,7 @@ SRCS( IStorage.cpp KeyDescription.cpp LiveView/StorageLiveView.cpp + LiveView/TemporaryLiveViewCleaner.cpp MergeTree/ActiveDataPartSet.cpp MergeTree/AllMergeSelector.cpp MergeTree/BackgroundProcessingPool.cpp From b8a2c1d2a29517c2bd0e8f791ce31c474f30f7d5 Mon Sep 17 00:00:00 2001 From: alesapin Date: Wed, 9 Sep 2020 22:45:37 +0300 Subject: [PATCH 043/121] Push pragma only for new gcc --- src/Storages/MergeTree/MergeTreePartition.cpp | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/src/Storages/MergeTree/MergeTreePartition.cpp b/src/Storages/MergeTree/MergeTreePartition.cpp index 8ef3e458871..2802b842f54 100644 --- a/src/Storages/MergeTree/MergeTreePartition.cpp +++ b/src/Storages/MergeTree/MergeTreePartition.cpp @@ -29,8 +29,10 @@ String MergeTreePartition::getID(const MergeTreeData & storage) const return getID(storage.getInMemoryMetadataPtr()->getPartitionKey().sample_block); } -#pragma GCC diagnostic push -#pragma GCC diagnostic ignored "-Wstringop-overflow" +#if defined (__GNUC__) && __GNUC__ >= 10 + #pragma GCC diagnostic push + #pragma GCC diagnostic ignored "-Wstringop-overflow" +#endif /// NOTE: This ID is used to create part names which are then persisted in ZK and as directory names on the file system. /// So if you want to change this method, be sure to guarantee compatibility with existing table data. @@ -90,7 +92,9 @@ String MergeTreePartition::getID(const Block & partition_key_sample) const return result; } -#pragma GCC diagnostic pop +#if defined (__GNUC__) && __GNUC__ >= 10 + #pragma GCC diagnostic pop +#endif void MergeTreePartition::serializeText(const MergeTreeData & storage, WriteBuffer & out, const FormatSettings & format_settings) const { From 27258c8e70213cf57e1bbf36176cda961d56e12f Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Wed, 9 Sep 2020 23:47:42 +0300 Subject: [PATCH 044/121] utils/list-licenses/list-licenses.sh: ignore more files - *.rtf They can have NULL byte, and StorageSystemLicenses.sh will warn: ./StorageSystemLicenses.sh: line 11: warning: command substitution: ignored null byte in input Found with: find contrib/ -type f -and '(' -iname 'LICENSE*' -or -iname 'COPYING*' -or -iname 'COPYRIGHT*' ')' -and -not -iname '*.html' | xargs grep -Pa '\x00' - *.h - *.cpp - *.htm And after verified with: $ find contrib/ -type f -and '(' -iname 'LICENSE*' -or -iname 'COPYING*' -or -iname 'COPYRIGHT*' ')' -and -not '(' -iname '*.html' -or -iname '*.htm' -or -iname '*.rtf' -or -name '*.cpp' -or -name '*.h' -or -iname '*.json' ')' | xargs file -b | sort -u ASCII text ASCII text, with CR line terminators ASCII text, with very long lines empty UTF-8 Unicode text --- utils/list-licenses/list-licenses.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/utils/list-licenses/list-licenses.sh b/utils/list-licenses/list-licenses.sh index 987179e26a8..8eee3f97253 100755 --- a/utils/list-licenses/list-licenses.sh +++ b/utils/list-licenses/list-licenses.sh @@ -7,7 +7,7 @@ ls -1 -d ${LIBS_PATH}/*/ | grep -F -v -- '-cmake' | while read LIB; do LIB_NAME=$(basename $LIB) LIB_LICENSE=$( - LC_ALL=C find "$LIB" -type f -and '(' -iname 'LICENSE*' -or -iname 'COPYING*' -or -iname 'COPYRIGHT*' ')' -and -not -iname '*.html' -printf "%d\t%p\n" | + LC_ALL=C find "$LIB" -type f -and '(' -iname 'LICENSE*' -or -iname 'COPYING*' -or -iname 'COPYRIGHT*' ')' -and -not '(' -iname '*.html' -or -iname '*.htm' -or -iname '*.rtf' -or -name '*.cpp' -or -name '*.h' -or -iname '*.json' ')' -printf "%d\t%p\n" | awk ' BEGIN { IGNORECASE=1; min_depth = 0 } /LICENSE/ { if (!min_depth || $1 <= min_depth) { min_depth = $1; license = $2 } } From 0f4fdcbf389909ed2e642263b0d6a65a3580d8e0 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Thu, 10 Sep 2020 02:05:41 +0300 Subject: [PATCH 045/121] Pass -fsanitize-blacklist for TSAN only under clang (gcc does not support this) And no such check for -fsnaitize=memory, since gcc does not support it anyway. --- cmake/sanitize.cmake | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/cmake/sanitize.cmake b/cmake/sanitize.cmake index 32443ed78c3..7c7e9c388a0 100644 --- a/cmake/sanitize.cmake +++ b/cmake/sanitize.cmake @@ -36,7 +36,15 @@ if (SANITIZE) endif () elseif (SANITIZE STREQUAL "thread") - set (TSAN_FLAGS "-fsanitize=thread -fsanitize-blacklist=${CMAKE_SOURCE_DIR}/tests/tsan_suppressions.txt") + set (TSAN_FLAGS "-fsanitize=thread") + if (COMPILER_CLANG) + set (TSAN_FLAGS "${TSAN_FLAGS} -fsanitize-blacklist=${CMAKE_SOURCE_DIR}/tests/tsan_suppressions.txt") + else() + message (WARNING "TSAN suppressions was not passed to the compiler (since the compiler is not clang)") + message (WARNING "Use the following command to pass them manually:") + message (WARNING " export TSAN_OPTIONS=\"$TSAN_OPTIONS suppressions=${CMAKE_SOURCE_DIR}/tests/tsan_suppressions.txt\"") + endif() + set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${SAN_FLAGS} ${TSAN_FLAGS}") set (CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${SAN_FLAGS} ${TSAN_FLAGS}") From a64473313971bbd3d461d5c7b68165b39d4515e0 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Thu, 10 Sep 2020 12:05:57 +0300 Subject: [PATCH 046/121] Attempt to make performance test more reliable --- programs/server/Server.cpp | 8 +- src/Common/remapExecutable.cpp | 213 +++++++++++++++++++++++++++++++++ src/Common/remapExecutable.h | 7 ++ 3 files changed, 227 insertions(+), 1 deletion(-) create mode 100644 src/Common/remapExecutable.cpp create mode 100644 src/Common/remapExecutable.h diff --git a/programs/server/Server.cpp b/programs/server/Server.cpp index f24ba444203..8149623ffce 100644 --- a/programs/server/Server.cpp +++ b/programs/server/Server.cpp @@ -32,6 +32,7 @@ #include #include #include +#include #include #include #include @@ -307,6 +308,11 @@ int Server::main(const std::vector & /*args*/) { if (config().getBool("mlock_executable", false)) { + LOG_DEBUG(log, "Will remap executable in memory."); + remapExecutable(); + LOG_DEBUG(log, "The code in memory has been successfully remapped."); + +/* if (hasLinuxCapability(CAP_IPC_LOCK)) { LOG_TRACE(log, "Will mlockall to prevent executable memory from being paged out. It may take a few seconds."); @@ -321,7 +327,7 @@ int Server::main(const std::vector & /*args*/) " It could happen due to incorrect ClickHouse package installation." " You could resolve the problem manually with 'sudo setcap cap_ipc_lock=+ep {}'." " Note that it will not work on 'nosuid' mounted filesystems.", executable_path); - } + }*/ } } #endif diff --git a/src/Common/remapExecutable.cpp b/src/Common/remapExecutable.cpp new file mode 100644 index 00000000000..f7f353a83c6 --- /dev/null +++ b/src/Common/remapExecutable.cpp @@ -0,0 +1,213 @@ +#if defined(__linux__) && defined(__amd64__) && defined(__SSE2__) + +#include +#include +#include + +#include + +#include + +#include +#include +#include +#include +#include + +#include "remapExecutable.h" + + +namespace DB +{ + +namespace ErrorCodes +{ + extern const int LOGICAL_ERROR; + extern const int CANNOT_ALLOCATE_MEMORY; +} + + +namespace +{ + +uintptr_t readAddressHex(DB::ReadBuffer & in) +{ + uintptr_t res = 0; + while (!in.eof()) + { + if (isHexDigit(*in.position())) + { + res *= 16; + res += unhex(*in.position()); + ++in.position(); + } + else + break; + } + return res; +} + + +/** Find the address and size of the mapped memory region pointed by ptr. + */ +std::pair getMappedArea(void * ptr) +{ + using namespace DB; + + uintptr_t uintptr = reinterpret_cast(ptr); + ReadBufferFromFile in("/proc/self/maps"); + + while (!in.eof()) + { + uintptr_t begin = readAddressHex(in); + assertChar('-', in); + uintptr_t end = readAddressHex(in); + skipToNextLineOrEOF(in); + + if (begin <= uintptr && uintptr < end) + return {reinterpret_cast(begin), end - begin}; + } + + throw Exception("Cannot find mapped area for pointer", ErrorCodes::LOGICAL_ERROR); +} + + +__attribute__((__noinline__)) int64_t our_syscall(...) +{ + __asm__ __volatile__ (R"( + movq %%rdi,%%rax; + movq %%rsi,%%rdi; + movq %%rdx,%%rsi; + movq %%rcx,%%rdx; + movq %%r8,%%r10; + movq %%r9,%%r8; + movq 8(%%rsp),%%r9; + syscall; + ret + )" : : : "memory"); + return 0; +} + + +__attribute__((__always_inline__)) void our_memcpy(char * __restrict dst, const char * __restrict src, ssize_t n) +{ + while (n > 0) + { + _mm_storeu_si128(reinterpret_cast<__m128i *>(dst), + _mm_loadu_si128(reinterpret_cast(src))); + + dst += 16; + src += 16; + n -= 16; + } +} + + +__attribute__((__noinline__)) void remapToHugeStep3(void * scratch, size_t size, size_t offset) +{ + /// The function should not use the stack, otherwise various optimizations, including "omit-frame-pointer" may break the code. + + /// Unmap the scratch area. + our_syscall(SYS_munmap, scratch, size); + + /** The return address of this function is pointing to scratch area (because it was called from there). + * But the scratch area no longer exists. We should correct the return address by subtracting the offset. + */ + __asm__ __volatile__("subq %0, 8(%%rsp)" : : "r"(offset) : "memory"); +} + + +__attribute__((__noinline__)) void remapToHugeStep2(void * begin, size_t size, void * scratch) +{ + /** Unmap old memory region with the code of our program. + * Our instruction pointer is located inside scratch area and this function can execute after old code is unmapped. + * But it cannot call any other functions because they are not available at usual addresses + * - that's why we have to use "our_syscall" and "our_memcpy" functions. + * (Relative addressing may continue to work but we should not assume that). + */ + + int64_t offset = reinterpret_cast(scratch) - reinterpret_cast(begin); + int64_t (*syscall_func)(...) = reinterpret_cast(reinterpret_cast(our_syscall) + offset); + + //char dot = '.'; + //syscall_func(SYS_write, 2, &dot, 1); + + int64_t munmap_res = syscall_func(SYS_munmap, begin, size); + if (munmap_res != 0) + return; + + //syscall_func(SYS_write, 2, &dot, 1); + + /// Map new anonymous memory region in place of old region with code. + + int64_t mmap_res = syscall_func(SYS_mmap, begin, size, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED, -1, 0); + if (-1 == mmap_res) + syscall_func(SYS_exit, 1); + //syscall_func(SYS_write, 2, &dot, 1); + + /// As the memory region is anonymous, we can do madvise with MADV_HUGEPAGE. + + syscall_func(SYS_madvise, begin, size, MADV_HUGEPAGE); + //syscall_func(SYS_write, 2, &dot, 1); + + /// Copy the code from scratch area to the old memory location. + + our_memcpy(reinterpret_cast(begin), reinterpret_cast(scratch), size); + //syscall_func(SYS_write, 2, &dot, 1); + + /// Make the memory area with the code executable and non-writable. + + syscall_func(SYS_mprotect, begin, size, PROT_READ | PROT_EXEC); + //syscall_func(SYS_write, 2, &dot, 1); + + /** Step 3 function should unmap the scratch area. + * The currently executed code is located in the scratch area and cannot be removed here. + * We have to call another function and use its address from the original location (not in scratch area). + * To do it, we obtain it's pointer and call by pointer. + */ + + void(* volatile step3)(void*, size_t, size_t) = remapToHugeStep3; + step3(scratch, size, offset); +} + + +__attribute__((__noinline__)) void remapToHugeStep1(void * begin, size_t size) +{ + /// Allocate scratch area and copy the code there. + + void * scratch = mmap(nullptr, size, PROT_READ | PROT_WRITE | PROT_EXEC, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0); + if (MAP_FAILED == scratch) + throwFromErrno(fmt::format("Cannot mmap {} bytes", size), ErrorCodes::CANNOT_ALLOCATE_MEMORY); + + memcpy(scratch, begin, size); + + /// Offset to the scratch area from previous location. + + int64_t offset = reinterpret_cast(scratch) - reinterpret_cast(begin); + + /// Jump to the next function inside the scratch area. + + reinterpret_cast(reinterpret_cast(remapToHugeStep2) + offset)(begin, size, scratch); +} + +} + + +void remapExecutable() +{ + auto [begin, size] = getMappedArea(reinterpret_cast(remapExecutable)); + remapToHugeStep1(begin, size); +} + +} + +#else + +namespace DB +{ + +void remapExecutable() {} + +} + +#endif diff --git a/src/Common/remapExecutable.h b/src/Common/remapExecutable.h new file mode 100644 index 00000000000..7acb61f13bd --- /dev/null +++ b/src/Common/remapExecutable.h @@ -0,0 +1,7 @@ +namespace DB +{ + +/// This function tries to reallocate the code of the running program in a more efficient way. +void remapExecutable(); + +} From 532d121100fc696fa2edb2d3dff863907850c218 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Thu, 10 Sep 2020 12:14:31 +0300 Subject: [PATCH 047/121] Fix typo --- src/Common/remapExecutable.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Common/remapExecutable.cpp b/src/Common/remapExecutable.cpp index f7f353a83c6..ec8b1703b0a 100644 --- a/src/Common/remapExecutable.cpp +++ b/src/Common/remapExecutable.cpp @@ -163,7 +163,7 @@ __attribute__((__noinline__)) void remapToHugeStep2(void * begin, size_t size, v /** Step 3 function should unmap the scratch area. * The currently executed code is located in the scratch area and cannot be removed here. * We have to call another function and use its address from the original location (not in scratch area). - * To do it, we obtain it's pointer and call by pointer. + * To do it, we obtain its pointer and call by pointer. */ void(* volatile step3)(void*, size_t, size_t) = remapToHugeStep3; From 65e9c678f8990756ac22ad2ad10bc08a9c0ec4b4 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Thu, 10 Sep 2020 17:47:02 +0300 Subject: [PATCH 048/121] Disable under certain conditions --- src/Common/remapExecutable.cpp | 2 +- src/Common/ya.make | 1 + 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/src/Common/remapExecutable.cpp b/src/Common/remapExecutable.cpp index ec8b1703b0a..b41fece0c79 100644 --- a/src/Common/remapExecutable.cpp +++ b/src/Common/remapExecutable.cpp @@ -1,4 +1,4 @@ -#if defined(__linux__) && defined(__amd64__) && defined(__SSE2__) +#if defined(__linux__) && defined(__amd64__) && defined(__SSE2__) && !defined(SANITIZER) && defined(NDEBUG) #include #include diff --git a/src/Common/ya.make b/src/Common/ya.make index d9a7a2ce4de..72f1fa42756 100644 --- a/src/Common/ya.make +++ b/src/Common/ya.make @@ -74,6 +74,7 @@ SRCS( QueryProfiler.cpp quoteString.cpp randomSeed.cpp + remapExecutable.cpp RemoteHostFilter.cpp renameat2.cpp RWLock.cpp From ca2a33008b291bc5d1507b568ac31d588a6aa3d8 Mon Sep 17 00:00:00 2001 From: Alexander Kuzmenkov Date: Wed, 2 Sep 2020 19:42:24 +0300 Subject: [PATCH 049/121] faster --- docker/test/performance-comparison/eqmed.sql | 4 ++-- docker/test/performance-comparison/perf.py | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/docker/test/performance-comparison/eqmed.sql b/docker/test/performance-comparison/eqmed.sql index f7f8d6ac40d..139f0758798 100644 --- a/docker/test/performance-comparison/eqmed.sql +++ b/docker/test/performance-comparison/eqmed.sql @@ -8,7 +8,7 @@ select from ( -- quantiles of randomization distributions - select quantileExactForEach(0.999)( + select quantileExactForEach(0.99)( arrayMap(x, y -> abs(x - y), metrics_by_label[1], metrics_by_label[2]) as d ) threshold ---- uncomment to see what the distribution is really like @@ -33,7 +33,7 @@ from -- strip the query away before the join -- it might be several kB long; (select metrics, run, version from table) no_query, -- duplicate input measurements into many virtual runs - numbers(1, 100000) nn + numbers(1, 10000) nn -- for each virtual run, randomly reorder measurements order by virtual_run, rand() ) virtual_runs diff --git a/docker/test/performance-comparison/perf.py b/docker/test/performance-comparison/perf.py index e1476d9aeb4..05e89c9e44c 100755 --- a/docker/test/performance-comparison/perf.py +++ b/docker/test/performance-comparison/perf.py @@ -20,7 +20,7 @@ parser = argparse.ArgumentParser(description='Run performance test.') parser.add_argument('file', metavar='FILE', type=argparse.FileType('r', encoding='utf-8'), nargs=1, help='test description file') parser.add_argument('--host', nargs='*', default=['localhost'], help="Server hostname(s). Corresponds to '--port' options.") parser.add_argument('--port', nargs='*', default=[9000], help="Server port(s). Corresponds to '--host' options.") -parser.add_argument('--runs', type=int, default=int(os.environ.get('CHPC_RUNS', 13)), help='Number of query runs per server. Defaults to CHPC_RUNS environment variable.') +parser.add_argument('--runs', type=int, default=int(os.environ.get('CHPC_RUNS', 7)), help='Number of query runs per server. Defaults to CHPC_RUNS environment variable.') parser.add_argument('--long', action='store_true', help='Do not skip the tests tagged as long.') parser.add_argument('--print-queries', action='store_true', help='Print test queries and exit.') parser.add_argument('--print-settings', action='store_true', help='Print test settings and exit.') From 26348ad0143f881c8d14e41e0c80d706614ab110 Mon Sep 17 00:00:00 2001 From: Alexander Kuzmenkov Date: Thu, 10 Sep 2020 18:48:39 +0300 Subject: [PATCH 050/121] fixup --- docker/test/performance-comparison/report.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/docker/test/performance-comparison/report.py b/docker/test/performance-comparison/report.py index 1003a6d0e1a..b3f8ef01138 100755 --- a/docker/test/performance-comparison/report.py +++ b/docker/test/performance-comparison/report.py @@ -372,7 +372,7 @@ if args.report == 'main': 'New, s', # 1 'Ratio of speedup (-) or slowdown (+)', # 2 'Relative difference (new − old) / old', # 3 - 'p < 0.001 threshold', # 4 + 'p < 0.01 threshold', # 4 # Failed # 5 'Test', # 6 '#', # 7 @@ -416,7 +416,7 @@ if args.report == 'main': 'Old, s', #0 'New, s', #1 'Relative difference (new - old)/old', #2 - 'p < 0.001 threshold', #3 + 'p < 0.01 threshold', #3 # Failed #4 'Test', #5 '#', #6 @@ -649,7 +649,7 @@ elif args.report == 'all-queries': 'New, s', #3 'Ratio of speedup (-) or slowdown (+)', #4 'Relative difference (new − old) / old', #5 - 'p < 0.001 threshold', #6 + 'p < 0.01 threshold', #6 'Test', #7 '#', #8 'Query', #9 From 45340c701dc517b29db5a1047c306f88ba891722 Mon Sep 17 00:00:00 2001 From: Alexander Kuzmenkov Date: Thu, 10 Sep 2020 19:49:57 +0300 Subject: [PATCH 051/121] changelog for 20.8 --- CHANGELOG.md | 148 ++++++++++++++++++++++ utils/simple-backport/backport.sh | 7 +- utils/simple-backport/format-changelog.py | 2 +- 3 files changed, 154 insertions(+), 3 deletions(-) diff --git a/CHANGELOG.md b/CHANGELOG.md index 345ee2c6213..f3266520eb1 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,3 +1,151 @@ +## ClickHouse release 20.8 + +### ClickHouse release v20.8.2.3-stable, 2020-09-08 + +#### Backward Incompatible Change + +* Now `OPTIMIZE FINAL` query doesn't recalculate TTL for parts that were added before TTL was created. Use `ALTER TABLE ... MATERIALIZE TTL` once to calculate them, after that `OPTIMIZE FINAL` will evaluate TTL's properly. This behavior never worked for replicated tables. [#14220](https://github.com/ClickHouse/ClickHouse/pull/14220) ([alesapin](https://github.com/alesapin)). +* Extend `parallel_distributed_insert_select` setting, adding an option to run `INSERT` into local table. The setting changes type from `Bool` to `UInt64`, so the values `false` and `true` are no longer supported. If you have these values in server configuration, the server will not start. Please replace them with `0` and `1`, respectively. [#14060](https://github.com/ClickHouse/ClickHouse/pull/14060) ([Azat Khuzhin](https://github.com/azat)). +* Remove support for the `ODBCDriver` input/output format. This was a deprecated format once used for communication with the ClickHouse ODBC driver, now long superseded by the `ODBCDriver2` format. Resolves [#13629](https://github.com/ClickHouse/ClickHouse/issues/13629). [#13847](https://github.com/ClickHouse/ClickHouse/pull/13847) ([hexiaoting](https://github.com/hexiaoting)). + +#### New Feature + +* Add `countDigits(x)` function that count number of decimal digits in integer or decimal column. Add `isDecimalOverflow(d, [p])` function that checks if the value in Decimal column is out of its (or specified) precision. [#14151](https://github.com/ClickHouse/ClickHouse/pull/14151) ([Artem Zuikov](https://github.com/4ertus2)). +* Add setting `min_index_granularity_bytes` that protects against accidentally creating a table with very low `index_granularity_bytes` setting. [#14139](https://github.com/ClickHouse/ClickHouse/pull/14139) ([Bharat Nallan](https://github.com/bharatnc)). +* Add the ability to specify `Default` compression codec for columns that correspond to settings specified in `config.xml`. Implements: [#9074](https://github.com/ClickHouse/ClickHouse/issues/9074). [#14049](https://github.com/ClickHouse/ClickHouse/pull/14049) ([alesapin](https://github.com/alesapin)). +* Added `date_trunc` function that truncates a date/time value to a specified date/time part. [#13888](https://github.com/ClickHouse/ClickHouse/pull/13888) ([Vladimir Golovchenko](https://github.com/vladimir-golovchenko)). +* Add `time_zones` table. [#13880](https://github.com/ClickHouse/ClickHouse/pull/13880) ([Bharat Nallan](https://github.com/bharatnc)). +* Add function `defaultValueOfTypeName` that returns the default value for a given type. [#13877](https://github.com/ClickHouse/ClickHouse/pull/13877) ([hcz](https://github.com/hczhcz)). +* Add `quantileExactLow` and `quantileExactHigh` implementations with respective aliases for `medianExactLow` and `medianExactHigh`. [#13818](https://github.com/ClickHouse/ClickHouse/pull/13818) ([Bharat Nallan](https://github.com/bharatnc)). +* Add function `normalizeQuery` that replaces literals, sequences of literals and complex aliases with placeholders. Add function `normalizedQueryHash` that returns identical 64bit hash values for similar queries. It helps to analyze query log. This closes [#11271](https://github.com/ClickHouse/ClickHouse/issues/11271). [#13816](https://github.com/ClickHouse/ClickHouse/pull/13816) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Add new optional section to the main config. [#13425](https://github.com/ClickHouse/ClickHouse/pull/13425) ([Vitaly Baranov](https://github.com/vitlibar)). +* Add `ALTER SAMPLE BY` statement that allows to change table sample clause. [#13280](https://github.com/ClickHouse/ClickHouse/pull/13280) ([Amos Bird](https://github.com/amosbird)). +* Function `position` now supports optional `start_pos` argument. [#13237](https://github.com/ClickHouse/ClickHouse/pull/13237) ([vdimir](https://github.com/vdimir)). +* Add types `Int128`, `Int256`, `UInt256` and related functions for them. Extend Decimals with Decimal256 (precision up to 76 digits). New types are under the setting `allow_experimental_bigint_types`. [#13097](https://github.com/ClickHouse/ClickHouse/pull/13097) ([Artem Zuikov](https://github.com/4ertus2)). +* Support Kerberos authentication in Kafka, using `krb5` and `cyrus-sasl` libraries. [#12771](https://github.com/ClickHouse/ClickHouse/pull/12771) ([Ilya Golshtein](https://github.com/ilejn)). +* Support `MaterializeMySQL` database engine. Implements [#4006](https://github.com/ClickHouse/ClickHouse/issues/4006). [#10851](https://github.com/ClickHouse/ClickHouse/pull/10851) ([Winter Zhang](https://github.com/zhang2014)). + +#### Bug Fix + +* Check for array size overflow in `topK` aggregate function. Without this check the user may send a query with carefully crafter parameters that will lead to server crash. This closes [#14452](https://github.com/ClickHouse/ClickHouse/issues/14452). [#14467](https://github.com/ClickHouse/ClickHouse/pull/14467) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Fix bug which leads to wrong merges assignment if table has partitions with a single part. [#14444](https://github.com/ClickHouse/ClickHouse/pull/14444) ([alesapin](https://github.com/alesapin)). +* Stop query execution if exception happened in `PipelineExecutor` itself. This could prevent rare possible query hung. Continuation of [#14334](https://github.com/ClickHouse/ClickHouse/issues/14334). [#14402](https://github.com/ClickHouse/ClickHouse/pull/14402) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Stop query execution if exception happened in `PipelineExecutor` itself. This could prevent rare possible query hung. [#14334](https://github.com/ClickHouse/ClickHouse/pull/14334) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Fix crash during `ALTER` query for table which was created `AS table_function`. Fixes [#14212](https://github.com/ClickHouse/ClickHouse/issues/14212). [#14326](https://github.com/ClickHouse/ClickHouse/pull/14326) ([alesapin](https://github.com/alesapin)). +* Fix exception during ALTER LIVE VIEW query with REFRESH command. [#14320](https://github.com/ClickHouse/ClickHouse/pull/14320) ([Bharat Nallan](https://github.com/bharatnc)). +* Fix QueryPlan lifetime (for EXPLAIN PIPELINE graph=1) for queries with nested interpreter. [#14315](https://github.com/ClickHouse/ClickHouse/pull/14315) ([Azat Khuzhin](https://github.com/azat)). +* Fix segfault in `clickhouse-odbc-bridge` during schema fetch from some external sources. This PR fixes https://github.com/ClickHouse/ClickHouse/issues/13861. [#14267](https://github.com/ClickHouse/ClickHouse/pull/14267) ([Vitaly Baranov](https://github.com/vitlibar)). +* Disallows `CODEC` on `ALIAS` column type. Fixes [#13911](https://github.com/ClickHouse/ClickHouse/issues/13911). [#14263](https://github.com/ClickHouse/ClickHouse/pull/14263) ([Bharat Nallan](https://github.com/bharatnc)). +* Fix handling of empty transactions in `MaterializeMySQL` database engine. This fixes [#14235](https://github.com/ClickHouse/ClickHouse/issues/14235). [#14253](https://github.com/ClickHouse/ClickHouse/pull/14253) ([BohuTANG](https://github.com/BohuTANG)). +* fixes [#14231](https://github.com/ClickHouse/ClickHouse/issues/14231) fix wrong lexer in MaterializeMySQL database engine dump stage. [#14232](https://github.com/ClickHouse/ClickHouse/pull/14232) ([Winter Zhang](https://github.com/zhang2014)). +* Fix crash in mark inclusion search introduced in https://github.com/ClickHouse/ClickHouse/pull/12277 . [#14225](https://github.com/ClickHouse/ClickHouse/pull/14225) ([Amos Bird](https://github.com/amosbird)). +* Fixed incorrect sorting order if LowCardinality column. This fixes [#13958](https://github.com/ClickHouse/ClickHouse/issues/13958). [#14223](https://github.com/ClickHouse/ClickHouse/pull/14223) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Fix creation of tables with named tuples. This fixes [#13027](https://github.com/ClickHouse/ClickHouse/issues/13027). [#14143](https://github.com/ClickHouse/ClickHouse/pull/14143) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Fix formatting of minimal negative decimal numbers. This fixes https://github.com/ClickHouse/ClickHouse/issues/14111. [#14119](https://github.com/ClickHouse/ClickHouse/pull/14119) ([Alexander Kuzmenkov](https://github.com/akuzm)). +* When waiting for a dictionary update to complete, use the timeout specified by `query_wait_timeout_milliseconds` setting instead of a hard-coded value. [#14105](https://github.com/ClickHouse/ClickHouse/pull/14105) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Fix DistributedFilesToInsert metric (zeroed when it should not). [#14095](https://github.com/ClickHouse/ClickHouse/pull/14095) ([Azat Khuzhin](https://github.com/azat)). +* Fix pointInPolygon with const 2d array as polygon. [#14079](https://github.com/ClickHouse/ClickHouse/pull/14079) ([Alexey Ilyukhov](https://github.com/livace)). +* Fixed wrong mount point in extra info for `Poco::Exception: no space left on device`. [#14050](https://github.com/ClickHouse/ClickHouse/pull/14050) ([tavplubix](https://github.com/tavplubix)). +* Fix GRANT ALL statement when executed on a non-global level. [#13987](https://github.com/ClickHouse/ClickHouse/pull/13987) ([Vitaly Baranov](https://github.com/vitlibar)). +* Fix parser to reject create table as table function with engine. [#13940](https://github.com/ClickHouse/ClickHouse/pull/13940) ([hcz](https://github.com/hczhcz)). +* Fix wrong results in select queries with `DISTINCT` keyword in case `optimize_duplicate_order_by_and_distinct` setting is enabled. [#13925](https://github.com/ClickHouse/ClickHouse/pull/13925) ([Artem Zuikov](https://github.com/4ertus2)). +* Fixed potential deadlock when renaming `Distributed` table. [#13922](https://github.com/ClickHouse/ClickHouse/pull/13922) ([tavplubix](https://github.com/tavplubix)). +* Fix incorrect sorting for `FixedString` columns. Fixes [#13182](https://github.com/ClickHouse/ClickHouse/issues/13182). [#13887](https://github.com/ClickHouse/ClickHouse/pull/13887) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Fix topK/topKWeighted merge (with non-default parameters). [#13817](https://github.com/ClickHouse/ClickHouse/pull/13817) ([Azat Khuzhin](https://github.com/azat)). +* Fix reading from MergeTree table with INDEX of type SET fails when comparing against NULL. This fixes [#13686](https://github.com/ClickHouse/ClickHouse/issues/13686). [#13793](https://github.com/ClickHouse/ClickHouse/pull/13793) ([Amos Bird](https://github.com/amosbird)). +* Fix arrayJoin() capturing in lambda (LOGICAL_ERROR). [#13792](https://github.com/ClickHouse/ClickHouse/pull/13792) ([Azat Khuzhin](https://github.com/azat)). +* Fix step overflow in range(). [#13790](https://github.com/ClickHouse/ClickHouse/pull/13790) ([Azat Khuzhin](https://github.com/azat)). +* Fixed `Directory not empty` error when concurrently executing `DROP DATABASE` and `CREATE TABLE`. [#13756](https://github.com/ClickHouse/ClickHouse/pull/13756) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Add range check for h3KRing function. This fixes [#13633](https://github.com/ClickHouse/ClickHouse/issues/13633). [#13752](https://github.com/ClickHouse/ClickHouse/pull/13752) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Fix race condition between DETACH and background merges. Parts may revive after detach. This is continuation of [#8602](https://github.com/ClickHouse/ClickHouse/issues/8602) that did not fix the issue but introduced a test that started to fail in very rare cases, demonstrating the issue. [#13746](https://github.com/ClickHouse/ClickHouse/pull/13746) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Fix logging Settings.Names/Values when log_queries_min_type > QUERY_START. [#13737](https://github.com/ClickHouse/ClickHouse/pull/13737) ([Azat Khuzhin](https://github.com/azat)). +* Fixes /replicas_status endpoint response status code when verbose=1. [#13722](https://github.com/ClickHouse/ClickHouse/pull/13722) ([javi santana](https://github.com/javisantana)). +* Fix incorrect message in `clickhouse-server.init` while checking user and group. [#13711](https://github.com/ClickHouse/ClickHouse/pull/13711) ([ylchou](https://github.com/ylchou)). +* Fix visible data clobbering by progress bar in client in interactive mode. This fixes [#12562](https://github.com/ClickHouse/ClickHouse/issues/12562) and [#13369](https://github.com/ClickHouse/ClickHouse/issues/13369) and [#13584](https://github.com/ClickHouse/ClickHouse/issues/13584) and fixes [#12964](https://github.com/ClickHouse/ClickHouse/issues/12964). [#13691](https://github.com/ClickHouse/ClickHouse/pull/13691) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Do not optimize any(arrayJoin()) -> arrayJoin() under optimize_move_functions_out_of_any. [#13681](https://github.com/ClickHouse/ClickHouse/pull/13681) ([Azat Khuzhin](https://github.com/azat)). +* Fix crash in JOIN with StorageMerge and `set enable_optimize_predicate_expression=1`. [#13679](https://github.com/ClickHouse/ClickHouse/pull/13679) ([Artem Zuikov](https://github.com/4ertus2)). +* Fix typo in error message about `The value of 'number_of_free_entries_in_pool_to_lower_max_size_of_merge' setting`. [#13678](https://github.com/ClickHouse/ClickHouse/pull/13678) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Concurrent `ALTER ... REPLACE/MOVE PARTITION ...` queries might cause deadlock. It's fixed. [#13626](https://github.com/ClickHouse/ClickHouse/pull/13626) ([tavplubix](https://github.com/tavplubix)). +* Fixed the behaviour when sometimes cache-dictionary returned default value instead of present value from source. [#13624](https://github.com/ClickHouse/ClickHouse/pull/13624) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Fix secondary indices corruption in compact parts. [#13538](https://github.com/ClickHouse/ClickHouse/pull/13538) ([Anton Popov](https://github.com/CurtizJ)). +* Fix premature `ON CLUSTER` timeouts for queries that must be executed on a single replica. Fixes [#6704](https://github.com/ClickHouse/ClickHouse/issues/6704), [#7228](https://github.com/ClickHouse/ClickHouse/issues/7228), [#13361](https://github.com/ClickHouse/ClickHouse/issues/13361), [#11884](https://github.com/ClickHouse/ClickHouse/issues/11884). [#13450](https://github.com/ClickHouse/ClickHouse/pull/13450) ([alesapin](https://github.com/alesapin)). +* Fix wrong code in function `netloc`. This fixes [#13335](https://github.com/ClickHouse/ClickHouse/issues/13335). [#13446](https://github.com/ClickHouse/ClickHouse/pull/13446) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Fix possible race in `StorageMemory`. https://clickhouse-test-reports.s3.yandex.net/0/9cac8a7244063d2092ad25d45502611e18d3749c/stress_test_(thread)/stderr.log Have no idea how to write a test. [#13416](https://github.com/ClickHouse/ClickHouse/pull/13416) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Fix missing or excessive headers in `TSV/CSVWithNames` formats. This fixes [#12504](https://github.com/ClickHouse/ClickHouse/issues/12504). [#13343](https://github.com/ClickHouse/ClickHouse/pull/13343) ([Azat Khuzhin](https://github.com/azat)). +* Fix parsing row policies from users.xml when names of databases or tables contain dots. This fixes https://github.com/ClickHouse/ClickHouse/issues/5779, https://github.com/ClickHouse/ClickHouse/issues/12527. [#13199](https://github.com/ClickHouse/ClickHouse/pull/13199) ([Vitaly Baranov](https://github.com/vitlibar)). +* Fix access to redis dictionary after connection was dropped once. It may happen with `cache` and `direct` dictionary layouts. [#13082](https://github.com/ClickHouse/ClickHouse/pull/13082) ([Anton Popov](https://github.com/CurtizJ)). +* Removed wrong auth access check when using ClickHouseDictionarySource to query remote tables. [#12756](https://github.com/ClickHouse/ClickHouse/pull/12756) ([sundyli](https://github.com/sundy-li)). +* subquery hash values are not enough to distinguish. https://github.com/ClickHouse/ClickHouse/issues/8333. [#8367](https://github.com/ClickHouse/ClickHouse/pull/8367) ([Amos Bird](https://github.com/amosbird)). + +#### Improvement + +* Now it's possible to `ALTER TABLE table_name FETCH PARTITION partition_expr FROM 'zk://:/path-in-zookeeper'`. It's useful for shipping data to new clusters. [#14155](https://github.com/ClickHouse/ClickHouse/pull/14155) ([Amos Bird](https://github.com/amosbird)). +* Slightly better performance of Memory table if it was constructed from a huge number of very small blocks (that's unlikely). Author of the idea: [Mark Papadakis](https://github.com/markpapadakis). Closes [#14043](https://github.com/ClickHouse/ClickHouse/issues/14043). [#14056](https://github.com/ClickHouse/ClickHouse/pull/14056) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Conditional aggregate functions (for example: `avgIf`, `sumIf`, `maxIf`) should return `NULL` when miss rows and use nullable arguments. [#13964](https://github.com/ClickHouse/ClickHouse/pull/13964) ([Winter Zhang](https://github.com/zhang2014)). +* Increase limit in -Resample combinator to 1M. [#13947](https://github.com/ClickHouse/ClickHouse/pull/13947) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Corrected an error in AvroConfluent format that caused the Kafka table engine to stop processing messages when an abnormally small, malformed, message was received. [#13941](https://github.com/ClickHouse/ClickHouse/pull/13941) ([Gervasio Varela](https://github.com/gervarela)). +* Fix wrong error for long queries. It was possible to get syntax error other than `Max query size exceeded` for correct query. [#13928](https://github.com/ClickHouse/ClickHouse/pull/13928) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Better error message for null value of TabSeparatedRow format. [#13906](https://github.com/ClickHouse/ClickHouse/pull/13906) ([jiang tao](https://github.com/tomjiang1987)). +* Function `arrayCompact` will compare NaNs bitwise if the type of array elements is Float32/Float64. In previous versions NaNs were always not equal if the type of array elements is Float32/Float64 and were always equal if the type is more complex, like Nullable(Float64). This closes [#13857](https://github.com/ClickHouse/ClickHouse/issues/13857). [#13868](https://github.com/ClickHouse/ClickHouse/pull/13868) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Fix data race in `lgamma` function. This race was caught only in `tsan`, no side effects a really happened. [#13842](https://github.com/ClickHouse/ClickHouse/pull/13842) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* 1. Add [GTID-Based Replication](https://dev.mysql.com/doc/refman/5.7/en/replication-gtids-concepts.html), it works even when replication topology changes, and supported/prefered in MySQL 5.6/5.7/8.0 2. Add BIT/SET filed type supports 3. Fix up varchar type meta length bug. [#13820](https://github.com/ClickHouse/ClickHouse/pull/13820) ([BohuTANG](https://github.com/BohuTANG)). +* Avoid too slow queries when arrays are manipulated as fields. Throw exception instead. [#13753](https://github.com/ClickHouse/ClickHouse/pull/13753) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Added Redis requirepass authorization. [#13688](https://github.com/ClickHouse/ClickHouse/pull/13688) ([Ivan Torgashov](https://github.com/it1804)). +* Add MergeTree Write-Ahead-Log(WAL) dump tool. [#13640](https://github.com/ClickHouse/ClickHouse/pull/13640) ([BohuTANG](https://github.com/BohuTANG)). +* In previous versions `lcm` function may produce assertion violation in debug build if called with specifically crafted arguments. This fixes [#13368](https://github.com/ClickHouse/ClickHouse/issues/13368). [#13510](https://github.com/ClickHouse/ClickHouse/pull/13510) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Provide monotonicity for `toDate/toDateTime` functions in more cases. Now the input arguments are saturated more naturally and provides better monotonicity. [#13497](https://github.com/ClickHouse/ClickHouse/pull/13497) ([Amos Bird](https://github.com/amosbird)). +* Support compound identifiers for custom settings. [#13496](https://github.com/ClickHouse/ClickHouse/pull/13496) ([Vitaly Baranov](https://github.com/vitlibar)). +* Move parts from DIskLocal to DiskS3 in parallel. [#13459](https://github.com/ClickHouse/ClickHouse/pull/13459) ([Pavel Kovalenko](https://github.com/Jokser)). +* Enable mixed granularity parts by default. [#13449](https://github.com/ClickHouse/ClickHouse/pull/13449) ([alesapin](https://github.com/alesapin)). +* Proper remote host checking in S3 redirects (security-related thing). [#13404](https://github.com/ClickHouse/ClickHouse/pull/13404) ([Vladimir Chebotarev](https://github.com/excitoon)). +* Add QueryTimeMicroseconds, SelectQueryTimeMicroseconds and InsertQueryTimeMicroseconds to system.events. [#13336](https://github.com/ClickHouse/ClickHouse/pull/13336) ([ianton-ru](https://github.com/ianton-ru)). +* Fix assert when decimal has too large negative exponent. Fixes [#13188](https://github.com/ClickHouse/ClickHouse/issues/13188). [#13228](https://github.com/ClickHouse/ClickHouse/pull/13228) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Added cache layer for DiskS3 (cache to local disk mark and index files). [#13076](https://github.com/ClickHouse/ClickHouse/pull/13076) ([Pavel Kovalenko](https://github.com/Jokser)). + +#### Performance Improvement + +* Slightly optimize very short queries with LowCardinality. [#14129](https://github.com/ClickHouse/ClickHouse/pull/14129) ([Anton Popov](https://github.com/CurtizJ)). +* Enable parallel INSERTs for table engines `Null`, `Memory`, `Distributed` and `Buffer`. [#14120](https://github.com/ClickHouse/ClickHouse/pull/14120) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Fail fast if `max_rows_to_read` limit is exceeded on parts scan. The motivation behind this change is to skip ranges scan for all selected parts if it is clear that `max_rows_to_read` is already exceeded. The change is quite noticeable for queries over big number of parts. [#13677](https://github.com/ClickHouse/ClickHouse/pull/13677) ([Roman Khavronenko](https://github.com/hagen1778)). +* Slightly improve performance of aggregation by UInt8/UInt16 keys. [#13099](https://github.com/ClickHouse/ClickHouse/pull/13099) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Optimize `has()`, `indexOf()` and `countEqual()` functions for `Array(LowCardinality(T))` and constant right arguments. [#12550](https://github.com/ClickHouse/ClickHouse/pull/12550) ([myrrc](https://github.com/myrrc)). +* When performing trivial `INSERT SELECT` queries, automatically set `max_threads` to 1 or `max_insert_threads`, and set `max_block_size` to `min_insert_block_size_rows`. Related to [#5907](https://github.com/ClickHouse/ClickHouse/issues/5907). [#12195](https://github.com/ClickHouse/ClickHouse/pull/12195) ([flynn](https://github.com/ucasFL)). + +#### Build/Testing/Packaging Improvement + +* Actually there are no symlinks there, so `-type f` is enough ``` ~/workspace/ClickHouse/contrib/cctz/testdata/zoneinfo$ find . -type l -ls | wc -l 0 ``` Closes [#14209](https://github.com/ClickHouse/ClickHouse/issues/14209). [#14215](https://github.com/ClickHouse/ClickHouse/pull/14215) ([filimonov](https://github.com/filimonov)). +* Switch tests docker images to use test-base parent. [#14167](https://github.com/ClickHouse/ClickHouse/pull/14167) ([Ilya Yatsishin](https://github.com/qoega)). +* Add the ability to write js-style comments in skip_list.json. [#14159](https://github.com/ClickHouse/ClickHouse/pull/14159) ([alesapin](https://github.com/alesapin)). +* * Adding retry logic when bringing up docker-compose cluster * Increasing COMPOSE_HTTP_TIMEOUT. [#14112](https://github.com/ClickHouse/ClickHouse/pull/14112) ([vzakaznikov](https://github.com/vzakaznikov)). +* Enabled text-log in stress test to find more bugs. [#13855](https://github.com/ClickHouse/ClickHouse/pull/13855) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Testflows LDAP module: adding missing certificates and dhparam.pem for openldap4. [#13780](https://github.com/ClickHouse/ClickHouse/pull/13780) ([vzakaznikov](https://github.com/vzakaznikov)). +* ZooKeeper cannot work reliably in unit tests in CI infrastructure. Using unit tests for ZooKeeper interaction with real ZooKeeper is bad idea from the start (unit tests are not supposed to verify complex distributed systems). We already using integration tests for this purpose and they are better suited. [#13745](https://github.com/ClickHouse/ClickHouse/pull/13745) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Added docker image for style check. Added style check that all docker and docker compose files are located in docker directory. [#13724](https://github.com/ClickHouse/ClickHouse/pull/13724) ([Ilya Yatsishin](https://github.com/qoega)). +* FIx cassandra build on Mac OS. [#13708](https://github.com/ClickHouse/ClickHouse/pull/13708) ([Ilya Yatsishin](https://github.com/qoega)). +* Fix link error in shared build. [#13700](https://github.com/ClickHouse/ClickHouse/pull/13700) ([Amos Bird](https://github.com/amosbird)). +* Add a CMake option to fail configuration instead of auto-reconfiguration, enabled by default. [#13687](https://github.com/ClickHouse/ClickHouse/pull/13687) ([Konstantin](https://github.com/podshumok)). +* Updating LDAP user authentication suite to check that it works with RBAC. [#13656](https://github.com/ClickHouse/ClickHouse/pull/13656) ([vzakaznikov](https://github.com/vzakaznikov)). +* Expose version of embedded tzdata via TZDATA_VERSION in system.build_options. [#13648](https://github.com/ClickHouse/ClickHouse/pull/13648) ([filimonov](https://github.com/filimonov)). +* Removed `-DENABLE_CURL_CLIENT` for `contrib/aws`. [#13628](https://github.com/ClickHouse/ClickHouse/pull/13628) ([Vladimir Chebotarev](https://github.com/excitoon)). +* Build ClickHouse with the most fresh tzdata from package repository. [#13623](https://github.com/ClickHouse/ClickHouse/pull/13623) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Increasing health-check timeouts for ClickHouse nodes and adding support to dump docker-compose logs if unhealthy containers found. [#13612](https://github.com/ClickHouse/ClickHouse/pull/13612) ([vzakaznikov](https://github.com/vzakaznikov)). +* Make sure https://github.com/ClickHouse/ClickHouse/issues/10977 is invalid. [#13539](https://github.com/ClickHouse/ClickHouse/pull/13539) ([Amos Bird](https://github.com/amosbird)). +* Enable Shellcheck in CI as a linter of .sh tests. This closes [#13168](https://github.com/ClickHouse/ClickHouse/issues/13168). [#13530](https://github.com/ClickHouse/ClickHouse/pull/13530) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Fix the remaining shellcheck notices. A preparation to enable Shellcheck. [#13529](https://github.com/ClickHouse/ClickHouse/pull/13529) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Added `clickhouse install` script, that is useful if you only have a single binary. [#13528](https://github.com/ClickHouse/ClickHouse/pull/13528) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Allow to run `clickhouse` binary without configuration. [#13515](https://github.com/ClickHouse/ClickHouse/pull/13515) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Ensure that there is no copy-pasted GPL code. [#13514](https://github.com/ClickHouse/ClickHouse/pull/13514) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Enable check for typos in code with `codespell`. [#13513](https://github.com/ClickHouse/ClickHouse/pull/13513) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Fix typos in code with codespell. [#13511](https://github.com/ClickHouse/ClickHouse/pull/13511) ([alexey-milovidov](https://github.com/alexey-milovidov)). +* Skip PR's from robot-clickhouse. [#13489](https://github.com/ClickHouse/ClickHouse/pull/13489) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Move Dockerfiles from integration tests to `docker/test` directory. docker_compose files are available in `runner` docker container. Docker images are built in CI and not in integration tests. [#13448](https://github.com/ClickHouse/ClickHouse/pull/13448) ([Ilya Yatsishin](https://github.com/qoega)). + +#### Other + +* Create `system` database with `Atomic` engine by default. [#13680](https://github.com/ClickHouse/ClickHouse/pull/13680) ([tavplubix](https://github.com/tavplubix)). +* Fix readline so it dumps history to file now. [#13600](https://github.com/ClickHouse/ClickHouse/pull/13600) ([Amos Bird](https://github.com/amosbird)). + + ## ClickHouse release 20.7 ### ClickHouse release v20.7.2.30-stable, 2020-08-31 diff --git a/utils/simple-backport/backport.sh b/utils/simple-backport/backport.sh index 71920304d56..64f8e6004bf 100755 --- a/utils/simple-backport/backport.sh +++ b/utils/simple-backport/backport.sh @@ -4,7 +4,10 @@ set -e branch="$1" merge_base=$(git merge-base origin/master "origin/$branch") master_git_cmd=(git log "$merge_base..origin/master" --first-parent) -branch_git_cmd=(git log "$merge_base..origin/$branch" --first-parent) +# The history in back branches shouldn't be too crazy, and sometimes we have a PR +# that merges several backport commits there (3f2cba6824fddf31c30bde8c6f4f860572f4f580), +# so don't use --first-parent +branch_git_cmd=(git log "$merge_base..origin/$branch") # Make lists of PRs that were merged into each branch. Use first parent here, or else # we'll get weird things like seeing older master that was merged into a PR branch @@ -30,7 +33,7 @@ fi # Search for PR numbers in commit messages. First variant is normal merge, and second # variant is squashed. Next are some backport message variants. find_prs=(sed -n "s/^.*merg[eding]*.*#\([[:digit:]]\+\).*$/\1/Ip; - s/^.*(#\([[:digit:]]\+\))$/\1/p; + s/^.*#\([[:digit:]]\+\))$/\1/p; s/^.*back[- ]*port[ed of]*.*#\([[:digit:]]\+\).*$/\1/Ip; s/^.*cherry[- ]*pick[ed of]*.*#\([[:digit:]]\+\).*$/\1/Ip") diff --git a/utils/simple-backport/format-changelog.py b/utils/simple-backport/format-changelog.py index ccda88c6809..5dff4f1c5e8 100755 --- a/utils/simple-backport/format-changelog.py +++ b/utils/simple-backport/format-changelog.py @@ -93,7 +93,7 @@ for line in args.file: # Normalize category name for c in categories_preferred_order: - if fuzzywuzzy.fuzz.ratio(pr['category'], c) >= 90: + if fuzzywuzzy.fuzz.ratio(pr['category'].lower(), c.lower()) >= 90: pr['category'] = c break From 5675efbd47fde50524463a14758c672091264897 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Thu, 10 Sep 2020 20:16:12 +0300 Subject: [PATCH 052/121] Fix build --- src/Common/remapExecutable.cpp | 30 ++++++++++++++---------------- 1 file changed, 14 insertions(+), 16 deletions(-) diff --git a/src/Common/remapExecutable.cpp b/src/Common/remapExecutable.cpp index b41fece0c79..6b86e8717a4 100644 --- a/src/Common/remapExecutable.cpp +++ b/src/Common/remapExecutable.cpp @@ -89,20 +89,6 @@ __attribute__((__noinline__)) int64_t our_syscall(...) } -__attribute__((__always_inline__)) void our_memcpy(char * __restrict dst, const char * __restrict src, ssize_t n) -{ - while (n > 0) - { - _mm_storeu_si128(reinterpret_cast<__m128i *>(dst), - _mm_loadu_si128(reinterpret_cast(src))); - - dst += 16; - src += 16; - n -= 16; - } -} - - __attribute__((__noinline__)) void remapToHugeStep3(void * scratch, size_t size, size_t offset) { /// The function should not use the stack, otherwise various optimizations, including "omit-frame-pointer" may break the code. @@ -122,7 +108,7 @@ __attribute__((__noinline__)) void remapToHugeStep2(void * begin, size_t size, v /** Unmap old memory region with the code of our program. * Our instruction pointer is located inside scratch area and this function can execute after old code is unmapped. * But it cannot call any other functions because they are not available at usual addresses - * - that's why we have to use "our_syscall" and "our_memcpy" functions. + * - that's why we have to use "our_syscall" function and a substitution for memcpy. * (Relative addressing may continue to work but we should not assume that). */ @@ -152,7 +138,19 @@ __attribute__((__noinline__)) void remapToHugeStep2(void * begin, size_t size, v /// Copy the code from scratch area to the old memory location. - our_memcpy(reinterpret_cast(begin), reinterpret_cast(scratch), size); + { + __m128i * __restrict dst = reinterpret_cast<__m128i *>(begin); + const __m128i * __restrict src = reinterpret_cast(scratch); + const __m128i * __restrict src_end = reinterpret_cast(reinterpret_cast(scratch) + size); + while (src < src_end) + { + _mm_storeu_si128(dst, _mm_loadu_si128(src)); + + ++dst; + ++src; + } + } + //syscall_func(SYS_write, 2, &dot, 1); /// Make the memory area with the code executable and non-writable. From f2a5216e97f1283b373720717f1d6f7ac79af86d Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Fri, 11 Sep 2020 02:24:16 +0300 Subject: [PATCH 053/121] add waiting for fsync in WAL --- src/Common/FileSyncGuard.h | 2 +- src/Storages/MergeTree/MergeTreeSettings.h | 1 + src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp | 14 +++++++++----- src/Storages/MergeTree/MergeTreeWriteAheadLog.h | 5 +++-- utils/durability-test/durability-test.sh | 12 ++++++++++++ 5 files changed, 26 insertions(+), 8 deletions(-) diff --git a/src/Common/FileSyncGuard.h b/src/Common/FileSyncGuard.h index 5ec9b1d0c98..6451f6ebf36 100644 --- a/src/Common/FileSyncGuard.h +++ b/src/Common/FileSyncGuard.h @@ -17,7 +17,7 @@ public: FileSyncGuard(const DiskPtr & disk_, int fd_) : disk(disk_), fd(fd_) {} FileSyncGuard(const DiskPtr & disk_, const String & path) - : disk(disk_), fd(disk_->open(path, O_RDONLY)) {} + : disk(disk_), fd(disk_->open(path, O_RDWR)) {} ~FileSyncGuard() { diff --git a/src/Storages/MergeTree/MergeTreeSettings.h b/src/Storages/MergeTree/MergeTreeSettings.h index 3f8f44dc11e..8652a6ef691 100644 --- a/src/Storages/MergeTree/MergeTreeSettings.h +++ b/src/Storages/MergeTree/MergeTreeSettings.h @@ -47,6 +47,7 @@ struct Settings; M(Bool, fsync_part_directory, false, "Do fsync for part directory after all part operations (writes, renames, etc.).", 0) \ M(UInt64, write_ahead_log_bytes_to_fsync, 100ULL * 1024 * 1024, "Amount of bytes, accumulated in WAL to do fsync.", 0) \ M(UInt64, write_ahead_log_interval_ms_to_fsync, 100, "Interval in milliseconds after which fsync for WAL is being done.", 0) \ + M(Bool, in_memory_parts_insert_sync, false, "If true insert of part with in-memory format will wait for fsync of WAL", 0) \ \ /** Inserts settings. */ \ M(UInt64, parts_to_delay_insert, 150, "If table contains at least that many active parts in single partition, artificially slow down insert into table.", 0) \ diff --git a/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp b/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp index 5cfe9017248..bc6738a8321 100644 --- a/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp +++ b/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp @@ -33,6 +33,7 @@ MergeTreeWriteAheadLog::MergeTreeWriteAheadLog( std::lock_guard lock(write_mutex); out->sync(); sync_scheduled = false; + sync_cv.notify_all(); }); } @@ -50,7 +51,7 @@ void MergeTreeWriteAheadLog::init() void MergeTreeWriteAheadLog::addPart(const Block & block, const String & part_name) { - std::lock_guard lock(write_mutex); + std::unique_lock lock(write_mutex); auto part_info = MergeTreePartInfo::fromPartName(part_name, storage.format_version); min_block_number = std::min(min_block_number, part_info.min_block); @@ -70,7 +71,7 @@ void MergeTreeWriteAheadLog::addPart(const Block & block, const String & part_na void MergeTreeWriteAheadLog::dropPart(const String & part_name) { - std::lock_guard lock(write_mutex); + std::unique_lock lock(write_mutex); writeIntBinary(static_cast(0), *out); writeIntBinary(static_cast(ActionType::DROP_PART), *out); @@ -78,7 +79,7 @@ void MergeTreeWriteAheadLog::dropPart(const String & part_name) sync(lock); } -void MergeTreeWriteAheadLog::rotate(const std::lock_guard &) +void MergeTreeWriteAheadLog::rotate(const std::unique_lock &) { String new_name = String(WAL_FILE_NAME) + "_" + toString(min_block_number) + "_" @@ -90,7 +91,7 @@ void MergeTreeWriteAheadLog::rotate(const std::lock_guard &) MergeTreeData::MutableDataPartsVector MergeTreeWriteAheadLog::restore(const StorageMetadataPtr & metadata_snapshot) { - std::lock_guard lock(write_mutex); + std::unique_lock lock(write_mutex); MergeTreeData::MutableDataPartsVector parts; auto in = disk->readFile(path, DBMS_DEFAULT_BUFFER_SIZE); @@ -185,7 +186,7 @@ MergeTreeData::MutableDataPartsVector MergeTreeWriteAheadLog::restore(const Stor return result; } -void MergeTreeWriteAheadLog::sync(const std::lock_guard &) +void MergeTreeWriteAheadLog::sync(std::unique_lock & lock) { size_t bytes_to_sync = storage.getSettings()->write_ahead_log_bytes_to_fsync; time_t time_to_sync = storage.getSettings()->write_ahead_log_interval_ms_to_fsync; @@ -201,6 +202,9 @@ void MergeTreeWriteAheadLog::sync(const std::lock_guard &) sync_task->scheduleAfter(time_to_sync); sync_scheduled = true; } + + if (storage.getSettings()->in_memory_parts_insert_sync) + sync_cv.wait(lock, [this] { return !sync_scheduled; }); } std::optional diff --git a/src/Storages/MergeTree/MergeTreeWriteAheadLog.h b/src/Storages/MergeTree/MergeTreeWriteAheadLog.h index 43abf3c04be..c5675eac696 100644 --- a/src/Storages/MergeTree/MergeTreeWriteAheadLog.h +++ b/src/Storages/MergeTree/MergeTreeWriteAheadLog.h @@ -44,8 +44,8 @@ public: private: void init(); - void rotate(const std::lock_guard & lock); - void sync(const std::lock_guard & lock); + void rotate(const std::unique_lock & lock); + void sync(std::unique_lock & lock); const MergeTreeData & storage; DiskPtr disk; @@ -60,6 +60,7 @@ private: BackgroundSchedulePool & pool; BackgroundSchedulePoolTaskHolder sync_task; + std::condition_variable sync_cv; size_t bytes_at_last_sync = 0; bool sync_scheduled = false; diff --git a/utils/durability-test/durability-test.sh b/utils/durability-test/durability-test.sh index c7f8936ec95..97c39473b69 100755 --- a/utils/durability-test/durability-test.sh +++ b/utils/durability-test/durability-test.sh @@ -1,5 +1,17 @@ #!/bin/bash +: ' +A simple test for durability. It starts up clickhouse server in qemu VM and runs +inserts via clickhouse benchmark tool. Then it kills VM in random moment and +checks whether table contains broken parts. With enabled fsync no broken parts +should be appeared. + +Usage: + +./install.sh +./durability-test.sh +' + URL=http://cloud-images.ubuntu.com/bionic/current IMAGE=bionic-server-cloudimg-amd64.img SSH_PORT=11022 From da2bb4e0d3d2e642993f070b923401a9db470d81 Mon Sep 17 00:00:00 2001 From: Amos Bird Date: Fri, 11 Sep 2020 15:46:14 +0800 Subject: [PATCH 054/121] Fix missing clone in replace column transformer --- src/Parsers/ASTColumnsTransformers.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Parsers/ASTColumnsTransformers.cpp b/src/Parsers/ASTColumnsTransformers.cpp index 2625a03830b..43d54f07ab8 100644 --- a/src/Parsers/ASTColumnsTransformers.cpp +++ b/src/Parsers/ASTColumnsTransformers.cpp @@ -110,7 +110,7 @@ void ASTColumnsReplaceTransformer::replaceChildren(ASTPtr & node, const ASTPtr & if (const auto * id = child->as()) { if (id->shortName() == name) - child = replacement; + child = replacement->clone(); } else replaceChildren(child, replacement, name); From 3b9ab3f1be330b5ae7ffd7c68fd629ad3ebc9f6b Mon Sep 17 00:00:00 2001 From: alesapin Date: Fri, 11 Sep 2020 12:23:31 +0300 Subject: [PATCH 055/121] Fix if --- src/Functions/if.cpp | 19 +++++++++++++++---- 1 file changed, 15 insertions(+), 4 deletions(-) diff --git a/src/Functions/if.cpp b/src/Functions/if.cpp index 20848bede32..584bed3f8c5 100644 --- a/src/Functions/if.cpp +++ b/src/Functions/if.cpp @@ -604,7 +604,6 @@ private: const ColumnUInt8 * cond_col, Block & block, const ColumnNumbers & arguments, size_t result, size_t input_rows_count) { /// Convert both columns to the common type (if needed). - const ColumnWithTypeAndName & arg1 = block.getByPosition(arguments[1]); const ColumnWithTypeAndName & arg2 = block.getByPosition(arguments[2]); @@ -765,10 +764,22 @@ private: return ColumnNullable::create(materialized, ColumnUInt8::create(column->size(), 0)); } - static ColumnPtr getNestedColumn(const ColumnPtr & column) + /// Return nested column recursively removing Nullable, examples: + /// Nullable(size = 1, Int32(size = 1), UInt8(size = 1)) -> Int32(size = 1) + /// Const(size = 0, Nullable(size = 1, Int32(size = 1), UInt8(size = 1))) -> + /// Const(size = 0, Int32(size = 1)) + static ColumnPtr recursiveGetNestedColumnWithoutNullable(const ColumnPtr & column) { if (const auto * nullable = checkAndGetColumn(*column)) + { + /// Nullable cannot contain Nullable return nullable->getNestedColumnPtr(); + } + else if (const auto * column_const = checkAndGetColumn(*column)) + { + /// Save Constant, but remove Nullable + return ColumnConst::create(recursiveGetNestedColumnWithoutNullable(column_const->getDataColumnPtr()), column->size()); + } return column; } @@ -826,12 +837,12 @@ private: { arg_cond, { - getNestedColumn(arg_then.column), + recursiveGetNestedColumnWithoutNullable(arg_then.column), removeNullable(arg_then.type), "" }, { - getNestedColumn(arg_else.column), + recursiveGetNestedColumnWithoutNullable(arg_else.column), removeNullable(arg_else.type), "" }, From e25b1da29fa168b24464c83c1f661b363916afad Mon Sep 17 00:00:00 2001 From: alesapin Date: Fri, 11 Sep 2020 13:53:26 +0300 Subject: [PATCH 056/121] Disable -Wstringop-overflow for gcc-10 --- cmake/warnings.cmake | 11 +++++++++-- src/Storages/MergeTree/MergeTreePartition.cpp | 8 -------- 2 files changed, 9 insertions(+), 10 deletions(-) diff --git a/cmake/warnings.cmake b/cmake/warnings.cmake index aec3e46ffa6..6b26b9b95a5 100644 --- a/cmake/warnings.cmake +++ b/cmake/warnings.cmake @@ -169,9 +169,16 @@ elseif (COMPILER_GCC) # Warn if vector operation is not implemented via SIMD capabilities of the architecture add_cxx_compile_options(-Wvector-operation-performance) - # XXX: gcc10 stuck with this option while compiling GatherUtils code - # (anyway there are builds with clang, that will warn) if (CMAKE_CXX_COMPILER_VERSION VERSION_GREATER_EQUAL 10) + # XXX: gcc10 stuck with this option while compiling GatherUtils code + # (anyway there are builds with clang, that will warn) add_cxx_compile_options(-Wno-sequence-point) + # XXX: gcc10 false positive with this warning in MergeTreePartition.cpp + # inlined from 'void writeHexByteLowercase(UInt8, void*)' at ../src/Common/hex.h:39:11, + # inlined from 'DB::String DB::MergeTreePartition::getID(const DB::Block&) const' at ../src/Storages/MergeTree/MergeTreePartition.cpp:85:30: + # ../contrib/libc-headers/x86_64-linux-gnu/bits/string_fortified.h:34:33: error: writing 2 bytes into a region of size 0 [-Werror=stringop-overflow=] + # 34 | return __builtin___memcpy_chk (__dest, __src, __len, __bos0 (__dest)); + # For some reason (bug in gcc?) macro 'GCC diagnostic ignored "-Wstringop-overflow"' doesn't help. + add_cxx_compile_options(-Wno-stringop-overflow) endif() endif () diff --git a/src/Storages/MergeTree/MergeTreePartition.cpp b/src/Storages/MergeTree/MergeTreePartition.cpp index 2802b842f54..880a3aa181d 100644 --- a/src/Storages/MergeTree/MergeTreePartition.cpp +++ b/src/Storages/MergeTree/MergeTreePartition.cpp @@ -29,11 +29,6 @@ String MergeTreePartition::getID(const MergeTreeData & storage) const return getID(storage.getInMemoryMetadataPtr()->getPartitionKey().sample_block); } -#if defined (__GNUC__) && __GNUC__ >= 10 - #pragma GCC diagnostic push - #pragma GCC diagnostic ignored "-Wstringop-overflow" -#endif - /// NOTE: This ID is used to create part names which are then persisted in ZK and as directory names on the file system. /// So if you want to change this method, be sure to guarantee compatibility with existing table data. String MergeTreePartition::getID(const Block & partition_key_sample) const @@ -92,9 +87,6 @@ String MergeTreePartition::getID(const Block & partition_key_sample) const return result; } -#if defined (__GNUC__) && __GNUC__ >= 10 - #pragma GCC diagnostic pop -#endif void MergeTreePartition::serializeText(const MergeTreeData & storage, WriteBuffer & out, const FormatSettings & format_settings) const { From c36192db233af7ce3f971a0cd950db4cfbb6175d Mon Sep 17 00:00:00 2001 From: alesapin Date: Fri, 11 Sep 2020 13:54:03 +0300 Subject: [PATCH 057/121] Remove diff --- src/Storages/MergeTree/MergeTreePartition.cpp | 1 - 1 file changed, 1 deletion(-) diff --git a/src/Storages/MergeTree/MergeTreePartition.cpp b/src/Storages/MergeTree/MergeTreePartition.cpp index 880a3aa181d..4a846f63b7c 100644 --- a/src/Storages/MergeTree/MergeTreePartition.cpp +++ b/src/Storages/MergeTree/MergeTreePartition.cpp @@ -87,7 +87,6 @@ String MergeTreePartition::getID(const Block & partition_key_sample) const return result; } - void MergeTreePartition::serializeText(const MergeTreeData & storage, WriteBuffer & out, const FormatSettings & format_settings) const { auto metadata_snapshot = storage.getInMemoryMetadataPtr(); From ebb9de1376d50e834b61b48cc2f4695513244ad9 Mon Sep 17 00:00:00 2001 From: alesapin Date: Fri, 11 Sep 2020 14:13:41 +0300 Subject: [PATCH 058/121] Supress strange warning --- src/Functions/negate.cpp | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/src/Functions/negate.cpp b/src/Functions/negate.cpp index 39ca434ea89..3101513648b 100644 --- a/src/Functions/negate.cpp +++ b/src/Functions/negate.cpp @@ -13,7 +13,14 @@ struct NegateImpl static inline NO_SANITIZE_UNDEFINED ResultType apply(A a) { - return -static_cast(a); +#if defined (__GNUC__) && __GNUC__ >= 10 + #pragma GCC diagnostic push + #pragma GCC diagnostic ignored "-Wvector-operation-performance" +#endif + return -(static_cast(a)); +#if defined (__GNUC__) && __GNUC__ >= 10 + #pragma GCC diagnostic pop +#endif } #if USE_EMBEDDED_COMPILER From 5ce0c21bbe3c08a0f5169bced9dcea208857c88a Mon Sep 17 00:00:00 2001 From: alesapin Date: Fri, 11 Sep 2020 14:24:42 +0300 Subject: [PATCH 059/121] Remove redundant change --- src/Functions/negate.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Functions/negate.cpp b/src/Functions/negate.cpp index 3101513648b..de3995927d3 100644 --- a/src/Functions/negate.cpp +++ b/src/Functions/negate.cpp @@ -17,7 +17,7 @@ struct NegateImpl #pragma GCC diagnostic push #pragma GCC diagnostic ignored "-Wvector-operation-performance" #endif - return -(static_cast(a)); + return -static_cast(a); #if defined (__GNUC__) && __GNUC__ >= 10 #pragma GCC diagnostic pop #endif From a64331d79f04bb9321383269150fe8302289e9b2 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Fri, 11 Sep 2020 16:09:26 +0300 Subject: [PATCH 060/121] fix syncing of WAL --- src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp | 7 +++++++ src/Storages/MergeTree/MergeTreeWriteAheadLog.h | 2 ++ 2 files changed, 9 insertions(+) diff --git a/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp b/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp index bc6738a8321..35fadb999b4 100644 --- a/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp +++ b/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp @@ -37,6 +37,13 @@ MergeTreeWriteAheadLog::MergeTreeWriteAheadLog( }); } +MergeTreeWriteAheadLog::~MergeTreeWriteAheadLog() +{ + std::unique_lock lock(write_mutex); + if (sync_scheduled) + sync_cv.wait(lock, [this] { return !sync_scheduled; }); +} + void MergeTreeWriteAheadLog::init() { out = disk->writeFile(path, DBMS_DEFAULT_BUFFER_SIZE, WriteMode::Append); diff --git a/src/Storages/MergeTree/MergeTreeWriteAheadLog.h b/src/Storages/MergeTree/MergeTreeWriteAheadLog.h index c5675eac696..77c7c7e11e7 100644 --- a/src/Storages/MergeTree/MergeTreeWriteAheadLog.h +++ b/src/Storages/MergeTree/MergeTreeWriteAheadLog.h @@ -35,6 +35,8 @@ public: MergeTreeWriteAheadLog(MergeTreeData & storage_, const DiskPtr & disk_, const String & name = DEFAULT_WAL_FILE_NAME); + ~MergeTreeWriteAheadLog(); + void addPart(const Block & block, const String & part_name); void dropPart(const String & part_name); std::vector restore(const StorageMetadataPtr & metadata_snapshot); From 7bbf7b295095cf6b9315ae9533b82d5ef9e519bc Mon Sep 17 00:00:00 2001 From: Vxider Date: Fri, 11 Sep 2020 23:26:01 +0800 Subject: [PATCH 061/121] improvement chinese translation of remote.md --- .../sql-reference/table-functions/remote.md | 29 +++++++------------ 1 file changed, 11 insertions(+), 18 deletions(-) diff --git a/docs/zh/sql-reference/table-functions/remote.md b/docs/zh/sql-reference/table-functions/remote.md index 1125353e2fa..3ec1da3cd2c 100644 --- a/docs/zh/sql-reference/table-functions/remote.md +++ b/docs/zh/sql-reference/table-functions/remote.md @@ -1,13 +1,6 @@ ---- -machine_translated: true -machine_translated_rev: 72537a2d527c63c07aa5d2361a8829f3895cf2bd -toc_priority: 40 -toc_title: "\u8FDC\u7A0B" ---- - # 远程，远程安全 {#remote-remotesecure} -允许您访问远程服务器，而无需创建 `Distributed` 桌子 +允许您访问远程服务器，而无需创建 `Distributed` 表 签名: @@ -18,10 +11,10 @@ remoteSecure('addresses_expr', db, table[, 'user'[, 'password']]) remoteSecure('addresses_expr', db.table[, 'user'[, 'password']]) ``` -`addresses_expr` – An expression that generates addresses of remote servers. This may be just one server address. The server address is `host:port`，或者只是 `host`. 主机可以指定为服务器名称，也可以指定为IPv4或IPv6地址。 IPv6地址在方括号中指定。 端口是远程服务器上的TCP端口。 如果省略端口，它使用 `tcp_port` 从服务器的配置文件（默认情况下，9000）。 +`addresses_expr` – 代表远程服务器地址的一个表达式。可以只是单个服务器地址。 服务器地址可以是 `host:port` 或 `host`。`host` 可以指定为服务器域名，或是IPV4或IPV6地址。IPv6地址在方括号中指定。`port` 是远程服务器上的TCP端口。 如果省略端口，则使用服务器配置文件中的 `tcp_port` （默认情况为，9000）。 !!! important "重要事项" - IPv6地址需要该端口。 + IPv6地址需要指定端口。 例: @@ -34,7 +27,7 @@ localhost [2a02:6b8:0:1111::11]:9000 ``` -多个地址可以用逗号分隔。 在这种情况下，ClickHouse将使用分布式处理，因此它将将查询发送到所有指定的地址（如具有不同数据的分片）。 +多个地址可以用逗号分隔。在这种情况下，ClickHouse将使用分布式处理，因此它将将查询发送到所有指定的地址（如具有不同数据的分片）。 示例: @@ -56,7 +49,7 @@ example01-{01..02}-1 如果您有多对大括号，它会生成相应集合的直接乘积。 -大括号中的地址和部分地址可以用管道符号(\|)分隔。 在这种情况下，相应的地址集被解释为副本，并且查询将被发送到第一个正常副本。 但是，副本将按照当前设置的顺序进行迭代 [load\_balancing](../../operations/settings/settings.md) 设置。 +大括号中的地址和部分地址可以用管道符号(\|)分隔。 在这种情况下，相应的地址集被解释为副本，并且查询将被发送到第一个正常副本。 但是，副本将按照当前[load\_balancing](../../operations/settings/settings.md)设置的顺序进行迭代。 示例: @@ -66,20 +59,20 @@ example01-{01..02}-{1|2} 此示例指定两个分片，每个分片都有两个副本。 -生成的地址数由常量限制。 现在这是1000个地址。 +生成的地址数由常量限制。目前这是1000个地址。 -使用 `remote` 表函数比创建一个不太优化 `Distributed` 表，因为在这种情况下，服务器连接被重新建立为每个请求。 此外，如果设置了主机名，则会解析这些名称，并且在使用各种副本时不会计算错误。 在处理大量查询时，始终创建 `Distributed` 表的时间提前，不要使用 `remote` 表功能。 +使用 `remote` 表函数没有创建一个 `Distributed` 表更优，因为在这种情况下，将为每个请求重新建立服务器连接。此外，如果设置了主机名，则会解析这些名称，并且在使用各种副本时不会计算错误。 在处理大量查询时，始终优先创建 `Distributed` 表，不要使用 `remote` 表功能。 该 `remote` 表函数可以在以下情况下是有用的: - 访问特定服务器进行数据比较、调试和测试。 -- 查询之间的各种ClickHouse群集用于研究目的。 -- 手动发出的罕见分布式请求。 +- 在多个ClickHouse集群之间的用户研究目的的查询。 +- 手动发出的不频繁分布式请求。 - 每次重新定义服务器集的分布式请求。 -如果未指定用户, `default` 被使用。 +如果未指定用户, 将会使用`default`。 如果未指定密码，则使用空密码。 -`remoteSecure` -相同 `remote` but with secured connection. Default port — [tcp\_port\_secure](../../operations/server-configuration-parameters/settings.md#server_configuration_parameters-tcp_port_secure) 从配置或9440. +`remoteSecure` - 与 `remote` 相同，但是会使用加密链接。默认端口 — [tcp\_port\_secure](../../operations/server-configuration-parameters/settings.md#server_configuration_parameters-tcp_port_secure) 配置文件或或9440. [原始文章](https://clickhouse.tech/docs/en/query_language/table_functions/remote/) From d9394fbf66b5313d5c07bfc3d2e9119837516525 Mon Sep 17 00:00:00 2001 From: alesapin Date: Fri, 11 Sep 2020 18:51:08 +0300 Subject: [PATCH 062/121] Fix code --- src/Core/MultiEnum.h | 4 ++-- tests/ci/ci_config.json | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/Core/MultiEnum.h b/src/Core/MultiEnum.h index 748550a8779..ddfc5b13e86 100644 --- a/src/Core/MultiEnum.h +++ b/src/Core/MultiEnum.h @@ -83,13 +83,13 @@ struct MultiEnum template >> friend bool operator==(ValueType left, MultiEnum right) { - return right == left; + return right.operator==(left); } template friend bool operator!=(L left, MultiEnum right) { - return !(right == left); + return !(right.operator==(left)); } private: diff --git a/tests/ci/ci_config.json b/tests/ci/ci_config.json index adb736a8df3..9a11a06db0d 100644 --- a/tests/ci/ci_config.json +++ b/tests/ci/ci_config.json @@ -92,7 +92,7 @@ "with_coverage": false }, { - "compiler": "gcc-10", + "compiler": "gcc-9", "build-type": "", "sanitizer": "", "package-type": "deb", From 31dbfd07e22a307992ac868590eb8a794178630d Mon Sep 17 00:00:00 2001 From: kssenii Date: Fri, 11 Sep 2020 16:16:24 +0000 Subject: [PATCH 063/121] remove tests crash reason --- tests/integration/test_storage_rabbitmq/test.py | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/tests/integration/test_storage_rabbitmq/test.py b/tests/integration/test_storage_rabbitmq/test.py index c5b65d60de6..370515956ea 100644 --- a/tests/integration/test_storage_rabbitmq/test.py +++ b/tests/integration/test_storage_rabbitmq/test.py @@ -24,8 +24,8 @@ import rabbitmq_pb2 cluster = ClickHouseCluster(__file__) instance = cluster.add_instance('instance', main_configs=['configs/rabbitmq.xml','configs/log_conf.xml'], - with_rabbitmq=True, - clickhouse_path_dir='clickhouse_path') + with_rabbitmq=True) +# clickhouse_path_dir='clickhouse_path') rabbitmq_id = '' @@ -431,6 +431,7 @@ def test_rabbitmq_many_materialized_views(rabbitmq_cluster): rabbitmq_check_result(result2, True) +@pytest.mark.skip(reason="clichouse_path with rabbitmq.proto fails to be exported") @pytest.mark.timeout(180) def test_rabbitmq_protobuf(rabbitmq_cluster): instance.query(''' From b96da75ead4e291d3ca6f9785ebe9b361688f311 Mon Sep 17 00:00:00 2001 From: nikitamikhaylov Date: Fri, 11 Sep 2020 19:44:14 +0300 Subject: [PATCH 064/121] done --- tests/queries/0_stateless/arcadia_skip_list.txt | 1 + 1 file changed, 1 insertion(+) diff --git a/tests/queries/0_stateless/arcadia_skip_list.txt b/tests/queries/0_stateless/arcadia_skip_list.txt index 71e67c811cd..aa8a9f48ce9 100644 --- a/tests/queries/0_stateless/arcadia_skip_list.txt +++ b/tests/queries/0_stateless/arcadia_skip_list.txt @@ -141,3 +141,4 @@ 01460_DistributedFilesToInsert 01474_executable_dictionary 01474_bad_global_join +01473_event_time_microseconds From 489b9c80aca2099e16b0d7380341f69b9633edd9 Mon Sep 17 00:00:00 2001 From: alesapin Date: Wed, 9 Sep 2020 15:18:16 +0300 Subject: [PATCH 065/121] Starting steps --- src/Parsers/ASTAlterQuery.h | 18 ++++++++++++++++++ src/Parsers/ParserAlterQuery.cpp | 24 ++++++++++++++++-------- 2 files changed, 34 insertions(+), 8 deletions(-) diff --git a/src/Parsers/ASTAlterQuery.h b/src/Parsers/ASTAlterQuery.h index df27ba0a3b0..00350d4efa1 100644 --- a/src/Parsers/ASTAlterQuery.h +++ b/src/Parsers/ASTAlterQuery.h @@ -28,11 +28,13 @@ public: ADD_COLUMN, DROP_COLUMN, MODIFY_COLUMN, + MODIFY_COLUMN_REMOVE_PROPERTY, COMMENT_COLUMN, RENAME_COLUMN, MODIFY_ORDER_BY, MODIFY_SAMPLE_BY, MODIFY_TTL, + REMOVE_TABLE_TTL, MATERIALIZE_TTL, MODIFY_SETTING, MODIFY_QUERY, @@ -61,6 +63,20 @@ public: LIVE_VIEW_REFRESH, }; + /// Which property user wants to remove from column + enum RemoveProperty + { + /// Default specifiers + DEFAULT, + MATERIALIZED, + ALIAS, + + /// Other properties + COMMENT, + CODEC, + TTL + }; + Type type = NO_TYPE; /** The ADD COLUMN query stores the name and type of the column to add @@ -167,6 +183,8 @@ public: /// Target column name ASTPtr rename_to; + RemoveProperty to_remove; + String getID(char delim) const override { return "AlterCommand" + (delim + std::to_string(static_cast(type))); } ASTPtr clone() const override; diff --git a/src/Parsers/ParserAlterQuery.cpp b/src/Parsers/ParserAlterQuery.cpp index 9930bb649b4..b7bde35139b 100644 --- a/src/Parsers/ParserAlterQuery.cpp +++ b/src/Parsers/ParserAlterQuery.cpp @@ -82,6 +82,8 @@ bool ParserAlterCommand::parseImpl(Pos & pos, ASTPtr & node, Expected & expected ParserKeyword s_where("WHERE"); ParserKeyword s_to("TO"); + ParserKeyword s_remove("REMOVE"); + ParserCompoundIdentifier parser_name; ParserStringLiteral parser_string_literal; ParserCompoundColumnDeclaration parser_col_decl; @@ -430,18 +432,24 @@ bool ParserAlterCommand::parseImpl(Pos & pos, ASTPtr & node, Expected & expected if (s_if_exists.ignore(pos, expected)) command->if_exists = true; - if (!parser_modify_col_decl.parse(pos, command->col_decl, expected)) - return false; - - if (s_first.ignore(pos, expected)) - command->first = true; - else if (s_after.ignore(pos, expected)) + if (s_remove.ignore(pos, expected)) { - if (!parser_name.parse(pos, command->column, expected)) + } + else + { + if (!parser_modify_col_decl.parse(pos, command->col_decl, expected)) return false; + + if (s_first.ignore(pos, expected)) + command->first = true; + else if (s_after.ignore(pos, expected)) + { + if (!parser_name.parse(pos, command->column, expected)) + return false; + } + command->type = ASTAlterCommand::MODIFY_COLUMN; } - command->type = ASTAlterCommand::MODIFY_COLUMN; } else if (s_modify_order_by.ignore(pos, expected)) { From a5f889987412404de0b5578492957e745c86782e Mon Sep 17 00:00:00 2001 From: alesapin Date: Fri, 11 Sep 2020 16:44:48 +0300 Subject: [PATCH 066/121] First implementation --- src/Parsers/ASTAlterQuery.cpp | 49 +++++++++++-- src/Parsers/ASTAlterQuery.h | 34 ++++----- src/Parsers/ParserAlterQuery.cpp | 36 +++++++-- src/Storages/AlterCommands.cpp | 121 +++++++++++++++++++++++++------ src/Storages/AlterCommands.h | 12 +-- 5 files changed, 193 insertions(+), 59 deletions(-) diff --git a/src/Parsers/ASTAlterQuery.cpp b/src/Parsers/ASTAlterQuery.cpp index d033cdc79a2..b7a55c58714 100644 --- a/src/Parsers/ASTAlterQuery.cpp +++ b/src/Parsers/ASTAlterQuery.cpp @@ -99,12 +99,42 @@ void ASTAlterCommand::formatImpl( settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << "MODIFY COLUMN " << (if_exists ? "IF EXISTS " : "") << (settings.hilite ? hilite_none : ""); col_decl->formatImpl(settings, state, frame); - if (first) - settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << " FIRST " << (settings.hilite ? hilite_none : ""); - else if (column) /// AFTER + if (to_remove != RemoveProperty::NO_PROPERTY) { - settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << " AFTER " << (settings.hilite ? hilite_none : ""); - column->formatImpl(settings, state, frame); + settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << "REMOVE "; + switch (to_remove) + { + case RemoveProperty::DEFAULT: + settings.ostr << "DEFAULT"; + break; + case RemoveProperty::MATERIALIZED: + settings.ostr << "MATERIALIZED"; + break; + case RemoveProperty::ALIAS: + settings.ostr << "ALIAS"; + break; + case RemoveProperty::COMMENT: + settings.ostr << "COMMENT"; + break; + case RemoveProperty::CODEC: + settings.ostr << "CODEC"; + break; + case RemoveProperty::TTL: + settings.ostr << "TTL"; + break; + default: + __builtin_unreachable(); + } + } + else + { + if (first) + settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << " FIRST " << (settings.hilite ? hilite_none : ""); + else if (column) /// AFTER + { + settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << " AFTER " << (settings.hilite ? hilite_none : ""); + column->formatImpl(settings, state, frame); + } } } else if (type == ASTAlterCommand::COMMENT_COLUMN) @@ -278,7 +308,14 @@ void ASTAlterCommand::formatImpl( else if (type == ASTAlterCommand::MODIFY_TTL) { settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << "MODIFY TTL " << (settings.hilite ? hilite_none : ""); - ttl->formatImpl(settings, state, frame); + if (ttl) + { + ttl->formatImpl(settings, state, frame); + } + else if (to_remove == RemoveProperty::TTL) + { + settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str<< " REMOVE " << (settings.hilite ? hilite_none : ""); + } } else if (type == ASTAlterCommand::MATERIALIZE_TTL) { diff --git a/src/Parsers/ASTAlterQuery.h b/src/Parsers/ASTAlterQuery.h index 00350d4efa1..a7822806797 100644 --- a/src/Parsers/ASTAlterQuery.h +++ b/src/Parsers/ASTAlterQuery.h @@ -9,6 +9,22 @@ namespace DB { +/// Which property user wants to remove from column +enum class RemoveProperty +{ + NO_PROPERTY, + /// Default specifiers + DEFAULT, + MATERIALIZED, + ALIAS, + + /// Other properties + COMMENT, + CODEC, + TTL +}; + + /** ALTER query: * ALTER TABLE [db.]name_type * ADD COLUMN col_name type [AFTER col_after], @@ -28,13 +44,11 @@ public: ADD_COLUMN, DROP_COLUMN, MODIFY_COLUMN, - MODIFY_COLUMN_REMOVE_PROPERTY, COMMENT_COLUMN, RENAME_COLUMN, MODIFY_ORDER_BY, MODIFY_SAMPLE_BY, MODIFY_TTL, - REMOVE_TABLE_TTL, MATERIALIZE_TTL, MODIFY_SETTING, MODIFY_QUERY, @@ -63,20 +77,6 @@ public: LIVE_VIEW_REFRESH, }; - /// Which property user wants to remove from column - enum RemoveProperty - { - /// Default specifiers - DEFAULT, - MATERIALIZED, - ALIAS, - - /// Other properties - COMMENT, - CODEC, - TTL - }; - Type type = NO_TYPE; /** The ADD COLUMN query stores the name and type of the column to add @@ -183,7 +183,7 @@ public: /// Target column name ASTPtr rename_to; - RemoveProperty to_remove; + RemoveProperty to_remove = RemoveProperty::NO_PROPERTY; String getID(char delim) const override { return "AlterCommand" + (delim + std::to_string(static_cast(type))); } diff --git a/src/Parsers/ParserAlterQuery.cpp b/src/Parsers/ParserAlterQuery.cpp index b7bde35139b..4a1418cbe6a 100644 --- a/src/Parsers/ParserAlterQuery.cpp +++ b/src/Parsers/ParserAlterQuery.cpp @@ -83,6 +83,12 @@ bool ParserAlterCommand::parseImpl(Pos & pos, ASTPtr & node, Expected & expected ParserKeyword s_to("TO"); ParserKeyword s_remove("REMOVE"); + ParserKeyword s_default("DEFAULT"); + ParserKeyword s_materialized("MATERIALIZED"); + ParserKeyword s_alias("ALIAS"); + ParserKeyword s_comment("COMMENT"); + ParserKeyword s_codec("CODEC"); + ParserKeyword s_ttl("TTL"); ParserCompoundIdentifier parser_name; ParserStringLiteral parser_string_literal; @@ -432,14 +438,28 @@ bool ParserAlterCommand::parseImpl(Pos & pos, ASTPtr & node, Expected & expected if (s_if_exists.ignore(pos, expected)) command->if_exists = true; + if (!parser_modify_col_decl.parse(pos, command->col_decl, expected)) + return false; + if (s_remove.ignore(pos, expected)) { + if (s_default.ignore(pos, expected)) + command->to_remove = RemoveProperty::DEFAULT; + else if (s_materialized.ignore(pos, expected)) + command->to_remove = RemoveProperty::MATERIALIZED; + else if (s_alias.ignore(pos, expected)) + command->to_remove = RemoveProperty::ALIAS; + else if (s_comment.ignore(pos, expected)) + command->to_remove = RemoveProperty::COMMENT; + else if (s_codec.ignore(pos, expected)) + command->to_remove = RemoveProperty::CODEC; + else if (s_ttl.ignore(pos, expected)) + command->to_remove = RemoveProperty::TTL; + else + return false; } else { - if (!parser_modify_col_decl.parse(pos, command->col_decl, expected)) - return false; - if (s_first.ignore(pos, expected)) command->first = true; else if (s_after.ignore(pos, expected)) @@ -447,9 +467,8 @@ bool ParserAlterCommand::parseImpl(Pos & pos, ASTPtr & node, Expected & expected if (!parser_name.parse(pos, command->column, expected)) return false; } - command->type = ASTAlterCommand::MODIFY_COLUMN; } - + command->type = ASTAlterCommand::MODIFY_COLUMN; } else if (s_modify_order_by.ignore(pos, expected)) { @@ -501,7 +520,12 @@ bool ParserAlterCommand::parseImpl(Pos & pos, ASTPtr & node, Expected & expected else if (s_modify_ttl.ignore(pos, expected)) { if (!parser_ttl_list.parse(pos, command->ttl, expected)) - return false; + { + if (s_remove.ignore(pos, expected)) + command->to_remove = RemoveProperty::TTL; + else + return false; + } command->type = ASTAlterCommand::MODIFY_TTL; } else if (s_materialize_ttl.ignore(pos, expected)) diff --git a/src/Storages/AlterCommands.cpp b/src/Storages/AlterCommands.cpp index caf98e911ab..5c7a45a27be 100644 --- a/src/Storages/AlterCommands.cpp +++ b/src/Storages/AlterCommands.cpp @@ -113,6 +113,8 @@ std::optional AlterCommand::parse(const ASTAlterCommand * command_ const auto & ast_col_decl = command_ast->col_decl->as(); command.column_name = ast_col_decl.name; + command.to_remove = command_ast->to_remove; + if (ast_col_decl.type) { command.data_type = data_type_factory.get(ast_col_decl.type); @@ -301,24 +303,45 @@ void AlterCommand::apply(StorageInMemoryMetadata & metadata, const Context & con { metadata.columns.modify(column_name, after_column, first, [&](ColumnDescription & column) { - if (codec) - column.codec = CompressionCodecFactory::instance().validateCodecAndGetPreprocessedAST(codec, data_type ? data_type : column.type, false); - - if (comment) - column.comment = *comment; - - if (ttl) - column.ttl = ttl; - - if (data_type) - column.type = data_type; - - /// User specified default expression or changed - /// datatype. We have to replace default. - if (default_expression || data_type) + if (to_remove == RemoveProperty::DEFAULT + || to_remove == RemoveProperty::MATERIALIZED + || to_remove == RemoveProperty::ALIAS) { - column.default_desc.kind = default_kind; - column.default_desc.expression = default_expression; + column.default_desc = ColumnDefault{}; + } + else if (to_remove == RemoveProperty::CODEC) + { + column.codec.reset(); + } + else if (to_remove == RemoveProperty::COMMENT) + { + column.comment = String{}; + } + else if (to_remove == RemoveProperty::TTL) + { + column.ttl.reset(); + } + else + { + if (codec) + column.codec = CompressionCodecFactory::instance().validateCodecAndGetPreprocessedAST(codec, data_type ? data_type : column.type, false); + + if (comment) + column.comment = *comment; + + if (ttl) + column.ttl = ttl; + + if (data_type) + column.type = data_type; + + /// User specified default expression or changed + /// datatype. We have to replace default. + if (default_expression || data_type) + { + column.default_desc.kind = default_kind; + column.default_desc.expression = default_expression; + } } }); @@ -448,7 +471,10 @@ void AlterCommand::apply(StorageInMemoryMetadata & metadata, const Context & con } else if (type == MODIFY_TTL) { - metadata.table_ttl = TTLTableDescription::getTTLForTableFromAST(ttl, metadata.columns, context, metadata.primary_key); + if (to_remove == RemoveProperty::TTL) + metadata.table_ttl = TTLTableDescription{}; + else + metadata.table_ttl = TTLTableDescription::getTTLForTableFromAST(ttl, metadata.columns, context, metadata.primary_key); } else if (type == MODIFY_QUERY) { @@ -590,6 +616,10 @@ bool AlterCommand::isRequireMutationStage(const StorageInMemoryMetadata & metada if (type != MODIFY_COLUMN || data_type == nullptr) return false; + /// We remove properties on metadata level + if (type == MODIFY_COLUMN && to_remove != RemoveProperty::NO_PROPERTY) + return false; + for (const auto & column : metadata.columns.getAllPhysical()) { if (column.name == column_name && !isMetadataOnlyConversion(column.type.get(), data_type.get())) @@ -783,14 +813,30 @@ void AlterCommands::prepare(const StorageInMemoryMetadata & metadata) if (!has_column && command.if_exists) command.ignore = true; - if (has_column && command.data_type) + if (has_column) { auto column_from_table = columns.get(command.column_name); - if (!command.default_expression && column_from_table.default_desc.expression) + if (command.to_remove != RemoveProperty::NO_PROPERTY) + { + auto column_default = columns.getDefault(command.column_name); + if (!column_default + && (command.to_remove == RemoveProperty::ALIAS || command.to_remove == RemoveProperty::DEFAULT + || command.to_remove == RemoveProperty::MATERIALIZED)) + command.ignore = true; + + if (command.to_remove == RemoveProperty::TTL && column_from_table.ttl == nullptr) + command.ignore = true; + if (command.to_remove == RemoveProperty::COMMENT && column_from_table.comment == "") + command.ignore = true; + if (command.to_remove == RemoveProperty::CODEC && column_from_table.codec == nullptr) + command.ignore = true; + } + else if (command.data_type && !command.default_expression && column_from_table.default_desc.expression) { command.default_kind = column_from_table.default_desc.kind; command.default_expression = column_from_table.default_desc.expression; } + } } else if (command.type == AlterCommand::ADD_COLUMN) @@ -805,6 +851,11 @@ void AlterCommands::prepare(const StorageInMemoryMetadata & metadata) if (!has_column && command.if_exists) command.ignore = true; } + else if (command.type == AlterCommand::MODIFY_TTL) + { + if (!metadata.hasAnyTTL()) + command.ignore = true; + } } prepared = true; } @@ -857,6 +908,34 @@ void AlterCommands::validate(const StorageInMemoryMetadata & metadata, const Con if (command.codec) CompressionCodecFactory::instance().validateCodecAndGetPreprocessedAST(command.codec, command.data_type, !context.getSettingsRef().allow_suspicious_codecs); + auto column_default = all_columns.getDefault(column_name); + if (column_default) + { + if (command.to_remove == RemoveProperty::DEFAULT && column_default->kind != ColumnDefaultKind::Default) + { + throw Exception{ + ErrorCodes::BAD_ARGUMENTS, + "Cannot remove DEFAULT from column {}, because column default type is {}. Use REMOVE {} to delete it.", + backQuote(column_name), toString(column_default->kind), toString(column_default->kind) + }; + } + if (command.to_remove == RemoveProperty::MATERIALIZED && column_default->kind != ColumnDefaultKind::Materialized) + { + throw Exception{ + ErrorCodes::BAD_ARGUMENTS, + "Cannot remove MATERIALIZED from column {}, because column default type is {}. Use REMOVE {} to delete it.", + backQuote(column_name), toString(column_default->kind), toString(column_default->kind) + }; + } + if (command.to_remove == RemoveProperty::ALIAS && column_default->kind != ColumnDefaultKind::Alias) + { + throw Exception{ + ErrorCodes::BAD_ARGUMENTS, + "Cannot remove ALIAS from column {}, because column default type is {}. Use REMOVE {} to delete it.", + backQuote(column_name), toString(column_default->kind), toString(column_default->kind) + }; + } + } modified_columns.emplace(column_name); } @@ -1048,7 +1127,7 @@ MutationCommands AlterCommands::getMutationCommands(StorageInMemoryMetadata meta { for (const auto & alter_cmd : *this) { - if (alter_cmd.isTTLAlter(metadata)) + if (alter_cmd.isTTLAlter(metadata) && alter_cmd.to_remove != RemoveProperty::TTL) { result.push_back(createMaterializeTTLCommand()); break; diff --git a/src/Storages/AlterCommands.h b/src/Storages/AlterCommands.h index 3578507a361..a4eff5523b7 100644 --- a/src/Storages/AlterCommands.h +++ b/src/Storages/AlterCommands.h @@ -107,16 +107,13 @@ struct AlterCommand /// Target column name String rename_to; + /// What to remove from column (or TTL) + RemoveProperty to_remove; + static std::optional parse(const ASTAlterCommand * command); void apply(StorageInMemoryMetadata & metadata, const Context & context) const; - /// Checks that alter query changes data. For MergeTree: - /// * column files (data and marks) - /// * each part meta (columns.txt) - /// in each part on disk (it's not lightweight alter). - bool isModifyingData(const StorageInMemoryMetadata & metadata) const; - /// Check that alter command require data modification (mutation) to be /// executed. For example, cast from Date to UInt16 type can be executed /// without any data modifications. But column drop or modify from UInt16 to @@ -164,9 +161,6 @@ public: /// Commands have to be prepared before apply. void apply(StorageInMemoryMetadata & metadata, const Context & context) const; - /// At least one command modify data on disk. - bool isModifyingData(const StorageInMemoryMetadata & metadata) const; - /// At least one command modify settings. bool isSettingsAlter() const; From a4c43e51b91eef3bf8337653dfec36478ad95ca7 Mon Sep 17 00:00:00 2001 From: alesapin Date: Fri, 11 Sep 2020 17:56:09 +0300 Subject: [PATCH 067/121] Add a test --- src/Parsers/ASTAlterQuery.cpp | 2 +- src/Parsers/ParserAlterQuery.cpp | 12 +++- src/Storages/AlterCommands.cpp | 1 - .../01493_alter_remove_properties.reference | 17 ++++++ .../01493_alter_remove_properties.sql | 58 +++++++++++++++++++ 5 files changed, 87 insertions(+), 3 deletions(-) create mode 100644 tests/queries/0_stateless/01493_alter_remove_properties.reference create mode 100644 tests/queries/0_stateless/01493_alter_remove_properties.sql diff --git a/src/Parsers/ASTAlterQuery.cpp b/src/Parsers/ASTAlterQuery.cpp index b7a55c58714..62f33b25c57 100644 --- a/src/Parsers/ASTAlterQuery.cpp +++ b/src/Parsers/ASTAlterQuery.cpp @@ -101,7 +101,7 @@ void ASTAlterCommand::formatImpl( if (to_remove != RemoveProperty::NO_PROPERTY) { - settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << "REMOVE "; + settings.ostr << (settings.hilite ? hilite_keyword : "") << " REMOVE "; switch (to_remove) { case RemoveProperty::DEFAULT: diff --git a/src/Parsers/ParserAlterQuery.cpp b/src/Parsers/ParserAlterQuery.cpp index 4a1418cbe6a..4d6e71e95cf 100644 --- a/src/Parsers/ParserAlterQuery.cpp +++ b/src/Parsers/ParserAlterQuery.cpp @@ -438,7 +438,9 @@ bool ParserAlterCommand::parseImpl(Pos & pos, ASTPtr & node, Expected & expected if (s_if_exists.ignore(pos, expected)) command->if_exists = true; - if (!parser_modify_col_decl.parse(pos, command->col_decl, expected)) + ASTPtr column_name; + Pos stop_pos = pos; + if (!parser_name.parse(pos, column_name, expected)) return false; if (s_remove.ignore(pos, expected)) @@ -457,9 +459,17 @@ bool ParserAlterCommand::parseImpl(Pos & pos, ASTPtr & node, Expected & expected command->to_remove = RemoveProperty::TTL; else return false; + + auto column_declaration = std::make_shared(); + tryGetIdentifierNameInto(column_name, column_declaration->name); + command->col_decl = column_declaration; } else { + pos = stop_pos; + if (!parser_modify_col_decl.parse(pos, command->col_decl, expected)) + return false; + if (s_first.ignore(pos, expected)) command->first = true; else if (s_after.ignore(pos, expected)) diff --git a/src/Storages/AlterCommands.cpp b/src/Storages/AlterCommands.cpp index 5c7a45a27be..bc6455ef420 100644 --- a/src/Storages/AlterCommands.cpp +++ b/src/Storages/AlterCommands.cpp @@ -111,7 +111,6 @@ std::optional AlterCommand::parse(const ASTAlterCommand * command_ command.type = AlterCommand::MODIFY_COLUMN; const auto & ast_col_decl = command_ast->col_decl->as(); - command.column_name = ast_col_decl.name; command.to_remove = command_ast->to_remove; diff --git a/tests/queries/0_stateless/01493_alter_remove_properties.reference b/tests/queries/0_stateless/01493_alter_remove_properties.reference new file mode 100644 index 00000000000..4ce7a574742 --- /dev/null +++ b/tests/queries/0_stateless/01493_alter_remove_properties.reference @@ -0,0 +1,17 @@ +CREATE TABLE default.prop_table\n(\n `column_default` UInt64 DEFAULT 42,\n `column_materialized` UInt64 MATERIALIZED column_default * 42,\n `column_alias` UInt64 ALIAS column_default + 1,\n `column_codec` String CODEC(ZSTD(10)),\n `column_comment` Date COMMENT \'Some comment\',\n `column_ttl` UInt64 TTL column_comment + toIntervalMonth(1)\n)\nENGINE = MergeTree()\nORDER BY tuple()\nTTL column_comment + toIntervalMonth(2)\nSETTINGS index_granularity = 8192 +42 1764 43 str 2019-10-01 1 +CREATE TABLE default.prop_table\n(\n `column_default` UInt64 DEFAULT 42,\n `column_materialized` UInt64 MATERIALIZED column_default * 42,\n `column_alias` UInt64 ALIAS column_default + 1,\n `column_codec` String CODEC(ZSTD(10)),\n `column_comment` Date,\n `column_ttl` UInt64 TTL column_comment + toIntervalMonth(1)\n)\nENGINE = MergeTree()\nORDER BY tuple()\nTTL column_comment + toIntervalMonth(2)\nSETTINGS index_granularity = 8192 +CREATE TABLE default.prop_table\n(\n `column_default` UInt64 DEFAULT 42,\n `column_materialized` UInt64 MATERIALIZED column_default * 42,\n `column_alias` UInt64 ALIAS column_default + 1,\n `column_codec` String,\n `column_comment` Date,\n `column_ttl` UInt64 TTL column_comment + toIntervalMonth(1)\n)\nENGINE = MergeTree()\nORDER BY tuple()\nTTL column_comment + toIntervalMonth(2)\nSETTINGS index_granularity = 8192 +42 1764 0 str 2019-10-01 1 +CREATE TABLE default.prop_table\n(\n `column_default` UInt64 DEFAULT 42,\n `column_materialized` UInt64 MATERIALIZED column_default * 42,\n `column_alias` UInt64,\n `column_codec` String,\n `column_comment` Date,\n `column_ttl` UInt64 TTL column_comment + toIntervalMonth(1)\n)\nENGINE = MergeTree()\nORDER BY tuple()\nTTL column_comment + toIntervalMonth(2)\nSETTINGS index_granularity = 8192 +42 1764 0 str 2019-10-01 1 +42 1764 33 trs 2020-01-01 2 +CREATE TABLE default.prop_table\n(\n `column_default` UInt64 DEFAULT 42,\n `column_materialized` UInt64,\n `column_alias` UInt64,\n `column_codec` String,\n `column_comment` Date,\n `column_ttl` UInt64 TTL column_comment + toIntervalMonth(1)\n)\nENGINE = MergeTree()\nORDER BY tuple()\nTTL column_comment + toIntervalMonth(2)\nSETTINGS index_granularity = 8192 +42 1764 0 str 2019-10-01 1 +42 1764 33 trs 2020-01-01 2 +42 11 44 rts 2020-02-01 3 +CREATE TABLE default.prop_table\n(\n `column_default` UInt64,\n `column_materialized` UInt64,\n `column_alias` UInt64,\n `column_codec` String,\n `column_comment` Date,\n `column_ttl` UInt64 TTL column_comment + toIntervalMonth(1)\n)\nENGINE = MergeTree()\nORDER BY tuple()\nTTL column_comment + toIntervalMonth(2)\nSETTINGS index_granularity = 8192 +42 1764 0 str 2019-10-01 1 +42 1764 33 trs 2020-01-01 2 +42 11 44 rts 2020-02-01 3 +0 22 55 tsr 2020-03-01 4 diff --git a/tests/queries/0_stateless/01493_alter_remove_properties.sql b/tests/queries/0_stateless/01493_alter_remove_properties.sql new file mode 100644 index 00000000000..25000a50235 --- /dev/null +++ b/tests/queries/0_stateless/01493_alter_remove_properties.sql @@ -0,0 +1,58 @@ +DROP TABLE IF EXISTS prop_table; + +CREATE TABLE prop_table +( + column_default UInt64 DEFAULT 42, + column_materialized UInt64 MATERIALIZED column_default * 42, + column_alias UInt64 ALIAS column_default + 1, + column_codec String CODEC(ZSTD(10)), + column_comment Date COMMENT 'Some comment', + column_ttl UInt64 TTL column_comment + INTERVAL 1 MONTH +) +ENGINE MergeTree() +ORDER BY tuple() +TTL column_comment + INTERVAL 2 MONTH; + +SHOW CREATE TABLE prop_table; + +SYSTEM STOP TTL MERGES prop_table; + +INSERT INTO prop_table (column_codec, column_comment, column_ttl) VALUES ('str', toDate('2019-10-01'), 1); + +SELECT column_default, column_materialized, column_alias, column_codec, column_comment, column_ttl FROM prop_table; + +ALTER TABLE prop_table MODIFY COLUMN column_comment REMOVE COMMENT; + +SHOW CREATE TABLE prop_table; + +ALTER TABLE prop_table MODIFY COLUMN column_codec REMOVE CODEC; + +SHOW CREATE TABLE prop_table; + +ALTER TABLE prop_table MODIFY COLUMN column_alias REMOVE ALIAS; + +SELECT column_default, column_materialized, column_alias, column_codec, column_comment, column_ttl FROM prop_table; + +SHOW CREATE TABLE prop_table; + +INSERT INTO prop_table (column_alias, column_codec, column_comment, column_ttl) VALUES (33, 'trs', toDate('2020-01-01'), 2); + +SELECT column_default, column_materialized, column_alias, column_codec, column_comment, column_ttl FROM prop_table ORDER BY column_ttl; + +ALTER TABLE prop_table MODIFY COLUMN column_materialized REMOVE MATERIALIZED; + +SHOW CREATE TABLE prop_table; + +INSERT INTO prop_table (column_materialized, column_alias, column_codec, column_comment, column_ttl) VALUES (11, 44, 'rts', toDate('2020-02-01'), 3); + +SELECT column_default, column_materialized, column_alias, column_codec, column_comment, column_ttl FROM prop_table ORDER BY column_ttl; + +ALTER TABLE prop_table MODIFY COLUMN column_default REMOVE DEFAULT; + +SHOW CREATE TABLE prop_table; + +INSERT INTO prop_table (column_materialized, column_alias, column_codec, column_comment, column_ttl) VALUES (22, 55, 'tsr', toDate('2020-03-01'), 4); + +SELECT column_default, column_materialized, column_alias, column_codec, column_comment, column_ttl FROM prop_table ORDER BY column_ttl; + +DROP TABLE IF EXISTS prop_table; From 6dd75182f07310a5fa5a0fd72618969ac51d9dad Mon Sep 17 00:00:00 2001 From: alesapin Date: Fri, 11 Sep 2020 19:21:08 +0300 Subject: [PATCH 068/121] Better --- src/Parsers/ParserAlterQuery.cpp | 11 ++++------- src/Storages/AlterCommands.cpp | 2 +- 2 files changed, 5 insertions(+), 8 deletions(-) diff --git a/src/Parsers/ParserAlterQuery.cpp b/src/Parsers/ParserAlterQuery.cpp index 4d6e71e95cf..a6032bd38db 100644 --- a/src/Parsers/ParserAlterQuery.cpp +++ b/src/Parsers/ParserAlterQuery.cpp @@ -529,13 +529,10 @@ bool ParserAlterCommand::parseImpl(Pos & pos, ASTPtr & node, Expected & expected } else if (s_modify_ttl.ignore(pos, expected)) { - if (!parser_ttl_list.parse(pos, command->ttl, expected)) - { - if (s_remove.ignore(pos, expected)) - command->to_remove = RemoveProperty::TTL; - else - return false; - } + if (s_remove.ignore(pos, expected)) + command->to_remove = RemoveProperty::TTL; + else if (!parser_ttl_list.parse(pos, command->ttl, expected)) + return false; command->type = ASTAlterCommand::MODIFY_TTL; } else if (s_materialize_ttl.ignore(pos, expected)) diff --git a/src/Storages/AlterCommands.cpp b/src/Storages/AlterCommands.cpp index bc6455ef420..01b56e0e128 100644 --- a/src/Storages/AlterCommands.cpp +++ b/src/Storages/AlterCommands.cpp @@ -825,7 +825,7 @@ void AlterCommands::prepare(const StorageInMemoryMetadata & metadata) if (command.to_remove == RemoveProperty::TTL && column_from_table.ttl == nullptr) command.ignore = true; - if (command.to_remove == RemoveProperty::COMMENT && column_from_table.comment == "") + if (command.to_remove == RemoveProperty::COMMENT && column_from_table.comment.empty()) command.ignore = true; if (command.to_remove == RemoveProperty::CODEC && column_from_table.codec == nullptr) command.ignore = true; From c0dafb0283d52d6a9df714105c75d1e8d00603a0 Mon Sep 17 00:00:00 2001 From: alesapin Date: Fri, 11 Sep 2020 20:04:41 +0300 Subject: [PATCH 069/121] Disable test --- tests/integration/test_adaptive_granularity/test.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/integration/test_adaptive_granularity/test.py b/tests/integration/test_adaptive_granularity/test.py index 21d65588de4..d5ac91671e0 100644 --- a/tests/integration/test_adaptive_granularity/test.py +++ b/tests/integration/test_adaptive_granularity/test.py @@ -298,7 +298,7 @@ def test_mixed_granularity_single_node(start_dynamic_cluster, node): #still works assert node.query("SELECT count() from table_with_default_granularity") == '6\n' - +@pytest.mark.skip(reason="flaky") def test_version_update_two_nodes(start_dynamic_cluster): node11.query("INSERT INTO table_with_default_granularity VALUES (toDate('2018-10-01'), 1, 333), (toDate('2018-10-02'), 2, 444)") node12.query("SYSTEM SYNC REPLICA table_with_default_granularity", timeout=20) From 2c4047b280555df2ef2f50240eeb519d4dde1154 Mon Sep 17 00:00:00 2001 From: alesapin Date: Fri, 11 Sep 2020 20:07:00 +0300 Subject: [PATCH 070/121] Revert accident changes --- src/Parsers/ASTAlterQuery.cpp | 49 +------- src/Parsers/ASTAlterQuery.h | 18 --- src/Parsers/ParserAlterQuery.cpp | 51 +------- src/Storages/AlterCommands.cpp | 116 +++--------------- src/Storages/AlterCommands.h | 12 +- .../test_adaptive_granularity/test.py | 2 +- .../01493_alter_remove_properties.reference | 17 --- .../01493_alter_remove_properties.sql | 58 --------- 8 files changed, 41 insertions(+), 282 deletions(-) delete mode 100644 tests/queries/0_stateless/01493_alter_remove_properties.reference delete mode 100644 tests/queries/0_stateless/01493_alter_remove_properties.sql diff --git a/src/Parsers/ASTAlterQuery.cpp b/src/Parsers/ASTAlterQuery.cpp index 62f33b25c57..d033cdc79a2 100644 --- a/src/Parsers/ASTAlterQuery.cpp +++ b/src/Parsers/ASTAlterQuery.cpp @@ -99,42 +99,12 @@ void ASTAlterCommand::formatImpl( settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << "MODIFY COLUMN " << (if_exists ? "IF EXISTS " : "") << (settings.hilite ? hilite_none : ""); col_decl->formatImpl(settings, state, frame); - if (to_remove != RemoveProperty::NO_PROPERTY) + if (first) + settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << " FIRST " << (settings.hilite ? hilite_none : ""); + else if (column) /// AFTER { - settings.ostr << (settings.hilite ? hilite_keyword : "") << " REMOVE "; - switch (to_remove) - { - case RemoveProperty::DEFAULT: - settings.ostr << "DEFAULT"; - break; - case RemoveProperty::MATERIALIZED: - settings.ostr << "MATERIALIZED"; - break; - case RemoveProperty::ALIAS: - settings.ostr << "ALIAS"; - break; - case RemoveProperty::COMMENT: - settings.ostr << "COMMENT"; - break; - case RemoveProperty::CODEC: - settings.ostr << "CODEC"; - break; - case RemoveProperty::TTL: - settings.ostr << "TTL"; - break; - default: - __builtin_unreachable(); - } - } - else - { - if (first) - settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << " FIRST " << (settings.hilite ? hilite_none : ""); - else if (column) /// AFTER - { - settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << " AFTER " << (settings.hilite ? hilite_none : ""); - column->formatImpl(settings, state, frame); - } + settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << " AFTER " << (settings.hilite ? hilite_none : ""); + column->formatImpl(settings, state, frame); } } else if (type == ASTAlterCommand::COMMENT_COLUMN) @@ -308,14 +278,7 @@ void ASTAlterCommand::formatImpl( else if (type == ASTAlterCommand::MODIFY_TTL) { settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str << "MODIFY TTL " << (settings.hilite ? hilite_none : ""); - if (ttl) - { - ttl->formatImpl(settings, state, frame); - } - else if (to_remove == RemoveProperty::TTL) - { - settings.ostr << (settings.hilite ? hilite_keyword : "") << indent_str<< " REMOVE " << (settings.hilite ? hilite_none : ""); - } + ttl->formatImpl(settings, state, frame); } else if (type == ASTAlterCommand::MATERIALIZE_TTL) { diff --git a/src/Parsers/ASTAlterQuery.h b/src/Parsers/ASTAlterQuery.h index a7822806797..df27ba0a3b0 100644 --- a/src/Parsers/ASTAlterQuery.h +++ b/src/Parsers/ASTAlterQuery.h @@ -9,22 +9,6 @@ namespace DB { -/// Which property user wants to remove from column -enum class RemoveProperty -{ - NO_PROPERTY, - /// Default specifiers - DEFAULT, - MATERIALIZED, - ALIAS, - - /// Other properties - COMMENT, - CODEC, - TTL -}; - - /** ALTER query: * ALTER TABLE [db.]name_type * ADD COLUMN col_name type [AFTER col_after], @@ -183,8 +167,6 @@ public: /// Target column name ASTPtr rename_to; - RemoveProperty to_remove = RemoveProperty::NO_PROPERTY; - String getID(char delim) const override { return "AlterCommand" + (delim + std::to_string(static_cast(type))); } ASTPtr clone() const override; diff --git a/src/Parsers/ParserAlterQuery.cpp b/src/Parsers/ParserAlterQuery.cpp index a6032bd38db..9930bb649b4 100644 --- a/src/Parsers/ParserAlterQuery.cpp +++ b/src/Parsers/ParserAlterQuery.cpp @@ -82,14 +82,6 @@ bool ParserAlterCommand::parseImpl(Pos & pos, ASTPtr & node, Expected & expected ParserKeyword s_where("WHERE"); ParserKeyword s_to("TO"); - ParserKeyword s_remove("REMOVE"); - ParserKeyword s_default("DEFAULT"); - ParserKeyword s_materialized("MATERIALIZED"); - ParserKeyword s_alias("ALIAS"); - ParserKeyword s_comment("COMMENT"); - ParserKeyword s_codec("CODEC"); - ParserKeyword s_ttl("TTL"); - ParserCompoundIdentifier parser_name; ParserStringLiteral parser_string_literal; ParserCompoundColumnDeclaration parser_col_decl; @@ -438,46 +430,17 @@ bool ParserAlterCommand::parseImpl(Pos & pos, ASTPtr & node, Expected & expected if (s_if_exists.ignore(pos, expected)) command->if_exists = true; - ASTPtr column_name; - Pos stop_pos = pos; - if (!parser_name.parse(pos, column_name, expected)) + if (!parser_modify_col_decl.parse(pos, command->col_decl, expected)) return false; - if (s_remove.ignore(pos, expected)) + if (s_first.ignore(pos, expected)) + command->first = true; + else if (s_after.ignore(pos, expected)) { - if (s_default.ignore(pos, expected)) - command->to_remove = RemoveProperty::DEFAULT; - else if (s_materialized.ignore(pos, expected)) - command->to_remove = RemoveProperty::MATERIALIZED; - else if (s_alias.ignore(pos, expected)) - command->to_remove = RemoveProperty::ALIAS; - else if (s_comment.ignore(pos, expected)) - command->to_remove = RemoveProperty::COMMENT; - else if (s_codec.ignore(pos, expected)) - command->to_remove = RemoveProperty::CODEC; - else if (s_ttl.ignore(pos, expected)) - command->to_remove = RemoveProperty::TTL; - else + if (!parser_name.parse(pos, command->column, expected)) return false; - - auto column_declaration = std::make_shared(); - tryGetIdentifierNameInto(column_name, column_declaration->name); - command->col_decl = column_declaration; } - else - { - pos = stop_pos; - if (!parser_modify_col_decl.parse(pos, command->col_decl, expected)) - return false; - if (s_first.ignore(pos, expected)) - command->first = true; - else if (s_after.ignore(pos, expected)) - { - if (!parser_name.parse(pos, command->column, expected)) - return false; - } - } command->type = ASTAlterCommand::MODIFY_COLUMN; } else if (s_modify_order_by.ignore(pos, expected)) @@ -529,9 +492,7 @@ bool ParserAlterCommand::parseImpl(Pos & pos, ASTPtr & node, Expected & expected } else if (s_modify_ttl.ignore(pos, expected)) { - if (s_remove.ignore(pos, expected)) - command->to_remove = RemoveProperty::TTL; - else if (!parser_ttl_list.parse(pos, command->ttl, expected)) + if (!parser_ttl_list.parse(pos, command->ttl, expected)) return false; command->type = ASTAlterCommand::MODIFY_TTL; } diff --git a/src/Storages/AlterCommands.cpp b/src/Storages/AlterCommands.cpp index 01b56e0e128..caf98e911ab 100644 --- a/src/Storages/AlterCommands.cpp +++ b/src/Storages/AlterCommands.cpp @@ -111,9 +111,8 @@ std::optional AlterCommand::parse(const ASTAlterCommand * command_ command.type = AlterCommand::MODIFY_COLUMN; const auto & ast_col_decl = command_ast->col_decl->as(); - command.column_name = ast_col_decl.name; - command.to_remove = command_ast->to_remove; + command.column_name = ast_col_decl.name; if (ast_col_decl.type) { command.data_type = data_type_factory.get(ast_col_decl.type); @@ -302,45 +301,24 @@ void AlterCommand::apply(StorageInMemoryMetadata & metadata, const Context & con { metadata.columns.modify(column_name, after_column, first, [&](ColumnDescription & column) { - if (to_remove == RemoveProperty::DEFAULT - || to_remove == RemoveProperty::MATERIALIZED - || to_remove == RemoveProperty::ALIAS) - { - column.default_desc = ColumnDefault{}; - } - else if (to_remove == RemoveProperty::CODEC) - { - column.codec.reset(); - } - else if (to_remove == RemoveProperty::COMMENT) - { - column.comment = String{}; - } - else if (to_remove == RemoveProperty::TTL) - { - column.ttl.reset(); - } - else - { - if (codec) - column.codec = CompressionCodecFactory::instance().validateCodecAndGetPreprocessedAST(codec, data_type ? data_type : column.type, false); + if (codec) + column.codec = CompressionCodecFactory::instance().validateCodecAndGetPreprocessedAST(codec, data_type ? data_type : column.type, false); - if (comment) - column.comment = *comment; + if (comment) + column.comment = *comment; - if (ttl) - column.ttl = ttl; + if (ttl) + column.ttl = ttl; - if (data_type) - column.type = data_type; + if (data_type) + column.type = data_type; - /// User specified default expression or changed - /// datatype. We have to replace default. - if (default_expression || data_type) - { - column.default_desc.kind = default_kind; - column.default_desc.expression = default_expression; - } + /// User specified default expression or changed + /// datatype. We have to replace default. + if (default_expression || data_type) + { + column.default_desc.kind = default_kind; + column.default_desc.expression = default_expression; } }); @@ -470,10 +448,7 @@ void AlterCommand::apply(StorageInMemoryMetadata & metadata, const Context & con } else if (type == MODIFY_TTL) { - if (to_remove == RemoveProperty::TTL) - metadata.table_ttl = TTLTableDescription{}; - else - metadata.table_ttl = TTLTableDescription::getTTLForTableFromAST(ttl, metadata.columns, context, metadata.primary_key); + metadata.table_ttl = TTLTableDescription::getTTLForTableFromAST(ttl, metadata.columns, context, metadata.primary_key); } else if (type == MODIFY_QUERY) { @@ -615,10 +590,6 @@ bool AlterCommand::isRequireMutationStage(const StorageInMemoryMetadata & metada if (type != MODIFY_COLUMN || data_type == nullptr) return false; - /// We remove properties on metadata level - if (type == MODIFY_COLUMN && to_remove != RemoveProperty::NO_PROPERTY) - return false; - for (const auto & column : metadata.columns.getAllPhysical()) { if (column.name == column_name && !isMetadataOnlyConversion(column.type.get(), data_type.get())) @@ -812,30 +783,14 @@ void AlterCommands::prepare(const StorageInMemoryMetadata & metadata) if (!has_column && command.if_exists) command.ignore = true; - if (has_column) + if (has_column && command.data_type) { auto column_from_table = columns.get(command.column_name); - if (command.to_remove != RemoveProperty::NO_PROPERTY) - { - auto column_default = columns.getDefault(command.column_name); - if (!column_default - && (command.to_remove == RemoveProperty::ALIAS || command.to_remove == RemoveProperty::DEFAULT - || command.to_remove == RemoveProperty::MATERIALIZED)) - command.ignore = true; - - if (command.to_remove == RemoveProperty::TTL && column_from_table.ttl == nullptr) - command.ignore = true; - if (command.to_remove == RemoveProperty::COMMENT && column_from_table.comment.empty()) - command.ignore = true; - if (command.to_remove == RemoveProperty::CODEC && column_from_table.codec == nullptr) - command.ignore = true; - } - else if (command.data_type && !command.default_expression && column_from_table.default_desc.expression) + if (!command.default_expression && column_from_table.default_desc.expression) { command.default_kind = column_from_table.default_desc.kind; command.default_expression = column_from_table.default_desc.expression; } - } } else if (command.type == AlterCommand::ADD_COLUMN) @@ -850,11 +805,6 @@ void AlterCommands::prepare(const StorageInMemoryMetadata & metadata) if (!has_column && command.if_exists) command.ignore = true; } - else if (command.type == AlterCommand::MODIFY_TTL) - { - if (!metadata.hasAnyTTL()) - command.ignore = true; - } } prepared = true; } @@ -907,34 +857,6 @@ void AlterCommands::validate(const StorageInMemoryMetadata & metadata, const Con if (command.codec) CompressionCodecFactory::instance().validateCodecAndGetPreprocessedAST(command.codec, command.data_type, !context.getSettingsRef().allow_suspicious_codecs); - auto column_default = all_columns.getDefault(column_name); - if (column_default) - { - if (command.to_remove == RemoveProperty::DEFAULT && column_default->kind != ColumnDefaultKind::Default) - { - throw Exception{ - ErrorCodes::BAD_ARGUMENTS, - "Cannot remove DEFAULT from column {}, because column default type is {}. Use REMOVE {} to delete it.", - backQuote(column_name), toString(column_default->kind), toString(column_default->kind) - }; - } - if (command.to_remove == RemoveProperty::MATERIALIZED && column_default->kind != ColumnDefaultKind::Materialized) - { - throw Exception{ - ErrorCodes::BAD_ARGUMENTS, - "Cannot remove MATERIALIZED from column {}, because column default type is {}. Use REMOVE {} to delete it.", - backQuote(column_name), toString(column_default->kind), toString(column_default->kind) - }; - } - if (command.to_remove == RemoveProperty::ALIAS && column_default->kind != ColumnDefaultKind::Alias) - { - throw Exception{ - ErrorCodes::BAD_ARGUMENTS, - "Cannot remove ALIAS from column {}, because column default type is {}. Use REMOVE {} to delete it.", - backQuote(column_name), toString(column_default->kind), toString(column_default->kind) - }; - } - } modified_columns.emplace(column_name); } @@ -1126,7 +1048,7 @@ MutationCommands AlterCommands::getMutationCommands(StorageInMemoryMetadata meta { for (const auto & alter_cmd : *this) { - if (alter_cmd.isTTLAlter(metadata) && alter_cmd.to_remove != RemoveProperty::TTL) + if (alter_cmd.isTTLAlter(metadata)) { result.push_back(createMaterializeTTLCommand()); break; diff --git a/src/Storages/AlterCommands.h b/src/Storages/AlterCommands.h index a4eff5523b7..3578507a361 100644 --- a/src/Storages/AlterCommands.h +++ b/src/Storages/AlterCommands.h @@ -107,13 +107,16 @@ struct AlterCommand /// Target column name String rename_to; - /// What to remove from column (or TTL) - RemoveProperty to_remove; - static std::optional parse(const ASTAlterCommand * command); void apply(StorageInMemoryMetadata & metadata, const Context & context) const; + /// Checks that alter query changes data. For MergeTree: + /// * column files (data and marks) + /// * each part meta (columns.txt) + /// in each part on disk (it's not lightweight alter). + bool isModifyingData(const StorageInMemoryMetadata & metadata) const; + /// Check that alter command require data modification (mutation) to be /// executed. For example, cast from Date to UInt16 type can be executed /// without any data modifications. But column drop or modify from UInt16 to @@ -161,6 +164,9 @@ public: /// Commands have to be prepared before apply. void apply(StorageInMemoryMetadata & metadata, const Context & context) const; + /// At least one command modify data on disk. + bool isModifyingData(const StorageInMemoryMetadata & metadata) const; + /// At least one command modify settings. bool isSettingsAlter() const; diff --git a/tests/integration/test_adaptive_granularity/test.py b/tests/integration/test_adaptive_granularity/test.py index d5ac91671e0..21d65588de4 100644 --- a/tests/integration/test_adaptive_granularity/test.py +++ b/tests/integration/test_adaptive_granularity/test.py @@ -298,7 +298,7 @@ def test_mixed_granularity_single_node(start_dynamic_cluster, node): #still works assert node.query("SELECT count() from table_with_default_granularity") == '6\n' -@pytest.mark.skip(reason="flaky") + def test_version_update_two_nodes(start_dynamic_cluster): node11.query("INSERT INTO table_with_default_granularity VALUES (toDate('2018-10-01'), 1, 333), (toDate('2018-10-02'), 2, 444)") node12.query("SYSTEM SYNC REPLICA table_with_default_granularity", timeout=20) diff --git a/tests/queries/0_stateless/01493_alter_remove_properties.reference b/tests/queries/0_stateless/01493_alter_remove_properties.reference deleted file mode 100644 index 4ce7a574742..00000000000 --- a/tests/queries/0_stateless/01493_alter_remove_properties.reference +++ /dev/null @@ -1,17 +0,0 @@ -CREATE TABLE default.prop_table\n(\n `column_default` UInt64 DEFAULT 42,\n `column_materialized` UInt64 MATERIALIZED column_default * 42,\n `column_alias` UInt64 ALIAS column_default + 1,\n `column_codec` String CODEC(ZSTD(10)),\n `column_comment` Date COMMENT \'Some comment\',\n `column_ttl` UInt64 TTL column_comment + toIntervalMonth(1)\n)\nENGINE = MergeTree()\nORDER BY tuple()\nTTL column_comment + toIntervalMonth(2)\nSETTINGS index_granularity = 8192 -42 1764 43 str 2019-10-01 1 -CREATE TABLE default.prop_table\n(\n `column_default` UInt64 DEFAULT 42,\n `column_materialized` UInt64 MATERIALIZED column_default * 42,\n `column_alias` UInt64 ALIAS column_default + 1,\n `column_codec` String CODEC(ZSTD(10)),\n `column_comment` Date,\n `column_ttl` UInt64 TTL column_comment + toIntervalMonth(1)\n)\nENGINE = MergeTree()\nORDER BY tuple()\nTTL column_comment + toIntervalMonth(2)\nSETTINGS index_granularity = 8192 -CREATE TABLE default.prop_table\n(\n `column_default` UInt64 DEFAULT 42,\n `column_materialized` UInt64 MATERIALIZED column_default * 42,\n `column_alias` UInt64 ALIAS column_default + 1,\n `column_codec` String,\n `column_comment` Date,\n `column_ttl` UInt64 TTL column_comment + toIntervalMonth(1)\n)\nENGINE = MergeTree()\nORDER BY tuple()\nTTL column_comment + toIntervalMonth(2)\nSETTINGS index_granularity = 8192 -42 1764 0 str 2019-10-01 1 -CREATE TABLE default.prop_table\n(\n `column_default` UInt64 DEFAULT 42,\n `column_materialized` UInt64 MATERIALIZED column_default * 42,\n `column_alias` UInt64,\n `column_codec` String,\n `column_comment` Date,\n `column_ttl` UInt64 TTL column_comment + toIntervalMonth(1)\n)\nENGINE = MergeTree()\nORDER BY tuple()\nTTL column_comment + toIntervalMonth(2)\nSETTINGS index_granularity = 8192 -42 1764 0 str 2019-10-01 1 -42 1764 33 trs 2020-01-01 2 -CREATE TABLE default.prop_table\n(\n `column_default` UInt64 DEFAULT 42,\n `column_materialized` UInt64,\n `column_alias` UInt64,\n `column_codec` String,\n `column_comment` Date,\n `column_ttl` UInt64 TTL column_comment + toIntervalMonth(1)\n)\nENGINE = MergeTree()\nORDER BY tuple()\nTTL column_comment + toIntervalMonth(2)\nSETTINGS index_granularity = 8192 -42 1764 0 str 2019-10-01 1 -42 1764 33 trs 2020-01-01 2 -42 11 44 rts 2020-02-01 3 -CREATE TABLE default.prop_table\n(\n `column_default` UInt64,\n `column_materialized` UInt64,\n `column_alias` UInt64,\n `column_codec` String,\n `column_comment` Date,\n `column_ttl` UInt64 TTL column_comment + toIntervalMonth(1)\n)\nENGINE = MergeTree()\nORDER BY tuple()\nTTL column_comment + toIntervalMonth(2)\nSETTINGS index_granularity = 8192 -42 1764 0 str 2019-10-01 1 -42 1764 33 trs 2020-01-01 2 -42 11 44 rts 2020-02-01 3 -0 22 55 tsr 2020-03-01 4 diff --git a/tests/queries/0_stateless/01493_alter_remove_properties.sql b/tests/queries/0_stateless/01493_alter_remove_properties.sql deleted file mode 100644 index 25000a50235..00000000000 --- a/tests/queries/0_stateless/01493_alter_remove_properties.sql +++ /dev/null @@ -1,58 +0,0 @@ -DROP TABLE IF EXISTS prop_table; - -CREATE TABLE prop_table -( - column_default UInt64 DEFAULT 42, - column_materialized UInt64 MATERIALIZED column_default * 42, - column_alias UInt64 ALIAS column_default + 1, - column_codec String CODEC(ZSTD(10)), - column_comment Date COMMENT 'Some comment', - column_ttl UInt64 TTL column_comment + INTERVAL 1 MONTH -) -ENGINE MergeTree() -ORDER BY tuple() -TTL column_comment + INTERVAL 2 MONTH; - -SHOW CREATE TABLE prop_table; - -SYSTEM STOP TTL MERGES prop_table; - -INSERT INTO prop_table (column_codec, column_comment, column_ttl) VALUES ('str', toDate('2019-10-01'), 1); - -SELECT column_default, column_materialized, column_alias, column_codec, column_comment, column_ttl FROM prop_table; - -ALTER TABLE prop_table MODIFY COLUMN column_comment REMOVE COMMENT; - -SHOW CREATE TABLE prop_table; - -ALTER TABLE prop_table MODIFY COLUMN column_codec REMOVE CODEC; - -SHOW CREATE TABLE prop_table; - -ALTER TABLE prop_table MODIFY COLUMN column_alias REMOVE ALIAS; - -SELECT column_default, column_materialized, column_alias, column_codec, column_comment, column_ttl FROM prop_table; - -SHOW CREATE TABLE prop_table; - -INSERT INTO prop_table (column_alias, column_codec, column_comment, column_ttl) VALUES (33, 'trs', toDate('2020-01-01'), 2); - -SELECT column_default, column_materialized, column_alias, column_codec, column_comment, column_ttl FROM prop_table ORDER BY column_ttl; - -ALTER TABLE prop_table MODIFY COLUMN column_materialized REMOVE MATERIALIZED; - -SHOW CREATE TABLE prop_table; - -INSERT INTO prop_table (column_materialized, column_alias, column_codec, column_comment, column_ttl) VALUES (11, 44, 'rts', toDate('2020-02-01'), 3); - -SELECT column_default, column_materialized, column_alias, column_codec, column_comment, column_ttl FROM prop_table ORDER BY column_ttl; - -ALTER TABLE prop_table MODIFY COLUMN column_default REMOVE DEFAULT; - -SHOW CREATE TABLE prop_table; - -INSERT INTO prop_table (column_materialized, column_alias, column_codec, column_comment, column_ttl) VALUES (22, 55, 'tsr', toDate('2020-03-01'), 4); - -SELECT column_default, column_materialized, column_alias, column_codec, column_comment, column_ttl FROM prop_table ORDER BY column_ttl; - -DROP TABLE IF EXISTS prop_table; From 36019596c1dd2667f1ee0900ba3d75b63a1c82c1 Mon Sep 17 00:00:00 2001 From: alesapin Date: Fri, 11 Sep 2020 20:08:23 +0300 Subject: [PATCH 071/121] Disable flaky test --- tests/integration/test_adaptive_granularity/test.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/integration/test_adaptive_granularity/test.py b/tests/integration/test_adaptive_granularity/test.py index 21d65588de4..d5ac91671e0 100644 --- a/tests/integration/test_adaptive_granularity/test.py +++ b/tests/integration/test_adaptive_granularity/test.py @@ -298,7 +298,7 @@ def test_mixed_granularity_single_node(start_dynamic_cluster, node): #still works assert node.query("SELECT count() from table_with_default_granularity") == '6\n' - +@pytest.mark.skip(reason="flaky") def test_version_update_two_nodes(start_dynamic_cluster): node11.query("INSERT INTO table_with_default_granularity VALUES (toDate('2018-10-01'), 1, 333), (toDate('2018-10-02'), 2, 444)") node12.query("SYSTEM SYNC REPLICA table_with_default_granularity", timeout=20) From e2c2a679ef6a932d0372ca0a6f019bdca64c19e8 Mon Sep 17 00:00:00 2001 From: Vitaly Baranov Date: Fri, 11 Sep 2020 19:54:22 +0300 Subject: [PATCH 072/121] Skip access storages with same path while reading the main config. --- src/Access/AccessControlManager.cpp | 30 +++++++++++++++++- src/Access/DiskAccessStorage.cpp | 31 +++++++++++++------ src/Access/DiskAccessStorage.h | 6 +++- .../configs/duplicates.xml | 13 ++++++++ .../configs/mixed_style.xml | 3 ++ .../integration/test_user_directories/test.py | 16 ++++++---- 6 files changed, 81 insertions(+), 18 deletions(-) create mode 100644 tests/integration/test_user_directories/configs/duplicates.xml diff --git a/src/Access/AccessControlManager.cpp b/src/Access/AccessControlManager.cpp index 1fa26c85354..41137867213 100644 --- a/src/Access/AccessControlManager.cpp +++ b/src/Access/AccessControlManager.cpp @@ -181,6 +181,15 @@ void AccessControlManager::addUsersConfigStorage( const String & preprocessed_dir_, const zkutil::GetZooKeeper & get_zookeeper_function_) { + auto storages = getStoragesPtr(); + for (const auto & storage : *storages) + { + if (auto users_config_storage = typeid_cast>(storage)) + { + if (users_config_storage->getStoragePath() == users_config_path_) + return; + } + } auto check_setting_name_function = [this](const std::string_view & setting_name) { checkSettingNameIsAllowed(setting_name); }; auto new_storage = std::make_shared(storage_name_, check_setting_name_function); new_storage->load(users_config_path_, include_from_path_, preprocessed_dir_, get_zookeeper_function_); @@ -210,17 +219,36 @@ void AccessControlManager::startPeriodicReloadingUsersConfigs() void AccessControlManager::addDiskStorage(const String & directory_, bool readonly_) { - addStorage(std::make_shared(directory_, readonly_)); + addDiskStorage(DiskAccessStorage::STORAGE_TYPE, directory_, readonly_); } void AccessControlManager::addDiskStorage(const String & storage_name_, const String & directory_, bool readonly_) { + auto storages = getStoragesPtr(); + for (const auto & storage : *storages) + { + if (auto disk_storage = typeid_cast>(storage)) + { + if (disk_storage->isStoragePathEqual(directory_)) + { + if (readonly_) + disk_storage->setReadOnly(readonly_); + return; + } + } + } addStorage(std::make_shared(storage_name_, directory_, readonly_)); } void AccessControlManager::addMemoryStorage(const String & storage_name_) { + auto storages = getStoragesPtr(); + for (const auto & storage : *storages) + { + if (auto memory_storage = typeid_cast>(storage)) + return; + } addStorage(std::make_shared(storage_name_)); } diff --git a/src/Access/DiskAccessStorage.cpp b/src/Access/DiskAccessStorage.cpp index fc80859885d..2fcb9480e67 100644 --- a/src/Access/DiskAccessStorage.cpp +++ b/src/Access/DiskAccessStorage.cpp @@ -218,6 +218,16 @@ namespace } + /// Converts a path to an absolute path and append it with a separator. + String makeDirectoryPathCanonical(const String & directory_path) + { + auto canonical_directory_path = std::filesystem::weakly_canonical(directory_path); + if (canonical_directory_path.has_filename()) + canonical_directory_path += std::filesystem::path::preferred_separator; + return canonical_directory_path; + } + + /// Calculates the path to a file named .sql for saving an access entity. String getEntityFilePath(const String & directory_path, const UUID & id) { @@ -298,22 +308,17 @@ DiskAccessStorage::DiskAccessStorage(const String & directory_path_, bool readon { } - DiskAccessStorage::DiskAccessStorage(const String & storage_name_, const String & directory_path_, bool readonly_) : IAccessStorage(storage_name_) { - auto canonical_directory_path = std::filesystem::weakly_canonical(directory_path_); - if (canonical_directory_path.has_filename()) - canonical_directory_path += std::filesystem::path::preferred_separator; + directory_path = makeDirectoryPathCanonical(directory_path_); + readonly = readonly_; std::error_code create_dir_error_code; - std::filesystem::create_directories(canonical_directory_path, create_dir_error_code); + std::filesystem::create_directories(directory_path, create_dir_error_code); - if (!std::filesystem::exists(canonical_directory_path) || !std::filesystem::is_directory(canonical_directory_path) || create_dir_error_code) - throw Exception("Couldn't create directory " + canonical_directory_path.string() + " reason: '" + create_dir_error_code.message() + "'", ErrorCodes::DIRECTORY_DOESNT_EXIST); - - directory_path = canonical_directory_path; - readonly = readonly_; + if (!std::filesystem::exists(directory_path) || !std::filesystem::is_directory(directory_path) || create_dir_error_code) + throw Exception("Couldn't create directory " + directory_path + " reason: '" + create_dir_error_code.message() + "'", ErrorCodes::DIRECTORY_DOESNT_EXIST); bool should_rebuild_lists = std::filesystem::exists(getNeedRebuildListsMarkFilePath(directory_path)); if (!should_rebuild_lists) @@ -337,6 +342,12 @@ DiskAccessStorage::~DiskAccessStorage() } +bool DiskAccessStorage::isStoragePathEqual(const String & directory_path_) const +{ + return getStoragePath() == makeDirectoryPathCanonical(directory_path_); +} + + void DiskAccessStorage::clear() { entries_by_id.clear(); diff --git a/src/Access/DiskAccessStorage.h b/src/Access/DiskAccessStorage.h index 11eb1c3b1ad..f136b046ace 100644 --- a/src/Access/DiskAccessStorage.h +++ b/src/Access/DiskAccessStorage.h @@ -18,7 +18,11 @@ public: ~DiskAccessStorage() override; const char * getStorageType() const override { return STORAGE_TYPE; } + String getStoragePath() const override { return directory_path; } + bool isStoragePathEqual(const String & directory_path_) const; + + void setReadOnly(bool readonly_) { readonly = readonly_; } bool isStorageReadOnly() const override { return readonly; } private: @@ -67,7 +71,7 @@ private: void prepareNotifications(const UUID & id, const Entry & entry, bool remove, Notifications & notifications) const; String directory_path; - bool readonly; + std::atomic readonly; std::unordered_map entries_by_id; std::unordered_map entries_by_name_and_type[static_cast(EntityType::MAX)]; boost::container::flat_set types_of_lists_to_write; diff --git a/tests/integration/test_user_directories/configs/duplicates.xml b/tests/integration/test_user_directories/configs/duplicates.xml new file mode 100644 index 00000000000..69bb06a112b --- /dev/null +++ b/tests/integration/test_user_directories/configs/duplicates.xml @@ -0,0 +1,13 @@ + + + + /var/lib/clickhouse/access7/ + + + /etc/clickhouse-server/users7.xml + + + + /etc/clickhouse-server/users7.xml + /var/lib/clickhouse/access7/ + diff --git a/tests/integration/test_user_directories/configs/mixed_style.xml b/tests/integration/test_user_directories/configs/mixed_style.xml index d6ddecf6f5d..f668140521a 100644 --- a/tests/integration/test_user_directories/configs/mixed_style.xml +++ b/tests/integration/test_user_directories/configs/mixed_style.xml @@ -1,5 +1,8 @@ + + /var/lib/clickhouse/access6a/ + diff --git a/tests/integration/test_user_directories/test.py b/tests/integration/test_user_directories/test.py index 218330cb1a5..71745502064 100644 --- a/tests/integration/test_user_directories/test.py +++ b/tests/integration/test_user_directories/test.py @@ -12,11 +12,8 @@ def started_cluster(): try: cluster.start() - node.exec_in_container("cp /etc/clickhouse-server/users.xml /etc/clickhouse-server/users2.xml") - node.exec_in_container("cp /etc/clickhouse-server/users.xml /etc/clickhouse-server/users3.xml") - node.exec_in_container("cp /etc/clickhouse-server/users.xml /etc/clickhouse-server/users4.xml") - node.exec_in_container("cp /etc/clickhouse-server/users.xml /etc/clickhouse-server/users5.xml") - node.exec_in_container("cp /etc/clickhouse-server/users.xml /etc/clickhouse-server/users6.xml") + for i in range(2, 8): + node.exec_in_container("cp /etc/clickhouse-server/users.xml /etc/clickhouse-server/users{}.xml".format(i)) yield cluster @@ -56,4 +53,11 @@ def test_mixed_style(): node.restart_clickhouse() assert node.query("SELECT * FROM system.user_directories") == TSV([["users.xml", "users.xml", "/etc/clickhouse-server/users6.xml", 1, 1], ["local directory", "local directory", "/var/lib/clickhouse/access6/", 0, 2], - ["memory", "memory", "", 0, 3]]) + ["local directory", "local directory", "/var/lib/clickhouse/access6a/", 0, 3], + ["memory", "memory", "", 0, 4]]) + +def test_duplicates(): + node.copy_file_to_container(os.path.join(SCRIPT_DIR, "configs/duplicates.xml"), '/etc/clickhouse-server/config.d/z.xml') + node.restart_clickhouse() + assert node.query("SELECT * FROM system.user_directories") == TSV([["users.xml", "users.xml", "/etc/clickhouse-server/users7.xml", 1, 1], + ["local directory", "local directory", "/var/lib/clickhouse/access7/", 0, 2]]) From 1e849f297549f90bd7671286cace24f36c14e801 Mon Sep 17 00:00:00 2001 From: Vitaly Baranov Date: Sat, 12 Sep 2020 03:16:50 +0300 Subject: [PATCH 073/121] Fix permission denied on opening file /var/lib/clickhouse/status in integration tests. --- tests/integration/helpers/cluster.py | 1 + 1 file changed, 1 insertion(+) diff --git a/tests/integration/helpers/cluster.py b/tests/integration/helpers/cluster.py index 44a22d3fe2e..a8704ee42b1 100644 --- a/tests/integration/helpers/cluster.py +++ b/tests/integration/helpers/cluster.py @@ -1165,6 +1165,7 @@ class ClickHouseInstance: db_dir = p.abspath(p.join(self.path, 'database')) print "Setup database dir {}".format(db_dir) + os.mkdir(db_dir) if self.clickhouse_path_dir is not None: print "Database files taken from {}".format(self.clickhouse_path_dir) shutil.copytree(self.clickhouse_path_dir, db_dir) From e12ae99bf7e4b717b30179ac1a65920954cb3656 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Sat, 12 Sep 2020 03:55:54 +0300 Subject: [PATCH 074/121] Added review suggestion --- programs/git-import/git-import.cpp | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/programs/git-import/git-import.cpp b/programs/git-import/git-import.cpp index d314969a1a8..6e92b88734d 100644 --- a/programs/git-import/git-import.cpp +++ b/programs/git-import/git-import.cpp @@ -335,7 +335,7 @@ struct LineChange */ void setLineInfo(std::string full_line) { - indent = 0; + uint32_t num_spaces = 0; const char * pos = full_line.data(); const char * end = pos + full_line.size(); @@ -343,14 +343,15 @@ struct LineChange while (pos < end) { if (*pos == ' ') - ++indent; + ++num_spaces; else if (*pos == '\t') - indent += 4; + num_spaces += 4; else break; ++pos; } + indent = std::max(255U, num_spaces); line.assign(pos, end); if (pos == end) From 5b952a369bec6ec4696e5e202e0b5ddedf6be72f Mon Sep 17 00:00:00 2001 From: zhang2014 Date: Sat, 12 Sep 2020 12:07:02 +0800 Subject: [PATCH 075/121] Fix build failure in OSX --- src/Functions/GatherUtils/CMakeLists.txt | 11 +++++++++++ src/Functions/GatherUtils/Sources.h | 4 ++++ 2 files changed, 15 insertions(+) diff --git a/src/Functions/GatherUtils/CMakeLists.txt b/src/Functions/GatherUtils/CMakeLists.txt index 3f7f08621a1..f941091667e 100644 --- a/src/Functions/GatherUtils/CMakeLists.txt +++ b/src/Functions/GatherUtils/CMakeLists.txt @@ -3,6 +3,17 @@ add_headers_and_sources(clickhouse_functions_gatherutils .) add_library(clickhouse_functions_gatherutils ${clickhouse_functions_gatherutils_sources} ${clickhouse_functions_gatherutils_headers}) target_link_libraries(clickhouse_functions_gatherutils PRIVATE dbms) +check_cxx_compiler_flag(suggest-override HAS_SUGGEST_OVERRIDE) +check_cxx_compiler_flag(suggest-destructor-override HAS_SUGGEST_DESTRUCTOR_OVERRIDE) + +if (HAS_SUGGEST_OVERRIDE) + target_compile_definitions(clickhouse_functions_gatherutils PRIVATE HAS_SUGGEST_OVERRIDE) +endif() + +if (HAS_SUGGEST_DESTRUCTOR_OVERRIDE) + target_compile_definitions(clickhouse_functions_gatherutils PRIVATE HAS_SUGGEST_DESTRUCTOR_OVERRIDE) +endif() + if (STRIP_DEBUG_SYMBOLS_FUNCTIONS) target_compile_options(clickhouse_functions_gatherutils PRIVATE "-g0") endif() diff --git a/src/Functions/GatherUtils/Sources.h b/src/Functions/GatherUtils/Sources.h index 299884e1c9e..fe71a1f8be3 100644 --- a/src/Functions/GatherUtils/Sources.h +++ b/src/Functions/GatherUtils/Sources.h @@ -129,9 +129,13 @@ struct NumericArraySource : public ArraySourceImpl> #pragma GCC diagnostic ignored "-Wsuggest-override" #elif __clang_major__ >= 11 #pragma GCC diagnostic push +#ifdef HAS_SUGGEST_OVERRIDE #pragma GCC diagnostic ignored "-Wsuggest-override" +#endif +#ifdef HAS_SUGGEST_DESTRUCTOR_OVERRIDE #pragma GCC diagnostic ignored "-Wsuggest-destructor-override" #endif +#endif template struct ConstSource : public Base From ecbcbad0d96f3d7173d535f9bb181f6104e67ff7 Mon Sep 17 00:00:00 2001 From: alesapin Date: Sat, 12 Sep 2020 10:07:08 +0300 Subject: [PATCH 076/121] Fix flaky test --- .../01465_ttl_recompression.reference | 42 +++++++++---------- .../0_stateless/01465_ttl_recompression.sql | 19 +++++---- 2 files changed, 32 insertions(+), 29 deletions(-) diff --git a/tests/queries/0_stateless/01465_ttl_recompression.reference b/tests/queries/0_stateless/01465_ttl_recompression.reference index 524c44ef972..1c576c04e45 100644 --- a/tests/queries/0_stateless/01465_ttl_recompression.reference +++ b/tests/queries/0_stateless/01465_ttl_recompression.reference @@ -1,24 +1,24 @@ CREATE TABLE default.recompression_table\n(\n `dt` DateTime,\n `key` UInt64,\n `value` String\n)\nENGINE = MergeTree()\nPARTITION BY key\nORDER BY tuple()\nTTL dt + toIntervalMonth(1) RECOMPRESS CODEC(ZSTD(17)), dt + toIntervalYear(1) RECOMPRESS CODEC(LZ4HC(10))\nSETTINGS min_rows_for_wide_part = 0, min_bytes_for_wide_part = 0, index_granularity = 8192 3000 -1_1_1_0 LZ4 -2_2_2_0 LZ4 -3_3_3_0 LZ4 -1_1_1_1 LZ4 -2_2_2_1 ZSTD(17) -3_3_3_1 LZ4HC(10) +1_1_1 LZ4 +2_2_2 LZ4 +3_3_3 LZ4 +1_1_1 LZ4 +2_2_2 ZSTD(17) +3_3_3 LZ4HC(10) CREATE TABLE default.recompression_table\n(\n `dt` DateTime,\n `key` UInt64,\n `value` String\n)\nENGINE = MergeTree()\nPARTITION BY key\nORDER BY tuple()\nTTL dt + toIntervalDay(1) RECOMPRESS CODEC(ZSTD(12))\nSETTINGS min_rows_for_wide_part = 0, min_bytes_for_wide_part = 0, index_granularity = 8192 -1_1_1_1_4 LZ4 -2_2_2_1_4 ZSTD(17) -3_3_3_1_4 LZ4HC(10) -1_1_1_2_4 LZ4 -2_2_2_2_4 ZSTD(12) -3_3_3_2_4 ZSTD(12) -1_1_1_2_4 ['plus(dt, toIntervalDay(1))'] -2_2_2_2_4 ['plus(dt, toIntervalDay(1))'] -3_3_3_2_4 ['plus(dt, toIntervalDay(1))'] -1_1_1_0 LZ4 -2_2_2_0 LZ4 -3_3_3_0 LZ4 -1_1_1_0_4 LZ4 -2_2_2_0_4 ZSTD(12) -3_3_3_0_4 ZSTD(12) +1_1_1 LZ4 +2_2_2 ZSTD(17) +3_3_3 LZ4HC(10) +1_1_1 LZ4 +2_2_2 ZSTD(12) +3_3_3 ZSTD(12) +1_1_1 ['plus(dt, toIntervalDay(1))'] +2_2_2 ['plus(dt, toIntervalDay(1))'] +3_3_3 ['plus(dt, toIntervalDay(1))'] +1_1_1 LZ4 +2_2_2 LZ4 +3_3_3 LZ4 +1_1_1 LZ4 +2_2_2 ZSTD(12) +3_3_3 ZSTD(12) diff --git a/tests/queries/0_stateless/01465_ttl_recompression.sql b/tests/queries/0_stateless/01465_ttl_recompression.sql index 78550582307..2388e727722 100644 --- a/tests/queries/0_stateless/01465_ttl_recompression.sql +++ b/tests/queries/0_stateless/01465_ttl_recompression.sql @@ -24,25 +24,27 @@ INSERT INTO recompression_table SELECT now() - INTERVAL 2 YEAR, 3, toString(numb SELECT COUNT() FROM recompression_table; -SELECT name, default_compression_codec FROM system.parts WHERE table = 'recompression_table' and active = 1 and database = currentDatabase() ORDER BY name; +SELECT substring(name, 1, length(name) - 2), default_compression_codec FROM system.parts WHERE table = 'recompression_table' and active = 1 and database = currentDatabase() ORDER BY name; OPTIMIZE TABLE recompression_table FINAL; -SELECT name, default_compression_codec FROM system.parts WHERE table = 'recompression_table' and active = 1 and database = currentDatabase() ORDER BY name; +-- merge level and mutation in part name is not important +SELECT substring(name, 1, length(name) - 2), default_compression_codec FROM system.parts WHERE table = 'recompression_table' and active = 1 and database = currentDatabase() ORDER BY name; ALTER TABLE recompression_table MODIFY TTL dt + INTERVAL 1 DAY RECOMPRESS CODEC(ZSTD(12)) SETTINGS mutations_sync = 2; SHOW CREATE TABLE recompression_table; -SELECT name, default_compression_codec FROM system.parts WHERE table = 'recompression_table' and active = 1 and database = currentDatabase() ORDER BY name; +SELECT substring(name, 1, length(name) - 4), default_compression_codec FROM system.parts WHERE table = 'recompression_table' and active = 1 and database = currentDatabase() ORDER BY name; SYSTEM START TTL MERGES recompression_table; - +-- Additional merge can happen here OPTIMIZE TABLE recompression_table FINAL; -SELECT name, default_compression_codec FROM system.parts WHERE table = 'recompression_table' and active = 1 and database = currentDatabase() ORDER BY name; +-- merge level and mutation in part name is not important +SELECT substring(name, 1, length(name) - 4), default_compression_codec FROM system.parts WHERE table = 'recompression_table' and active = 1 and database = currentDatabase() ORDER BY name; -SELECT name, recompression_ttl_info.expression FROM system.parts WHERE table = 'recompression_table' and active = 1 and database = currentDatabase() ORDER BY name; +SELECT substring(name, 1, length(name) - 4), recompression_ttl_info.expression FROM system.parts WHERE table = 'recompression_table' and active = 1 and database = currentDatabase() ORDER BY name; DROP TABLE IF EXISTS recompression_table; @@ -66,10 +68,11 @@ INSERT INTO recompression_table_compact SELECT now() - INTERVAL 2 MONTH, 2, toSt INSERT INTO recompression_table_compact SELECT now() - INTERVAL 2 YEAR, 3, toString(number) from numbers(2000, 1000); -SELECT name, default_compression_codec FROM system.parts WHERE table = 'recompression_table_compact' and active = 1 and database = currentDatabase() ORDER BY name; +SELECT substring(name, 1, length(name) - 2), default_compression_codec FROM system.parts WHERE table = 'recompression_table_compact' and active = 1 and database = currentDatabase() ORDER BY name; ALTER TABLE recompression_table_compact MODIFY TTL dt + INTERVAL 1 MONTH RECOMPRESS CODEC(ZSTD(12)) SETTINGS mutations_sync = 2; -- mutation affect all columns, so codec changes -SELECT name, default_compression_codec FROM system.parts WHERE table = 'recompression_table_compact' and active = 1 and database = currentDatabase() ORDER BY name; +-- merge level and mutation in part name is not important +SELECT substring(name, 1, length(name) - 4), default_compression_codec FROM system.parts WHERE table = 'recompression_table_compact' and active = 1 and database = currentDatabase() ORDER BY name; DROP TABLE recompression_table_compact; From 8242a948804622f71eeaba1ad91a6e1cd14ab683 Mon Sep 17 00:00:00 2001 From: alesapin Date: Sat, 12 Sep 2020 15:42:07 +0300 Subject: [PATCH 077/121] Update ci_config.json --- tests/ci/ci_config.json | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/ci/ci_config.json b/tests/ci/ci_config.json index 9a11a06db0d..504b554029b 100644 --- a/tests/ci/ci_config.json +++ b/tests/ci/ci_config.json @@ -323,7 +323,7 @@ }, "Functional stateless tests (unbundled)": { "required_build_properties": { - "compiler": "gcc-10", + "compiler": "gcc-9", "package_type": "deb", "build_type": "relwithdebuginfo", "sanitizer": "none", From 8075ce28099ea34f26209ab5eba7c8eb9bc603b2 Mon Sep 17 00:00:00 2001 From: alesapin Date: Sat, 12 Sep 2020 15:42:32 +0300 Subject: [PATCH 078/121] Update warnings.cmake --- cmake/warnings.cmake | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/cmake/warnings.cmake b/cmake/warnings.cmake index 6b26b9b95a5..425972f00d8 100644 --- a/cmake/warnings.cmake +++ b/cmake/warnings.cmake @@ -23,7 +23,7 @@ option (WEVERYTHING "Enables -Weverything option with some exceptions. This is i # Control maximum size of stack frames. It can be important if the code is run in fibers with small stack size. # Only in release build because debug has too large stack frames. if ((NOT CMAKE_BUILD_TYPE_UC STREQUAL "DEBUG") AND (NOT SANITIZE)) - add_warning(frame-larger-than=32768) + add_warning(frame-larger-than=16384) endif () if (COMPILER_CLANG) From 421eeeccef7622f8f1462f9bce87303d51b880be Mon Sep 17 00:00:00 2001 From: Vitaly Baranov Date: Mon, 17 Aug 2020 17:38:10 +0300 Subject: [PATCH 079/121] Add the section user_directories to the default config. --- programs/server/config.xml | 16 +++++++++++----- .../helpers/0_common_instance_config.xml | 3 +++ 2 files changed, 14 insertions(+), 5 deletions(-) diff --git a/programs/server/config.xml b/programs/server/config.xml index af01e880dc2..3d7ebf0cd96 100644 --- a/programs/server/config.xml +++ b/programs/server/config.xml @@ -212,8 +212,17 @@ /var/lib/clickhouse/user_files/ - - /var/lib/clickhouse/access/ + + + + + users.xml + + + + /var/lib/clickhouse/access/ + + @@ -256,9 +265,6 @@ --> - - users.xml - default diff --git a/tests/integration/helpers/0_common_instance_config.xml b/tests/integration/helpers/0_common_instance_config.xml index 5377efbc241..b27ecf0c3ef 100644 --- a/tests/integration/helpers/0_common_instance_config.xml +++ b/tests/integration/helpers/0_common_instance_config.xml @@ -4,4 +4,7 @@ custom_ /var/lib/clickhouse/ /var/lib/clickhouse/tmp/ + + + users.xml From c2d79bc5ccb04aeef881379797c05d57e290782b Mon Sep 17 00:00:00 2001 From: Amos Bird Date: Thu, 10 Sep 2020 22:56:15 +0800 Subject: [PATCH 080/121] Add merge_algorithm to system.merges --- docs/en/operations/system-tables/merges.md | 6 ++++- src/Storages/MergeTree/MergeAlgorithm.cpp | 26 +++++++++++++++++++ src/Storages/MergeTree/MergeAlgorithm.h | 17 ++++++++++++ src/Storages/MergeTree/MergeList.cpp | 2 ++ src/Storages/MergeTree/MergeList.h | 3 +++ .../MergeTree/MergeTreeDataMergerMutator.cpp | 7 ++--- .../MergeTree/MergeTreeDataMergerMutator.h | 7 +---- src/Storages/System/StorageSystemMerges.cpp | 7 +++++ src/Storages/ya.make | 1 + 9 files changed, 64 insertions(+), 12 deletions(-) create mode 100644 src/Storages/MergeTree/MergeAlgorithm.cpp create mode 100644 src/Storages/MergeTree/MergeAlgorithm.h diff --git a/docs/en/operations/system-tables/merges.md b/docs/en/operations/system-tables/merges.md index fb98a2b9e34..3e712e2962c 100644 --- a/docs/en/operations/system-tables/merges.md +++ b/docs/en/operations/system-tables/merges.md @@ -10,12 +10,16 @@ Columns: - `progress` (Float64) — The percentage of completed work from 0 to 1. - `num_parts` (UInt64) — The number of pieces to be merged. - `result_part_name` (String) — The name of the part that will be formed as the result of merging. -- `is_mutation` (UInt8) - 1 if this process is a part mutation. +- `is_mutation` (UInt8) — 1 if this process is a part mutation. - `total_size_bytes_compressed` (UInt64) — The total size of the compressed data in the merged chunks. - `total_size_marks` (UInt64) — The total number of marks in the merged parts. - `bytes_read_uncompressed` (UInt64) — Number of bytes read, uncompressed. - `rows_read` (UInt64) — Number of rows read. - `bytes_written_uncompressed` (UInt64) — Number of bytes written, uncompressed. - `rows_written` (UInt64) — Number of rows written. +- `memory_usage` (UInt64) — Memory consumption of the merge process. +- `thread_id` (UInt64) — Thread ID of the merge process. +- `merge_type` — The type of current merge. Empty if it's an mutation. +- `merge_algorithm` — The algorithm used in current merge. Empty if it's an mutation. [Original article](https://clickhouse.tech/docs/en/operations/system_tables/merges) diff --git a/src/Storages/MergeTree/MergeAlgorithm.cpp b/src/Storages/MergeTree/MergeAlgorithm.cpp new file mode 100644 index 00000000000..9f73557e701 --- /dev/null +++ b/src/Storages/MergeTree/MergeAlgorithm.cpp @@ -0,0 +1,26 @@ +#include +#include + +namespace DB +{ +namespace ErrorCodes +{ + extern const int NOT_IMPLEMENTED; +} + +String toString(MergeAlgorithm merge_algorithm) +{ + switch (merge_algorithm) + { + case MergeAlgorithm::Undecided: + return "Undecided"; + case MergeAlgorithm::Horizontal: + return "Horizontal"; + case MergeAlgorithm::Vertical: + return "Vertical"; + } + + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Unknown MergeAlgorithm {}", static_cast(merge_algorithm)); +} + +} diff --git a/src/Storages/MergeTree/MergeAlgorithm.h b/src/Storages/MergeTree/MergeAlgorithm.h new file mode 100644 index 00000000000..813767f9fb1 --- /dev/null +++ b/src/Storages/MergeTree/MergeAlgorithm.h @@ -0,0 +1,17 @@ +#pragma once + +#include + +namespace DB +{ +/// Algorithm of Merge. +enum class MergeAlgorithm +{ + Undecided, /// Not running yet + Horizontal, /// per-row merge of all columns + Vertical /// per-row merge of PK and secondary indices columns, per-column gather for non-PK columns +}; + +String toString(MergeAlgorithm merge_algorithm); + +} diff --git a/src/Storages/MergeTree/MergeList.cpp b/src/Storages/MergeTree/MergeList.cpp index 30324bd5d9e..05d4cc6f963 100644 --- a/src/Storages/MergeTree/MergeList.cpp +++ b/src/Storages/MergeTree/MergeList.cpp @@ -24,6 +24,7 @@ MergeListElement::MergeListElement(const std::string & database_, const std::str , num_parts{future_part.parts.size()} , thread_id{getThreadId()} , merge_type{future_part.merge_type} + , merge_algorithm{MergeAlgorithm::Undecided} { for (const auto & source_part : future_part.parts) { @@ -74,6 +75,7 @@ MergeInfo MergeListElement::getInfo() const res.memory_usage = memory_tracker.get(); res.thread_id = thread_id; res.merge_type = toString(merge_type); + res.merge_algorithm = toString(merge_algorithm); for (const auto & source_part_name : source_part_names) res.source_part_names.emplace_back(source_part_name); diff --git a/src/Storages/MergeTree/MergeList.h b/src/Storages/MergeTree/MergeList.h index 4d080ff3569..c1166c55703 100644 --- a/src/Storages/MergeTree/MergeList.h +++ b/src/Storages/MergeTree/MergeList.h @@ -3,6 +3,7 @@ #include #include #include +#include #include #include #include @@ -47,6 +48,7 @@ struct MergeInfo UInt64 memory_usage; UInt64 thread_id; std::string merge_type; + std::string merge_algorithm; }; struct FutureMergedMutatedPart; @@ -90,6 +92,7 @@ struct MergeListElement : boost::noncopyable UInt64 thread_id; MergeType merge_type; + MergeAlgorithm merge_algorithm; MergeListElement(const std::string & database, const std::string & table, const FutureMergedMutatedPart & future_part); diff --git a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp index a8f7e265f68..99be79390be 100644 --- a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp +++ b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp @@ -62,10 +62,6 @@ namespace ErrorCodes extern const int ABORTED; } - -using MergeAlgorithm = MergeTreeDataMergerMutator::MergeAlgorithm; - - /// Do not start to merge parts, if free space is less than sum size of parts times specified coefficient. /// This value is chosen to not allow big merges to eat all free space. Thus allowing small merges to proceed. static const double DISK_USAGE_COEFFICIENT_TO_SELECT = 2; @@ -699,6 +695,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mergePartsToTempor size_t sum_input_rows_upper_bound = merge_entry->total_rows_count; MergeAlgorithm merge_alg = chooseMergeAlgorithm(parts, sum_input_rows_upper_bound, gathering_columns, deduplicate, need_remove_expired_values); + merge_entry->merge_algorithm = merge_alg; LOG_DEBUG(log, "Selected MergeAlgorithm: {}", ((merge_alg == MergeAlgorithm::Vertical) ? "Vertical" : "Horizontal")); @@ -1238,7 +1235,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeDataMergerMutator::mutatePartToTempor } -MergeTreeDataMergerMutator::MergeAlgorithm MergeTreeDataMergerMutator::chooseMergeAlgorithm( +MergeAlgorithm MergeTreeDataMergerMutator::chooseMergeAlgorithm( const MergeTreeData::DataPartsVector & parts, size_t sum_rows_upper_bound, const NamesAndTypesList & gathering_columns, bool deduplicate, bool need_remove_expired_values) const { diff --git a/src/Storages/MergeTree/MergeTreeDataMergerMutator.h b/src/Storages/MergeTree/MergeTreeDataMergerMutator.h index 96ab14ba57b..2ba6b04e082 100644 --- a/src/Storages/MergeTree/MergeTreeDataMergerMutator.h +++ b/src/Storages/MergeTree/MergeTreeDataMergerMutator.h @@ -6,6 +6,7 @@ #include #include #include +#include #include @@ -226,12 +227,6 @@ public : ActionBlocker merges_blocker; ActionBlocker ttl_merges_blocker; - enum class MergeAlgorithm - { - Horizontal, /// per-row merge of all columns - Vertical /// per-row merge of PK and secondary indices columns, per-column gather for non-PK columns - }; - private: MergeAlgorithm chooseMergeAlgorithm( diff --git a/src/Storages/System/StorageSystemMerges.cpp b/src/Storages/System/StorageSystemMerges.cpp index 3b9e39c1ef8..b61324818e4 100644 --- a/src/Storages/System/StorageSystemMerges.cpp +++ b/src/Storages/System/StorageSystemMerges.cpp @@ -31,6 +31,7 @@ NamesAndTypesList StorageSystemMerges::getNamesAndTypes() {"memory_usage", std::make_shared()}, {"thread_id", std::make_shared()}, {"merge_type", std::make_shared()}, + {"merge_algorithm", std::make_shared()}, }; } @@ -67,9 +68,15 @@ void StorageSystemMerges::fillData(MutableColumns & res_columns, const Context & res_columns[i++]->insert(merge.memory_usage); res_columns[i++]->insert(merge.thread_id); if (!merge.is_mutation) + { res_columns[i++]->insert(merge.merge_type); + res_columns[i++]->insert(merge.merge_algorithm); + } else + { res_columns[i++]->insertDefault(); + res_columns[i++]->insertDefault(); + } } } diff --git a/src/Storages/ya.make b/src/Storages/ya.make index 597e0c6f975..20377428857 100644 --- a/src/Storages/ya.make +++ b/src/Storages/ya.make @@ -36,6 +36,7 @@ SRCS( MergeTree/KeyCondition.cpp MergeTree/LevelMergeSelector.cpp MergeTree/localBackup.cpp + MergeTree/MergeAlgorithm.cpp MergeTree/MergedBlockOutputStream.cpp MergeTree/MergedColumnOnlyOutputStream.cpp MergeTree/MergeList.cpp From 016f707ea133f323ffd135a91ac86959112c6a8e Mon Sep 17 00:00:00 2001 From: Amos Bird Date: Fri, 4 Sep 2020 01:51:16 +0800 Subject: [PATCH 081/121] column transformers in insert select --- src/Interpreters/InterpreterInsertQuery.cpp | 26 ++++++++++++++++++- src/Parsers/ParserInsertQuery.cpp | 9 ++++++- src/Parsers/ParserInsertQuery.h | 9 +++++++ ...1470_test_insert_select_asterisk.reference | 6 +++++ .../01470_test_insert_select_asterisk.sql | 18 +++++++++++++ 5 files changed, 66 insertions(+), 2 deletions(-) create mode 100644 tests/queries/0_stateless/01470_test_insert_select_asterisk.reference create mode 100644 tests/queries/0_stateless/01470_test_insert_select_asterisk.sql diff --git a/src/Interpreters/InterpreterInsertQuery.cpp b/src/Interpreters/InterpreterInsertQuery.cpp index 9d33650405a..01fee30a445 100644 --- a/src/Interpreters/InterpreterInsertQuery.cpp +++ b/src/Interpreters/InterpreterInsertQuery.cpp @@ -17,6 +17,7 @@ #include #include #include +#include #include #include #include @@ -29,6 +30,8 @@ #include #include #include +#include +#include namespace { @@ -90,9 +93,30 @@ Block InterpreterInsertQuery::getSampleBlock( } Block table_sample = metadata_snapshot->getSampleBlock(); + const auto & columns = metadata_snapshot->getColumns(); + auto names_and_types = columns.getOrdinary(); + removeDuplicateColumns(names_and_types); + auto table_expr = std::make_shared(); + table_expr->database_and_table_name = createTableIdentifier(table->getStorageID()); + table_expr->children.push_back(table_expr->database_and_table_name); + TablesWithColumns tables_with_columns; + tables_with_columns.emplace_back(DatabaseAndTableWithAlias(*table_expr, context.getCurrentDatabase()), names_and_types); + + tables_with_columns[0].addHiddenColumns(columns.getMaterialized()); + tables_with_columns[0].addHiddenColumns(columns.getAliases()); + tables_with_columns[0].addHiddenColumns(table->getVirtuals()); + + NameSet source_columns_set; + for (const auto & identifier : query.columns->children) + source_columns_set.insert(identifier->getColumnName()); + TranslateQualifiedNamesVisitor::Data visitor_data(source_columns_set, tables_with_columns); + TranslateQualifiedNamesVisitor visitor(visitor_data); + auto columns_ast = query.columns->clone(); + visitor.visit(columns_ast); + /// Form the block based on the column names from the query Block res; - for (const auto & identifier : query.columns->children) + for (const auto & identifier : columns_ast->children) { std::string current_name = identifier->getColumnName(); diff --git a/src/Parsers/ParserInsertQuery.cpp b/src/Parsers/ParserInsertQuery.cpp index dc25954c71f..50baf7566d1 100644 --- a/src/Parsers/ParserInsertQuery.cpp +++ b/src/Parsers/ParserInsertQuery.cpp @@ -36,7 +36,7 @@ bool ParserInsertQuery::parseImpl(Pos & pos, ASTPtr & node, Expected & expected) ParserToken s_lparen(TokenType::OpeningRoundBracket); ParserToken s_rparen(TokenType::ClosingRoundBracket); ParserIdentifier name_p; - ParserList columns_p(std::make_unique(), std::make_unique(TokenType::Comma), false); + ParserList columns_p(std::make_unique(), std::make_unique(TokenType::Comma), false); ParserFunction table_function_p{false}; ASTPtr database; @@ -189,5 +189,12 @@ bool ParserInsertQuery::parseImpl(Pos & pos, ASTPtr & node, Expected & expected) return true; } +bool ParserInsertElement::parseImpl(Pos & pos, ASTPtr & node, Expected & expected) +{ + return ParserColumnsMatcher().parse(pos, node, expected) + || ParserQualifiedAsterisk().parse(pos, node, expected) + || ParserAsterisk().parse(pos, node, expected) + || ParserCompoundIdentifier().parse(pos, node, expected); +} } diff --git a/src/Parsers/ParserInsertQuery.h b/src/Parsers/ParserInsertQuery.h index b69bc645c15..b6a199c9d71 100644 --- a/src/Parsers/ParserInsertQuery.h +++ b/src/Parsers/ParserInsertQuery.h @@ -33,4 +33,13 @@ public: ParserInsertQuery(const char * end_) : end(end_) {} }; +/** Insert accepts an identifier and an asterisk with variants. + */ +class ParserInsertElement : public IParserBase +{ +protected: + const char * getName() const override { return "insert element"; } + bool parseImpl(Pos & pos, ASTPtr & node, Expected & expected) override; +}; + } diff --git a/tests/queries/0_stateless/01470_test_insert_select_asterisk.reference b/tests/queries/0_stateless/01470_test_insert_select_asterisk.reference new file mode 100644 index 00000000000..c5d97af6937 --- /dev/null +++ b/tests/queries/0_stateless/01470_test_insert_select_asterisk.reference @@ -0,0 +1,6 @@ +1 0 0 2 +3 0 0 4 +1 0 0 2 +3 0 0 4 +1 0 0 2 +3 0 0 4 diff --git a/tests/queries/0_stateless/01470_test_insert_select_asterisk.sql b/tests/queries/0_stateless/01470_test_insert_select_asterisk.sql new file mode 100644 index 00000000000..607b8a25f82 --- /dev/null +++ b/tests/queries/0_stateless/01470_test_insert_select_asterisk.sql @@ -0,0 +1,18 @@ +DROP TABLE IF EXISTS insert_select_dst; +DROP TABLE IF EXISTS insert_select_src; + +CREATE TABLE insert_select_dst (i int, middle_a int, middle_b int, j int) ENGINE = Log; + +CREATE TABLE insert_select_src (i int, j int) ENGINE = Log; + +INSERT INTO insert_select_src VALUES (1, 2), (3, 4); + +INSERT INTO insert_select_dst(* EXCEPT (middle_a, middle_b)) SELECT * FROM insert_select_src; +INSERT INTO insert_select_dst(insert_select_dst.* EXCEPT (middle_a, middle_b)) SELECT * FROM insert_select_src; +INSERT INTO insert_select_dst(COLUMNS('.*') EXCEPT (middle_a, middle_b)) SELECT * FROM insert_select_src; +INSERT INTO insert_select_dst(insert_select_src.* EXCEPT (middle_a, middle_b)) SELECT * FROM insert_select_src; -- { serverError 47 } + +SELECT * FROM insert_select_dst; + +DROP TABLE IF EXISTS insert_select_dst; +DROP TABLE IF EXISTS insert_select_src; From 34b9547ce1e51c729489f9555d6a60c8c8b7b078 Mon Sep 17 00:00:00 2001 From: Amos Bird Date: Sat, 5 Sep 2020 22:12:47 +0800 Subject: [PATCH 082/121] Binary operator monotonicity --- src/Functions/FunctionBinaryArithmetic.h | 186 +++++++++++++++++- src/Functions/bitAnd.cpp | 2 +- src/Functions/bitBoolMaskAnd.cpp | 2 +- src/Functions/bitBoolMaskOr.cpp | 2 +- src/Functions/bitOr.cpp | 2 +- src/Functions/bitRotateLeft.cpp | 2 +- src/Functions/bitRotateRight.cpp | 2 +- src/Functions/bitShiftLeft.cpp | 2 +- src/Functions/bitShiftRight.cpp | 2 +- src/Functions/bitTest.cpp | 2 +- src/Functions/bitXor.cpp | 2 +- src/Functions/divide.cpp | 2 +- src/Functions/gcd.cpp | 2 +- src/Functions/intDiv.cpp | 2 +- src/Functions/intDivOrZero.cpp | 2 +- src/Functions/lcm.cpp | 2 +- src/Functions/minus.cpp | 2 +- src/Functions/modulo.cpp | 2 +- src/Functions/moduloOrZero.cpp | 2 +- src/Functions/multiply.cpp | 2 +- src/Functions/plus.cpp | 2 +- src/Storages/MergeTree/KeyCondition.cpp | 42 +++- ...480_binary_operator_monotonicity.reference | 0 .../01480_binary_operator_monotonicity.sql | 10 + 24 files changed, 247 insertions(+), 31 deletions(-) create mode 100644 tests/queries/0_stateless/01480_binary_operator_monotonicity.reference create mode 100644 tests/queries/0_stateless/01480_binary_operator_monotonicity.sql diff --git a/src/Functions/FunctionBinaryArithmetic.h b/src/Functions/FunctionBinaryArithmetic.h index ca0cc876035..f30b564d677 100644 --- a/src/Functions/FunctionBinaryArithmetic.h +++ b/src/Functions/FunctionBinaryArithmetic.h @@ -28,6 +28,7 @@ #include "FunctionFactory.h" #include #include +#include #if !defined(ARCADIA_BUILD) # include @@ -51,6 +52,7 @@ namespace ErrorCodes extern const int LOGICAL_ERROR; extern const int DECIMAL_OVERFLOW; extern const int CANNOT_ADD_DIFFERENT_AGGREGATE_STATES; + extern const int NUMBER_OF_ARGUMENTS_DOESNT_MATCH; } @@ -602,7 +604,8 @@ class FunctionBinaryArithmetic : public IFunction return castType(left, [&](const auto & left_) { return castType(right, [&](const auto & right_) { return f(left_, right_); }); }); } - FunctionOverloadResolverPtr getFunctionForIntervalArithmetic(const DataTypePtr & type0, const DataTypePtr & type1) const + static FunctionOverloadResolverPtr + getFunctionForIntervalArithmetic(const DataTypePtr & type0, const DataTypePtr & type1, const Context & context) { bool first_is_date_or_datetime = isDateOrDateTime(type0); bool second_is_date_or_datetime = isDateOrDateTime(type1); @@ -632,7 +635,7 @@ class FunctionBinaryArithmetic : public IFunction } if (second_is_date_or_datetime && is_minus) - throw Exception("Wrong order of arguments for function " + getName() + ": argument of type Interval cannot be first.", + throw Exception("Wrong order of arguments for function " + String(name) + ": argument of type Interval cannot be first.", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); std::string function_name; @@ -651,7 +654,7 @@ class FunctionBinaryArithmetic : public IFunction return FunctionFactory::instance().get(function_name, context); } - bool isAggregateMultiply(const DataTypePtr & type0, const DataTypePtr & type1) const + static bool isAggregateMultiply(const DataTypePtr & type0, const DataTypePtr & type1) { if constexpr (!is_multiply) return false; @@ -663,7 +666,7 @@ class FunctionBinaryArithmetic : public IFunction || (which0.isNativeUInt() && which1.isAggregateFunction()); } - bool isAggregateAddition(const DataTypePtr & type0, const DataTypePtr & type1) const + static bool isAggregateAddition(const DataTypePtr & type0, const DataTypePtr & type1) { if constexpr (!is_plus) return false; @@ -812,6 +815,11 @@ public: size_t getNumberOfArguments() const override { return 2; } DataTypePtr getReturnTypeImpl(const DataTypes & arguments) const override + { + return getReturnTypeImplStatic(arguments, context); + } + + static DataTypePtr getReturnTypeImplStatic(const DataTypes & arguments, const Context & context) { /// Special case when multiply aggregate function state if (isAggregateMultiply(arguments[0], arguments[1])) @@ -832,7 +840,7 @@ public: } /// Special case when the function is plus or minus, one of arguments is Date/DateTime and another is Interval. - if (auto function_builder = getFunctionForIntervalArithmetic(arguments[0], arguments[1])) + if (auto function_builder = getFunctionForIntervalArithmetic(arguments[0], arguments[1], context)) { ColumnsWithTypeAndName new_arguments(2); @@ -903,7 +911,7 @@ public: return false; }); if (!valid) - throw Exception("Illegal types " + arguments[0]->getName() + " and " + arguments[1]->getName() + " of arguments of function " + getName(), + throw Exception("Illegal types " + arguments[0]->getName() + " and " + arguments[1]->getName() + " of arguments of function " + String(name), ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); return type_res; } @@ -1110,7 +1118,8 @@ public: } /// Special case when the function is plus or minus, one of arguments is Date/DateTime and another is Interval. - if (auto function_builder = getFunctionForIntervalArithmetic(block.getByPosition(arguments[0]).type, block.getByPosition(arguments[1]).type)) + if (auto function_builder + = getFunctionForIntervalArithmetic(block.getByPosition(arguments[0]).type, block.getByPosition(arguments[1]).type, context)) { executeDateTimeIntervalPlusMinus(block, arguments, result, input_rows_count, function_builder); return; @@ -1200,4 +1209,167 @@ public: bool canBeExecutedOnDefaultArguments() const override { return valid_on_default_arguments; } }; + +template