thevar1able/ClickHouse
1
0
Fork 0
mirror of https://github.com/ClickHouse/ClickHouse.git synced 2024-11-23 16:12:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #51917 from ClickHouse/revert-system-drop-filesystem-cache-by-key

Browse Source 
Revert system drop filesystem cache by key
This commit is contained in:
Alexey Milovidov 2023-07-07 19:44:28 +03:00 committed by GitHub
commit 520195f7ef
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
17 changed files with 40 additions and 183 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
src/Disks/IO/CachedOnDiskReadBufferFromFile.cpp
View File

@ -74,22 +74,19 @@ CachedOnDiskReadBufferFromFile::CachedOnDiskReadBufferFromFile(
}
void CachedOnDiskReadBufferFromFile::appendFilesystemCacheLog(
const FileSegment & file_segment, CachedOnDiskReadBufferFromFile::ReadType type)
const FileSegment::Range & file_segment_range, CachedOnDiskReadBufferFromFile::ReadType type)
{
if (!cache_log)
return;
const auto range = file_segment.range();
FilesystemCacheLogElement elem
{
.event_time = std::chrono::system_clock::to_time_t(std::chrono::system_clock::now()),
.query_id = query_id,
.source_file_path = source_file_path,
.file_segment_range = { range.left, range.right },
.file_segment_range = { file_segment_range.left, file_segment_range.right },
.requested_range = { first_offset, read_until_position },
.file_segment_key = file_segment.key().toString(),
.file_segment_offset = file_segment.offset(),
.file_segment_size = range.size(),
.file_segment_size = file_segment_range.size(),
.read_from_cache_attempted = true,
.read_buffer_id = current_buffer_id,
.profile_counters = std::make_shared<ProfileEvents::Counters::Snapshot>(
@ -498,7 +495,7 @@ bool CachedOnDiskReadBufferFromFile::completeFileSegmentAndGetNext()
auto completed_range = current_file_segment->range();
if (cache_log)
appendFilesystemCacheLog(*current_file_segment, read_type);
appendFilesystemCacheLog(completed_range, read_type);
chassert(file_offset_of_buffer_end > completed_range.right);
@ -521,7 +518,7 @@ CachedOnDiskReadBufferFromFile::~CachedOnDiskReadBufferFromFile()
{
if (cache_log && file_segments && !file_segments->empty())
{
appendFilesystemCacheLog(file_segments->front(), read_type);
appendFilesystemCacheLog(file_segments->front().range(), read_type);
}
}

2
src/Disks/IO/CachedOnDiskReadBufferFromFile.h
View File

@ -90,7 +90,7 @@ private:
bool completeFileSegmentAndGetNext();
void appendFilesystemCacheLog(const FileSegment & file_segment, ReadType read_type);
void appendFilesystemCacheLog(const FileSegment::Range & file_segment_range, ReadType read_type);
bool writeCache(char * data, size_t size, size_t offset, FileSegment & file_segment);

2
src/Disks/IO/ReadBufferFromRemoteFSGather.cpp
View File

@ -90,8 +90,6 @@ void ReadBufferFromRemoteFSGather::appendUncachedReadInfo()
.source_file_path = current_object.remote_path,
.file_segment_range = { 0, current_object.bytes_size },
.cache_type = FilesystemCacheLogElement::CacheType::READ_FROM_FS_BYPASSING_CACHE,
.file_segment_key = {},
.file_segment_offset = {},
.file_segment_size = current_object.bytes_size,
.read_from_cache_attempted = false,
};

34
src/Interpreters/Cache/FileCache.cpp
View File

@ -806,13 +806,6 @@ bool FileCache::tryReserve(FileSegment & file_segment, const size_t size)
return true;
}
void FileCache::removeKey(const Key & key)
{
assertInitialized();
auto locked_key = metadata.lockKeyMetadata(key, CacheMetadata::KeyNotFoundPolicy::THROW);
locked_key->removeAll();
}
void FileCache::removeKeyIfExists(const Key & key)
{
assertInitialized();
@ -825,14 +818,7 @@ void FileCache::removeKeyIfExists(const Key & key)
/// But if we have multiple replicated zero-copy tables on the same server
/// it became possible to start removing something from cache when it is used
/// by other "zero-copy" tables. That is why it's not an error.
locked_key->removeAll(/* if_releasable */true);
}
void FileCache::removeFileSegment(const Key & key, size_t offset)
{
assertInitialized();
auto locked_key = metadata.lockKeyMetadata(key, CacheMetadata::KeyNotFoundPolicy::THROW);
locked_key->removeFileSegment(offset);
locked_key->removeAllReleasable();
}
void FileCache::removePathIfExists(const String & path)
@ -844,12 +830,22 @@ void FileCache::removeAllReleasable()
{
assertInitialized();
metadata.iterate([](LockedKey & locked_key) { locked_key.removeAll(/* if_releasable */true); });
auto lock = lockCache();
main_priority->iterate([&](LockedKey & locked_key, const FileSegmentMetadataPtr & segment_metadata)
{
if (segment_metadata->releasable())
{
auto file_segment = segment_metadata->file_segment;
locked_key.removeFileSegment(file_segment->offset(), file_segment->lock());
return PriorityIterationResult::REMOVE_AND_CONTINUE;
}
return PriorityIterationResult::CONTINUE;
}, lock);
if (stash)
{
/// Remove all access information.
auto lock = lockCache();
stash->records.clear();
stash->queue->removeAll(lock);
}
@ -919,7 +915,7 @@ void FileCache::loadMetadata()
continue;
}
const auto key = Key::fromKeyString(key_directory.filename().string());
const auto key = Key(unhexUInt<UInt128>(key_directory.filename().string().data()));
auto locked_key = metadata.lockKeyMetadata(key, CacheMetadata::KeyNotFoundPolicy::CREATE_EMPTY, /* is_initial_load */true);
for (fs::directory_iterator offset_it{key_directory}; offset_it != fs::directory_iterator(); ++offset_it)
@ -1074,7 +1070,7 @@ FileSegmentsHolderPtr FileCache::getSnapshot()
FileSegmentsHolderPtr FileCache::getSnapshot(const Key & key)
{
FileSegments file_segments;
auto locked_key = metadata.lockKeyMetadata(key, CacheMetadata::KeyNotFoundPolicy::THROW_LOGICAL);
auto locked_key = metadata.lockKeyMetadata(key, CacheMetadata::KeyNotFoundPolicy::THROW);
for (const auto & [_, file_segment_metadata] : *locked_key->getKeyMetadata())
file_segments.push_back(FileSegment::getSnapshot(file_segment_metadata->file_segment));
return std::make_unique<FileSegmentsHolder>(std::move(file_segments));

12
src/Interpreters/Cache/FileCache.h
View File

@ -83,19 +83,13 @@ public:
FileSegmentsHolderPtr set(const Key & key, size_t offset, size_t size, const CreateFileSegmentSettings & settings);
/// Remove file segment by `key` and `offset`. Throws if file segment does not exist.
void removeFileSegment(const Key & key, size_t offset);
/// Remove files by `key`. Throws if key does not exist.
void removeKey(const Key & key);
/// Remove files by `key`.
/// Remove files by `key`. Removes files which might be used at the moment.
void removeKeyIfExists(const Key & key);
/// Removes files by `path`.
/// Removes files by `path`. Removes files which might be used at the moment.
void removePathIfExists(const String & path);
/// Remove files by `key`.
/// Remove files by `key`. Will not remove files which are used at the moment.
void removeAllReleasable();
std::vector<String> tryGetCachePaths(const Key & key);

5
src/Interpreters/Cache/FileCacheKey.cpp
View File

@ -28,9 +28,4 @@ FileCacheKey FileCacheKey::random()
return FileCacheKey(UUIDHelpers::generateV4().toUnderType());
}
FileCacheKey FileCacheKey::fromKeyString(const std::string & key_str)
{
return FileCacheKey(unhexUInt<UInt128>(key_str.data()));
}
}

2
src/Interpreters/Cache/FileCacheKey.h
View File

@ -21,8 +21,6 @@ struct FileCacheKey
static FileCacheKey random();
bool operator==(const FileCacheKey & other) const { return key == other.key; }
static FileCacheKey fromKeyString(const std::string & key_str);
};
using FileCacheKeyAndOffset = std::pair<FileCacheKey, size_t>;

26
src/Interpreters/Cache/Metadata.cpp
View File

@ -25,7 +25,6 @@ namespace DB
namespace ErrorCodes
{
extern const int LOGICAL_ERROR;
extern const int BAD_ARGUMENTS;
}
FileSegmentMetadata::FileSegmentMetadata(FileSegmentPtr && file_segment_)
@ -192,8 +191,6 @@ LockedKeyPtr CacheMetadata::lockKeyMetadata(
if (it == end())
{
if (key_not_found_policy == KeyNotFoundPolicy::THROW)
throw Exception(ErrorCodes::BAD_ARGUMENTS, "No such key `{}` in cache", key);
else if (key_not_found_policy == KeyNotFoundPolicy::THROW_LOGICAL)
throw Exception(ErrorCodes::LOGICAL_ERROR, "No such key `{}` in cache", key);
else if (key_not_found_policy == KeyNotFoundPolicy::RETURN_NULL)
return nullptr;
@ -218,8 +215,6 @@ LockedKeyPtr CacheMetadata::lockKeyMetadata(
return locked_metadata;
if (key_not_found_policy == KeyNotFoundPolicy::THROW)
throw Exception(ErrorCodes::BAD_ARGUMENTS, "No such key `{}` in cache", key);
else if (key_not_found_policy == KeyNotFoundPolicy::THROW_LOGICAL)
throw Exception(ErrorCodes::LOGICAL_ERROR, "No such key `{}` in cache", key);
if (key_not_found_policy == KeyNotFoundPolicy::RETURN_NULL)
@ -563,11 +558,11 @@ bool LockedKey::isLastOwnerOfFileSegment(size_t offset) const
return file_segment_metadata->file_segment.use_count() == 2;
}
void LockedKey::removeAll(bool if_releasable)
void LockedKey::removeAllReleasable()
{
for (auto it = key_metadata->begin(); it != key_metadata->end();)
{
if (if_releasable && !it->second->releasable())
if (!it->second->releasable())
{
++it;
continue;
@ -588,32 +583,17 @@ void LockedKey::removeAll(bool if_releasable)
}
}
KeyMetadata::iterator LockedKey::removeFileSegment(size_t offset)
{
auto it = key_metadata->find(offset);
if (it == key_metadata->end())
throw Exception(ErrorCodes::BAD_ARGUMENTS, "There is no offset {}", offset);
auto file_segment = it->second->file_segment;
return removeFileSegmentImpl(it, file_segment->lock());
}
KeyMetadata::iterator LockedKey::removeFileSegment(size_t offset, const FileSegmentGuard::Lock & segment_lock)
{
auto it = key_metadata->find(offset);
if (it == key_metadata->end())
throw Exception(ErrorCodes::LOGICAL_ERROR, "There is no offset {}", offset);
return removeFileSegmentImpl(it, segment_lock);
}
KeyMetadata::iterator LockedKey::removeFileSegmentImpl(KeyMetadata::iterator it, const FileSegmentGuard::Lock & segment_lock)
{
auto file_segment = it->second->file_segment;
LOG_DEBUG(
key_metadata->log, "Remove from cache. Key: {}, offset: {}, size: {}",
getKey(), file_segment->offset(), file_segment->reserved_size);
getKey(), offset, file_segment->reserved_size);
chassert(file_segment->assertCorrectnessUnlocked(segment_lock));

8
src/Interpreters/Cache/Metadata.h
View File

@ -87,7 +87,7 @@ struct CacheMetadata : public std::unordered_map<FileCacheKey, KeyMetadataPtr>,
{
public:
using Key = FileCacheKey;
using IterateCacheMetadataFunc = std::function<void(LockedKey &)>;
using IterateCacheMetadataFunc = std::function<void(const LockedKey &)>;
explicit CacheMetadata(const std::string & path_);
@ -106,7 +106,6 @@ public:
enum class KeyNotFoundPolicy
{
THROW,
THROW_LOGICAL,
CREATE_EMPTY,
RETURN_NULL,
};
@ -170,10 +169,9 @@ struct LockedKey : private boost::noncopyable
std::shared_ptr<const KeyMetadata> getKeyMetadata() const { return key_metadata; }
std::shared_ptr<KeyMetadata> getKeyMetadata() { return key_metadata; }
void removeAll(bool if_releasable = true);
void removeAllReleasable();
KeyMetadata::iterator removeFileSegment(size_t offset, const FileSegmentGuard::Lock &);
KeyMetadata::iterator removeFileSegment(size_t offset);
void shrinkFileSegmentToDownloadedSize(size_t offset, const FileSegmentGuard::Lock &);
@ -190,8 +188,6 @@ struct LockedKey : private boost::noncopyable
std::string toString() const;
private:
KeyMetadata::iterator removeFileSegmentImpl(KeyMetadata::iterator it, const FileSegmentGuard::Lock &);
const std::shared_ptr<KeyMetadata> key_metadata;
KeyGuard::Lock lock; /// `lock` must be destructed before `key_metadata`.
};

4
src/Interpreters/FilesystemCacheLog.cpp
View File

@ -40,8 +40,6 @@ NamesAndTypesList FilesystemCacheLogElement::getNamesAndTypes()
{"source_file_path", std::make_shared<DataTypeString>()},
{"file_segment_range", std::make_shared<DataTypeTuple>(types)},
{"total_requested_range", std::make_shared<DataTypeTuple>(types)},
{"key", std::make_shared<DataTypeString>()},
{"offset", std::make_shared<DataTypeUInt64>()},
{"size", std::make_shared<DataTypeUInt64>()},
{"read_type", std::make_shared<DataTypeString>()},
{"read_from_cache_attempted", std::make_shared<DataTypeUInt8>()},
@ -62,8 +60,6 @@ void FilesystemCacheLogElement::appendToBlock(MutableColumns & columns) const
columns[i++]->insert(source_file_path);
columns[i++]->insert(Tuple{file_segment_range.first, file_segment_range.second});
columns[i++]->insert(Tuple{requested_range.first, requested_range.second});
columns[i++]->insert(file_segment_key);
columns[i++]->insert(file_segment_offset);
columns[i++]->insert(file_segment_size);
columns[i++]->insert(typeToString(cache_type));
columns[i++]->insert(read_from_cache_attempted);

13
src/Interpreters/FilesystemCacheLog.h
View File

@ -11,7 +11,16 @@
namespace DB
{
///
/// -------- Column --------- Type ------
/// | event_date | DateTime |
/// | event_time | UInt64 |
/// | query_id | String |
/// | remote_file_path | String |
/// | segment_range | Tuple |
/// | read_type | String |
/// -------------------------------------
///
struct FilesystemCacheLogElement
{
enum class CacheType
@ -30,8 +39,6 @@ struct FilesystemCacheLogElement
std::pair<size_t, size_t> file_segment_range{};
std::pair<size_t, size_t> requested_range{};
CacheType cache_type{};
std::string file_segment_key;
size_t file_segment_offset;
size_t file_segment_size;
bool read_from_cache_attempted;
String read_buffer_id;

13
src/Interpreters/InterpreterSystemQuery.cpp
View File

@ -370,18 +370,7 @@ BlockIO InterpreterSystemQuery::execute()
else
{
auto cache = FileCacheFactory::instance().getByName(query.filesystem_cache_name).cache;
if (query.key_to_drop.empty())
{
cache->removeAllReleasable();
}
else
{
auto key = FileCacheKey::fromKeyString(query.key_to_drop);
if (query.offset_to_drop.has_value())
cache->removeFileSegment(key, query.offset_to_drop.value());
else
cache->removeKey(key);
}
cache->removeAllReleasable();
}
break;
}

8
src/Parsers/ASTSystemQuery.cpp
View File

@ -210,15 +210,7 @@ void ASTSystemQuery::formatImpl(const FormatSettings & settings, FormatState &,
else if (type == Type::DROP_FILESYSTEM_CACHE)
{
if (!filesystem_cache_name.empty())
{
settings.ostr << (settings.hilite ? hilite_none : "") << " " << filesystem_cache_name;
if (!key_to_drop.empty())
{
settings.ostr << (settings.hilite ? hilite_none : "") << " KEY " << key_to_drop;
if (offset_to_drop.has_value())
settings.ostr << (settings.hilite ? hilite_none : "") << " OFFSET " << offset_to_drop.value();
}
}
}
else if (type == Type::UNFREEZE)
{

2
src/Parsers/ASTSystemQuery.h
View File

@ -107,8 +107,6 @@ public:
UInt64 seconds{};
String filesystem_cache_name;
std::string key_to_drop;
std::optional<size_t> offset_to_drop;
String backup_name;

8
src/Parsers/ParserSystemQuery.cpp
View File

@ -405,15 +405,7 @@ bool ParserSystemQuery::parseImpl(IParser::Pos & pos, ASTPtr & node, Expected &
ParserLiteral path_parser;
ASTPtr ast;
if (path_parser.parse(pos, ast, expected))
{
res->filesystem_cache_name = ast->as<ASTLiteral>()->value.safeGet<String>();
if (ParserKeyword{"KEY"}.ignore(pos, expected) && ParserIdentifier().parse(pos, ast, expected))
{
res->key_to_drop = ast->as<ASTIdentifier>()->name();
if (ParserKeyword{"OFFSET"}.ignore(pos, expected) && ParserLiteral().parse(pos, ast, expected))
res->offset_to_drop = ast->as<ASTLiteral>()->value.safeGet<UInt64>();
}
}
if (!parseQueryWithOnCluster(res, pos, expected))
return false;
break;

4
tests/queries/0_stateless/02808_filesystem_cache_drop_query.reference
View File

@ -1,4 +0,0 @@
1
0
1
0

67
tests/queries/0_stateless/02808_filesystem_cache_drop_query.sh
View File

@ -1,67 +0,0 @@
#!/usr/bin/env bash
# Tags: no-fasttest, no-parallel, no-s3-storage, no-random-settings
# set -x
CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
# shellcheck source=../shell_config.sh
. "$CUR_DIR"/../shell_config.sh
disk_name="${CLICKHOUSE_TEST_UNIQUE_NAME}"
$CLICKHOUSE_CLIENT -nm --query """
DROP TABLE IF EXISTS test;
CREATE TABLE test (a Int32, b String)
ENGINE = MergeTree() ORDER BY tuple()
SETTINGS disk = disk_$disk_name(type = cache, max_size = '100Ki', path = ${CLICKHOUSE_TEST_UNIQUE_NAME}, disk = s3_disk);
INSERT INTO test SELECT 1, 'test';
"""
query_id=$RANDOM
$CLICKHOUSE_CLIENT --query_id "$query_id" --query "SELECT * FROM test FORMAT Null SETTINGS enable_filesystem_cache_log = 1"
${CLICKHOUSE_CLIENT} -q " system flush logs"
key=$($CLICKHOUSE_CLIENT -nm --query """
SELECT key FROM system.filesystem_cache_log WHERE query_id = '$query_id' ORDER BY size DESC LIMIT 1;
""")
offset=$($CLICKHOUSE_CLIENT -nm --query """
SELECT offset FROM system.filesystem_cache_log WHERE query_id = '$query_id' ORDER BY size DESC LIMIT 1;
""")
$CLICKHOUSE_CLIENT -nm --query """
SELECT count() FROM system.filesystem_cache WHERE key = '$key' AND file_segment_range_begin = $offset;
"""
$CLICKHOUSE_CLIENT -nm --query """
SYSTEM DROP FILESYSTEM CACHE '$disk_name' KEY $key OFFSET $offset;
"""
$CLICKHOUSE_CLIENT -nm --query """
SELECT count() FROM system.filesystem_cache WHERE key = '$key' AND file_segment_range_begin = $offset;
"""
query_id=$RANDOM$RANDOM
$CLICKHOUSE_CLIENT --query_id "$query_id" --query "SELECT * FROM test FORMAT Null SETTINGS enable_filesystem_cache_log = 1"
${CLICKHOUSE_CLIENT} -q " system flush logs"
key=$($CLICKHOUSE_CLIENT -nm --query """
SELECT key FROM system.filesystem_cache_log WHERE query_id = '$query_id' ORDER BY size DESC LIMIT 1;
""")
$CLICKHOUSE_CLIENT -nm --query """
SELECT count() FROM system.filesystem_cache WHERE key = '$key';
"""
$CLICKHOUSE_CLIENT -nm --query """
SYSTEM DROP FILESYSTEM CACHE '$disk_name' KEY $key
"""
$CLICKHOUSE_CLIENT -nm --query """
SELECT count() FROM system.filesystem_cache WHERE key = '$key';
"""