thevar1able/ClickHouse
1
0
Fork 0
mirror of https://github.com/ClickHouse/ClickHouse.git synced 2024-11-21 23:21:59 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'virtualutil' of github.com:amosbird/ClickHouse into virtualutil

Browse Source
This commit is contained in:
Alexey Milovidov 2021-03-31 08:19:39 +03:00
commit 7f9c0e621a
182 changed files with 1820 additions and 903 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
CMakeLists.txt
View File

@ -248,19 +248,27 @@ if (ARCH_NATIVE)
set (COMPILER_FLAGS "${COMPILER_FLAGS} -march=native")
endif ()
if (COMPILER_GCC OR COMPILER_CLANG)
# to make numeric_limits<__int128> works with GCC
set (_CXX_STANDARD "gnu++2a")
else()
set (_CXX_STANDARD "c++2a")
endif()
if (${CMAKE_VERSION} VERSION_LESS "3.12.4")
# CMake < 3.12 doesn't support setting 20 as a C++ standard version.
# We will add C++ standard controlling flag in CMAKE_CXX_FLAGS manually for now.
# cmake < 3.12 doesn't support 20. We'll set CMAKE_CXX_FLAGS for now
# set (CMAKE_CXX_STANDARD 20)
set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=${_CXX_STANDARD}")
if (COMPILER_GCC OR COMPILER_CLANG)
# to make numeric_limits<__int128> works with GCC
set (_CXX_STANDARD "gnu++2a")
else ()
set (_CXX_STANDARD "c++2a")
endif ()
set (CMAKE_CXX_EXTENSIONS 0) # https://cmake.org/cmake/help/latest/prop_tgt/CXX_EXTENSIONS.html#prop_tgt:CXX_EXTENSIONS
set (CMAKE_CXX_STANDARD_REQUIRED ON)
set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=${_CXX_STANDARD}")
else ()
set (CMAKE_CXX_STANDARD 20)
set (CMAKE_CXX_EXTENSIONS ON) # Same as gnu++2a (ON) vs c++2a (OFF): https://cmake.org/cmake/help/latest/prop_tgt/CXX_EXTENSIONS.html
set (CMAKE_CXX_STANDARD_REQUIRED ON)
endif ()
set (CMAKE_C_STANDARD 11)
set (CMAKE_C_EXTENSIONS ON)
set (CMAKE_C_STANDARD_REQUIRED ON)
if (COMPILER_GCC OR COMPILER_CLANG)
# Enable C++14 sized global deallocation functions. It should be enabled by setting -std=c++14 but I'm not sure.

25
contrib/CMakeLists.txt vendored
View File

@ -215,15 +215,17 @@ if (USE_EMBEDDED_COMPILER AND USE_INTERNAL_LLVM_LIBRARY)
set (LLVM_ENABLE_RTTI 1 CACHE INTERNAL "")
set (LLVM_ENABLE_PIC 0 CACHE INTERNAL "")
set (LLVM_TARGETS_TO_BUILD "X86;AArch64" CACHE STRING "")
# Yes it is set globally, but this is not enough, since llvm will add -std=c++11 after default
# And c++2a cannot be used, due to ambiguous operator !=
if (COMPILER_GCC OR COMPILER_CLANG)
set (_CXX_STANDARD "gnu++17")
else()
set (_CXX_STANDARD "c++17")
endif()
set (LLVM_CXX_STD ${_CXX_STANDARD} CACHE STRING "" FORCE)
# Need to use C++17 since the compilation is not possible with C++20 currently, due to ambiguous operator != etc.
# LLVM project will set its default value for the -std=... but our global setting from CMake will override it.
set (CMAKE_CXX_STANDARD_bak ${CMAKE_CXX_STANDARD})
set (CMAKE_CXX_STANDARD 17)
add_subdirectory (llvm/llvm)
set (CMAKE_CXX_STANDARD ${CMAKE_CXX_STANDARD_bak})
unset (CMAKE_CXX_STANDARD_bak)
target_include_directories(LLVMSupport SYSTEM BEFORE PRIVATE ${ZLIB_INCLUDE_DIR})
endif ()
@ -280,7 +282,14 @@ if (USE_AMQPCPP)
add_subdirectory (amqpcpp-cmake)
endif()
if (USE_CASSANDRA)
# Need to use C++17 since the compilation is not possible with C++20 currently.
set (CMAKE_CXX_STANDARD_bak ${CMAKE_CXX_STANDARD})
set (CMAKE_CXX_STANDARD 17)
add_subdirectory (cassandra)
set (CMAKE_CXX_STANDARD ${CMAKE_CXX_STANDARD_bak})
unset (CMAKE_CXX_STANDARD_bak)
endif()
# Should go before:

2
contrib/grpc vendored

@ -1 +1 @@
Subproject commit 7436366ceb341ba5c00ea29f1645e02a2b70bf93
Subproject commit 8d558f03fe370240081424fafa76cdc9301ea14b

11
docker/packager/binary/Dockerfile
View File

@ -14,12 +14,8 @@ RUN apt-get update \
lsb-release \
wget \
--yes --no-install-recommends --verbose-versions \
&& cat /etc/resolv.conf \
&& echo "nameserver 1.1.1.1" >> /etc/resolv.conf \
&& nslookup -debug apt.llvm.org \
&& ping -c1 apt.llvm.org \
&& wget -nv --retry-connrefused --tries=10 -O /tmp/llvm-snapshot.gpg.key https://apt.llvm.org/llvm-snapshot.gpg.key \
&& export LLVM_PUBKEY_HASH="bda960a8da687a275a2078d43c111d66b1c6a893a3275271beedf266c1ff4a0cdecb429c7a5cccf9f486ea7aa43fd27f" \
&& wget -nv -O /tmp/llvm-snapshot.gpg.key https://apt.llvm.org/llvm-snapshot.gpg.key \
&& echo "${LLVM_PUBKEY_HASH} /tmp/llvm-snapshot.gpg.key" | sha384sum -c \
&& apt-key add /tmp/llvm-snapshot.gpg.key \
&& export CODENAME="$(lsb_release --codename --short | tr 'A-Z' 'a-z')" \
@ -36,10 +32,7 @@ RUN apt-get update \
software-properties-common \
--yes --no-install-recommends
RUN cat /etc/resolv.conf \
&& echo "nameserver 1.1.1.1" >> /etc/resolv.conf \
&& nslookup -debug apt.llvm.org \
&& apt-get update \
RUN apt-get update \
&& apt-get install \
bash \
cmake \

6
docker/test/base/Dockerfile
View File

@ -4,9 +4,8 @@ FROM ubuntu:20.04
ENV DEBIAN_FRONTEND=noninteractive LLVM_VERSION=11
RUN apt-get update \
&& apt-get install apt-utils ca-certificates lsb-release wget gnupg apt-transport-https \
&& apt-get install ca-certificates lsb-release wget gnupg apt-transport-https \
--yes --no-install-recommends --verbose-versions \
&& echo "nameserver 1.1.1.1" >> /etc/resolv.conf \
&& export LLVM_PUBKEY_HASH="bda960a8da687a275a2078d43c111d66b1c6a893a3275271beedf266c1ff4a0cdecb429c7a5cccf9f486ea7aa43fd27f" \
&& wget -nv -O /tmp/llvm-snapshot.gpg.key https://apt.llvm.org/llvm-snapshot.gpg.key \
&& echo "${LLVM_PUBKEY_HASH} /tmp/llvm-snapshot.gpg.key" | sha384sum -c \
@ -32,8 +31,7 @@ RUN curl -O https://clickhouse-builds.s3.yandex.net/utils/1/dpkg-deb \
&& chmod +x dpkg-deb \
&& cp dpkg-deb /usr/bin
RUN echo "nameserver 1.1.1.1" >> /etc/resolv.conf \
&& apt-get update \
RUN apt-get update \
&& apt-get install \
clang-${LLVM_VERSION} \
debhelper \

2
docker/test/fasttest/run.sh
View File

@ -70,7 +70,7 @@ function start_server
--path "$FASTTEST_DATA"
--user_files_path "$FASTTEST_DATA/user_files"
--top_level_domains_path "$FASTTEST_DATA/top_level_domains"
--test_keeper_server.log_storage_path "$FASTTEST_DATA/coordination"
--keeper_server.log_storage_path "$FASTTEST_DATA/coordination"
)
clickhouse-server "${opts[@]}" &>> "$FASTTEST_OUTPUT/server.log" &
server_pid=$!

9
docker/test/stateless/run.sh
View File

@ -51,6 +51,7 @@ function run_tests()
# Skip these tests, because they fail when we rerun them multiple times
if [ "$NUM_TRIES" -gt "1" ]; then
ADDITIONAL_OPTIONS+=('--order=random')
ADDITIONAL_OPTIONS+=('--skip')
ADDITIONAL_OPTIONS+=('00000_no_tests_to_skip')
ADDITIONAL_OPTIONS+=('--jobs')
@ -74,7 +75,13 @@ timeout "$MAX_RUN_TIME" bash -c run_tests ||:
./process_functional_tests_result.py || echo -e "failure\tCannot parse results" > /test_output/check_status.tsv
pigz < /var/log/clickhouse-server/clickhouse-server.log > /test_output/clickhouse-server.log.gz ||:
clickhouse-client -q "system flush logs" ||:
pigz < /var/log/clickhouse-server/clickhouse-server.log > /test_output/clickhouse-server.log.gz &
clickhouse-client -q "select * from system.query_log format TSVWithNamesAndTypes" | pigz > /test_output/query-log.tsv.gz &
clickhouse-client -q "select * from system.query_thread_log format TSVWithNamesAndTypes" | pigz > /test_output/query-thread-log.tsv.gz &
wait ||:
mv /var/log/clickhouse-server/stderr.log /test_output/ ||:
if [[ -n "$WITH_COVERAGE" ]] && [[ "$WITH_COVERAGE" -eq 1 ]]; then
tar -chf /test_output/clickhouse_coverage.tar.gz /profraw ||:

32
docs/en/operations/settings/settings.md
View File

@ -769,6 +769,38 @@ Example:
log_query_threads=1
```
## log_comment {#settings-log-comment}
Specifies the value for the `log_comment` field of the [system.query_log](../system-tables/query_log.md) table and comment text for the server log.
It can be used to improve the readability of server logs. Additionally, it helps to select queries related to the test from the `system.query_log` after running [clickhouse-test](../../development/tests.md).
Possible values:
- Any string no longer than [max_query_size](#settings-max_query_size). If length is exceeded, the server throws an exception.
Default value: empty string.
**Example**
Query:
``` sql
SET log_comment = 'log_comment test', log_queries = 1;
SELECT 1;
SYSTEM FLUSH LOGS;
SELECT type, query FROM system.query_log WHERE log_comment = 'log_comment test' AND event_date >= yesterday() ORDER BY event_time DESC LIMIT 2;
```
Result:
``` text
┌─type────────┬─query─────┐
│ QueryStart │ SELECT 1; │
│ QueryFinish │ SELECT 1; │
└─────────────┴───────────┘
```
## max_insert_block_size {#settings-max_insert_block_size}
The size of blocks (in a count of rows) to form for insertion into a table.

10
docs/en/sql-reference/aggregate-functions/parametric-functions.md
View File

@ -243,7 +243,7 @@ The function works according to the algorithm:
**Syntax**
``` sql
windowFunnel(window, [mode])(timestamp, cond1, cond2, ..., condN)
windowFunnel(window, [mode, [mode, ... ]])(timestamp, cond1, cond2, ..., condN)
```
**Arguments**
@ -253,9 +253,11 @@ windowFunnel(window, [mode])(timestamp, cond1, cond2, ..., condN)
**Parameters**
- `window` — Length of the sliding window. The unit of `window` depends on the `timestamp` itself and varies. Determined using the expression `timestamp of cond2 <= timestamp of cond1 + window`.
- `mode` — It is an optional argument.
- `'strict'` — When the `'strict'` is set, the windowFunnel() applies conditions only for the unique values.
- `window` — Length of the sliding window, it is the time interval between first condition and last condition. The unit of `window` depends on the `timestamp` itself and varies. Determined using the expression `timestamp of cond1 <= timestamp of cond2 <= ... <= timestamp of condN <= timestamp of cond1 + window`.
- `mode` — It is an optional argument. One or more modes can be set.
- `'strict'` — If same condition holds for sequence of events then such non-unique events would be skipped.
- `'strict_order'` — Don't allow interventions of other events. E.g. in the case of `A->B->D->C`, it stops finding `A->B->C` at the `D` and the max event level is 2.
- `'strict_increase'` — Apply conditions only to events with strictly increasing timestamps.
**Returned value**

52
docs/en/sql-reference/functions/ip-address-functions.md
View File

@ -394,3 +394,55 @@ Result:
└──────────────────┴────────────────────┘
```
## isIPAddressInRange {#isipaddressinrange}
Determines if an IP address is contained in a network represented in the [CIDR](https://en.wikipedia.org/wiki/Classless_Inter-Domain_Routing) notation. Returns `1` if true, or `0` otherwise.
**Syntax**
``` sql
isIPAddressInRange(address, prefix)
```
This function accepts both IPv4 and IPv6 addresses (and networks) represented as strings. It returns `0` if the IP version of the address and the CIDR don't match.
**Arguments**
- `address` — An IPv4 or IPv6 address. [String](../../sql-reference/data-types/string.md).
- `prefix` — An IPv4 or IPv6 network prefix in CIDR. [String](../../sql-reference/data-types/string.md).
**Returned value**
- `1` or `0`.
Type: [UInt8](../../sql-reference/data-types/int-uint.md).
**Example**
Query:
``` sql
SELECT isIPAddressInRange('127.0.0.1', '127.0.0.0/8')
```
Result:
``` text
┌─isIPAddressInRange('127.0.0.1', '127.0.0.0/8')─┐
│ 1 │
└────────────────────────────────────────────────┘
```
Query:
``` sql
SELECT isIPAddressInRange('127.0.0.1', 'ffff::/16')
```
Result:
``` text
┌─isIPAddressInRange('127.0.0.1', 'ffff::/16')─┐
│ 0 │
└──────────────────────────────────────────────┘
```

46
docs/en/sql-reference/functions/string-search-functions.md
View File

@ -12,7 +12,9 @@ The search is case-sensitive by default in all these functions. There are separa
## position(haystack, needle), locate(haystack, needle) {#position}
Returns the position (in bytes) of the found substring in the string, starting from 1.
Searches for the substring `needle` in the string `haystack`.
Returns the position (in bytes) of the found substring in the string, starting from 1.
For a case-insensitive search, use the function [positionCaseInsensitive](#positioncaseinsensitive).
@ -20,15 +22,22 @@ For a case-insensitive search, use the function [positionCaseInsensitive](#posit
``` sql
position(haystack, needle[, start_pos])
```
```
``` sql
position(needle IN haystack)
```
Alias: `locate(haystack, needle[, start_pos])`.
!!! note "Note"
Syntax of `position(needle IN haystack)` provides SQL-compatibility, the function works the same way as to `position(haystack, needle)`.
**Arguments**
- `haystack` — String, in which substring will to be searched. [String](../../sql-reference/syntax.md#syntax-string-literal).
- `needle` — Substring to be searched. [String](../../sql-reference/syntax.md#syntax-string-literal).
- `start_pos` — Optional parameter, position of the first character in the string to start search. [UInt](../../sql-reference/data-types/int-uint.md).
- `start_pos` – Position of the first character in the string to start search. [UInt](../../sql-reference/data-types/int-uint.md). Optional.
**Returned values**
@ -83,6 +92,36 @@ Result:
└───────────────────────────────┘
```
**Examples for POSITION(needle IN haystack) syntax**
Query:
```sql
SELECT 3 = position('c' IN 'abc');
```
Result:
```text
┌─equals(3, position('abc', 'c'))─┐
│ 1 │
└─────────────────────────────────┘
```
Query:
```sql
SELECT 6 = position('/' IN s) FROM (SELECT 'Hello/World' AS s);
```
Result:
```text
┌─equals(6, position(s, '/'))─┐
│ 1 │
└─────────────────────────────┘
```
## positionCaseInsensitive {#positioncaseinsensitive}
The same as [position](#position) returns the position (in bytes) of the found substring in the string, starting from 1. Use the function for a case-insensitive search.
@ -772,4 +811,3 @@ Result:
│ 2 │
└───────────────────────────────┘
```

35
docs/ru/operations/settings/settings.md
View File

@ -759,6 +759,38 @@ log_queries_min_type='EXCEPTION_WHILE_PROCESSING'
log_query_threads=1
```
## log_comment {#settings-log-comment}
Задаёт значение поля `log_comment` таблицы [system.query_log](../system-tables/query_log.md) и текст комментария в логе сервера.
Может быть использована для улучшения читабельности логов сервера. Кроме того, помогает быстро выделить связанные с тестом запросы из `system.query_log` после запуска [clickhouse-test](../../development/tests.md).
Возможные значения:
- Любая строка не длиннее [max_query_size](#settings-max_query_size). При превышении длины сервер сгенерирует исключение.
Значение по умолчанию: пустая строка.
**Пример**
Запрос:
``` sql
SET log_comment = 'log_comment test', log_queries = 1;
SELECT 1;
SYSTEM FLUSH LOGS;
SELECT type, query FROM system.query_log WHERE log_comment = 'log_comment test' AND event_date >= yesterday() ORDER BY event_time DESC LIMIT 2;
```
Результат:
``` text
┌─type────────┬─query─────┐
│ QueryStart │ SELECT 1; │
│ QueryFinish │ SELECT 1; │
└─────────────┴───────────┘
```
## max_insert_block_size {#settings-max_insert_block_size}
Формировать блоки указанного размера, при вставке в таблицу.
@ -2655,7 +2687,6 @@ SELECT * FROM test2;
Значение по умолчанию: `0`.
## live_view_heartbeat_interval {#live-view-heartbeat-interval}
Задает интервал в секундах для периодической проверки существования [LIVE VIEW](../../sql-reference/statements/create/view.md#live-view).
@ -2680,4 +2711,4 @@ SELECT * FROM test2;
Значение по умолчанию: `60`.
[Оригинальная статья](https://clickhouse.tech/docs/ru/operations/settings/settings/) <!--hide-->
[Оригинальная статья](https://clickhouse.tech/docs/ru/operations/settings/settings/) <!--hide-->

7
docs/ru/sql-reference/aggregate-functions/parametric-functions.md
View File

@ -243,7 +243,7 @@ SELECT sequenceCount('(?1).*(?2)')(time, number = 1, number = 2) FROM t
**Синтаксис**
``` sql
windowFunnel(window, [mode])(timestamp, cond1, cond2, ..., condN)
windowFunnel(window, [mode, [mode, ... ]])(timestamp, cond1, cond2, ..., condN)
```
**Аргументы**
@ -254,7 +254,10 @@ windowFunnel(window, [mode])(timestamp, cond1, cond2, ..., condN)
**Параметры**
- `window` — ширина скользящего окна по времени. Единица измерения зависит от `timestamp` и может варьироваться. Должно соблюдаться условие `timestamp события cond2 <= timestamp события cond1 + window`.
- `mode` — необязательный параметр. Если установлено значение `'strict'`, то функция `windowFunnel()` применяет условия только для уникальных значений.
- `mode` — необязательный параметр. Может быть установленно несколько значений одновременно.
- `'strict'` — не учитывать подряд идущие повторяющиеся события.
- `'strict_order'` — запрещает посторонние события в искомой последовательности. Например, при поиске цепочки `A->B->C` в `A->B->D->C` поиск будет остановлен на `D` и функция вернет 2.
- `'strict_increase'` — условия прменяются только для событий со строго возрастающими временными метками.
**Возвращаемое значение**

51
docs/ru/sql-reference/functions/ip-address-functions.md
View File

@ -395,3 +395,54 @@ SELECT addr, isIPv6String(addr) FROM ( SELECT ['::', '1111::ffff', '::ffff:127.0
└──────────────────┴────────────────────┘
```
## isIPAddressInRange {#isipaddressinrange}
Проверяет попадает ли IP адрес в интервал, заданный в [CIDR](https://en.wikipedia.org/wiki/Classless_Inter-Domain_Routing) нотации.
**Syntax**
``` sql
isIPAddressInRange(address, prefix)
```
Функция принимает IPv4 или IPv6 адрес виде строки. Возвращает `0`, если версия адреса и интервала не совпадают.
**Аргументы**
- `address` — IPv4 или IPv6 адрес. [String](../../sql-reference/data-types/string.md).
- `prefix` — IPv4 или IPv6 подсеть, заданная в CIDR нотации. [String](../../sql-reference/data-types/string.md).
**Возвращаемое значение**
- `1` или `0`.
Тип: [UInt8](../../sql-reference/data-types/int-uint.md).
**Примеры**
Запрос:
``` sql
SELECT isIPAddressInRange('127.0.0.1', '127.0.0.0/8')
```
Результат:
``` text
┌─isIPAddressInRange('127.0.0.1', '127.0.0.0/8')─┐
│ 1 │
└────────────────────────────────────────────────┘
```
Запрос:
``` sql
SELECT isIPAddressInRange('127.0.0.1', 'ffff::/16')
```
Результат:
``` text
┌─isIPAddressInRange('127.0.0.1', 'ffff::/16')─┐
│ 0 │
└──────────────────────────────────────────────┘
```

40
docs/ru/sql-reference/functions/string-search-functions.md
View File

@ -7,7 +7,7 @@ toc_title: "Функции поиска в строках"
Во всех функциях, поиск регистрозависимый по умолчанию. Существуют варианты функций для регистронезависимого поиска.
## position(haystack, needle) {#position}
## position(haystack, needle), locate(haystack, needle) {#position}
Поиск подстроки `needle` в строке `haystack`.
@ -21,8 +21,15 @@ toc_title: "Функции поиска в строках"
position(haystack, needle[, start_pos])
```
``` sql
position(needle IN haystack)
```
Алиас: `locate(haystack, needle[, start_pos])`.
!!! note "Примечание"
Синтаксис `position(needle IN haystack)` обеспечивает совместимость с SQL, функция работает так же, как `position(haystack, needle)`.
**Аргументы**
- `haystack` — строка, по которой выполняется поиск. [Строка](../syntax.md#syntax-string-literal).
@ -70,6 +77,36 @@ SELECT position('Привет, мир!', '!');
└───────────────────────────────┘
```
**Примеры работы функции с синтаксисом POSITION(needle IN haystack)**
Запрос:
```sql
SELECT 1 = position('абв' IN 'абв');
```
Результат:
```text
┌─equals(1, position('абв', 'абв'))─┐
│ 1 │
└───────────────────────────────────┘
```
Запрос:
```sql
SELECT 0 = position('абв' IN '');
```
Результат:
```text
┌─equals(0, position('', 'абв'))─┐
│ 1 │
└────────────────────────────────┘
```
## positionCaseInsensitive {#positioncaseinsensitive}
Такая же, как и [position](#position), но работает без учета регистра. Возвращает позицию в байтах найденной подстроки в строке, начиная с 1.
@ -758,4 +795,3 @@ SELECT countSubstringsCaseInsensitiveUTF8('аБв__АбВ__абв', 'Абв');
│ 3 │
└────────────────────────────────────────────────────────────┘
```

169
programs/copier/ClusterCopier.cpp
View File

@ -599,11 +599,13 @@ TaskStatus ClusterCopier::tryMoveAllPiecesToDestinationTable(const TaskTable & t
toString(current_piece_number));
Settings settings_push = task_cluster->settings_push;
/// It is important, ALTER ATTACH PARTITION must be done synchronously
/// And we will execute this ALTER query on each replica of a shard.
/// It is correct, because this query is idempotent.
settings_push.replication_alter_partitions_sync = 2;
ClusterExecutionMode execution_mode = ClusterExecutionMode::ON_EACH_NODE;
UInt64 max_successful_executions_per_shard = 0;
if (settings_push.replication_alter_partitions_sync == 1)
{
execution_mode = ClusterExecutionMode::ON_EACH_SHARD;
max_successful_executions_per_shard = 1;
}
query_alter_ast_string += " ALTER TABLE " + getQuotedTable(original_table) +
((partition_name == "'all'") ? " ATTACH PARTITION ID " : " ATTACH PARTITION ") + partition_name +
@ -613,14 +615,33 @@ TaskStatus ClusterCopier::tryMoveAllPiecesToDestinationTable(const TaskTable & t
try
{
size_t num_nodes = executeQueryOnCluster(
task_table.cluster_push,
query_alter_ast_string,
settings_push,
PoolMode::GET_MANY,
ClusterExecutionMode::ON_EACH_NODE);
/// Try attach partition on each shard
UInt64 num_nodes = executeQueryOnCluster(
task_table.cluster_push,
query_alter_ast_string,
task_cluster->settings_push,
PoolMode::GET_MANY,
execution_mode,
max_successful_executions_per_shard);
LOG_INFO(log, "Number of nodes that executed ALTER query successfully : {}", toString(num_nodes));
if (settings_push.replication_alter_partitions_sync == 1)
{
LOG_INFO(
log,
"Destination tables {} have been executed alter query successfully on {} shards of {}",
getQuotedTable(task_table.table_push),
num_nodes,
task_table.cluster_push->getShardCount());
if (num_nodes != task_table.cluster_push->getShardCount())
{
return TaskStatus::Error;
}
}
else
{
LOG_INFO(log, "Number of nodes that executed ALTER query successfully : {}", toString(num_nodes));
}
}
catch (...)
{
@ -856,6 +877,16 @@ bool ClusterCopier::tryDropPartitionPiece(
bool ClusterCopier::tryProcessTable(const ConnectionTimeouts & timeouts, TaskTable & task_table)
{
/// Create destination table
TaskStatus task_status = TaskStatus::Error;
task_status = tryCreateDestinationTable(timeouts, task_table);
/// Exit if success
if (task_status != TaskStatus::Finished)
{
LOG_WARNING(log, "Create destination Tale Failed ");
return false;
}
/// An heuristic: if previous shard is already done, then check next one without sleeps due to max_workers constraint
bool previous_shard_is_instantly_finished = false;
@ -932,7 +963,7 @@ bool ClusterCopier::tryProcessTable(const ConnectionTimeouts & timeouts, TaskTab
/// Do not sleep if there is a sequence of already processed shards to increase startup
bool is_unprioritized_task = !previous_shard_is_instantly_finished && shard->priority.is_remote;
TaskStatus task_status = TaskStatus::Error;
task_status = TaskStatus::Error;
bool was_error = false;
has_shard_to_process = true;
for (UInt64 try_num = 0; try_num < max_shard_partition_tries; ++try_num)
@ -1050,6 +1081,44 @@ bool ClusterCopier::tryProcessTable(const ConnectionTimeouts & timeouts, TaskTab
return table_is_done;
}
TaskStatus ClusterCopier::tryCreateDestinationTable(const ConnectionTimeouts & timeouts, TaskTable & task_table)
{
/// Try create original table (if not exists) on each shard
//TaskTable & task_table = task_shard.task_table;
const TaskShardPtr task_shard = task_table.all_shards.at(0);
/// We need to update table definitions for each part, it could be changed after ALTER
task_shard->current_pull_table_create_query = getCreateTableForPullShard(timeouts, *task_shard);
try
{
auto create_query_push_ast
= rewriteCreateQueryStorage(task_shard->current_pull_table_create_query, task_table.table_push, task_table.engine_push_ast);
auto & create = create_query_push_ast->as<ASTCreateQuery &>();
create.if_not_exists = true;
InterpreterCreateQuery::prepareOnClusterQuery(create, context, task_table.cluster_push_name);
String query = queryToString(create_query_push_ast);
LOG_DEBUG(log, "Create destination tables. Query: {}", query);
UInt64 shards = executeQueryOnCluster(task_table.cluster_push, query, task_cluster->settings_push, PoolMode::GET_MANY);
LOG_INFO(
log,
"Destination tables {} have been created on {} shards of {}",
getQuotedTable(task_table.table_push),
shards,
task_table.cluster_push->getShardCount());
if (shards != task_table.cluster_push->getShardCount())
{
return TaskStatus::Error;
}
}
catch (...)
{
tryLogCurrentException(log, "Error while creating original table. Maybe we are not first.");
}
return TaskStatus::Finished;
}
/// Job for copying partition from particular shard.
TaskStatus ClusterCopier::tryProcessPartitionTask(const ConnectionTimeouts & timeouts, ShardPartition & task_partition, bool is_unprioritized_task)
{
@ -1366,8 +1435,17 @@ TaskStatus ClusterCopier::processPartitionPieceTaskImpl(
LOG_DEBUG(log, "Create destination tables. Query: {}", query);
UInt64 shards = executeQueryOnCluster(task_table.cluster_push, query, task_cluster->settings_push, PoolMode::GET_MANY);
LOG_DEBUG(log, "Destination tables {} have been created on {} shards of {}",
getQuotedTable(task_table.table_push), shards, task_table.cluster_push->getShardCount());
LOG_INFO(
log,
"Destination tables {} have been created on {} shards of {}",
getQuotedTable(task_table.table_push),
shards,
task_table.cluster_push->getShardCount());
if (shards != task_table.cluster_push->getShardCount())
{
return TaskStatus::Error;
}
}
/// Do the copying
@ -1477,26 +1555,6 @@ TaskStatus ClusterCopier::processPartitionPieceTaskImpl(
LOG_INFO(log, "Partition {} piece {} copied. But not moved to original destination table.", task_partition.name, toString(current_piece_number));
/// Try create original table (if not exists) on each shard
try
{
auto create_query_push_ast = rewriteCreateQueryStorage(task_shard.current_pull_table_create_query,
task_table.table_push, task_table.engine_push_ast);
auto & create = create_query_push_ast->as<ASTCreateQuery &>();
create.if_not_exists = true;
InterpreterCreateQuery::prepareOnClusterQuery(create, context, task_table.cluster_push_name);
String query = queryToString(create_query_push_ast);
LOG_DEBUG(log, "Create destination tables. Query: {}", query);
UInt64 shards = executeQueryOnCluster(task_table.cluster_push, query, task_cluster->settings_push, PoolMode::GET_MANY);
LOG_DEBUG(log, "Destination tables {} have been created on {} shards of {}", getQuotedTable(task_table.table_push), shards, task_table.cluster_push->getShardCount());
}
catch (...)
{
tryLogCurrentException(log, "Error while creating original table. Maybe we are not first.");
}
/// Finalize the processing, change state of current partition task (and also check is_dirty flag)
{
String state_finished = TaskStateWithOwner::getData(TaskState::Finished, host_id);
@ -1538,33 +1596,36 @@ void ClusterCopier::dropLocalTableIfExists(const DatabaseAndTableName & table_na
interpreter.execute();
}
void ClusterCopier::dropHelpingTablesByPieceNumber(const TaskTable & task_table, size_t current_piece_number)
{
LOG_DEBUG(log, "Removing helping tables piece {}", current_piece_number);
DatabaseAndTableName original_table = task_table.table_push;
DatabaseAndTableName helping_table
= DatabaseAndTableName(original_table.first, original_table.second + "_piece_" + toString(current_piece_number));
String query = "DROP TABLE IF EXISTS " + getQuotedTable(helping_table);
const ClusterPtr & cluster_push = task_table.cluster_push;
Settings settings_push = task_cluster->settings_push;
LOG_DEBUG(log, "Execute distributed DROP TABLE: {}", query);
/// We have to drop partition_piece on each replica
UInt64 num_nodes = executeQueryOnCluster(cluster_push, query, settings_push, PoolMode::GET_MANY, ClusterExecutionMode::ON_EACH_NODE);
LOG_INFO(log, "DROP TABLE query was successfully executed on {} nodes.", toString(num_nodes));
}
void ClusterCopier::dropHelpingTables(const TaskTable & task_table)
{
LOG_DEBUG(log, "Removing helping tables");
for (size_t current_piece_number = 0; current_piece_number < task_table.number_of_splits; ++current_piece_number)
{
DatabaseAndTableName original_table = task_table.table_push;
DatabaseAndTableName helping_table = DatabaseAndTableName(original_table.first, original_table.second + "_piece_" + toString(current_piece_number));
String query = "DROP TABLE IF EXISTS " + getQuotedTable(helping_table);
const ClusterPtr & cluster_push = task_table.cluster_push;
Settings settings_push = task_cluster->settings_push;
LOG_DEBUG(log, "Execute distributed DROP TABLE: {}", query);
/// We have to drop partition_piece on each replica
UInt64 num_nodes = executeQueryOnCluster(
cluster_push, query,
settings_push,
PoolMode::GET_MANY,
ClusterExecutionMode::ON_EACH_NODE);
LOG_DEBUG(log, "DROP TABLE query was successfully executed on {} nodes.", toString(num_nodes));
dropHelpingTablesByPieceNumber(task_table, current_piece_number);
}
}
void ClusterCopier::dropParticularPartitionPieceFromAllHelpingTables(const TaskTable & task_table, const String & partition_name)
{
LOG_DEBUG(log, "Try drop partition partition from all helping tables.");
@ -1586,7 +1647,7 @@ void ClusterCopier::dropParticularPartitionPieceFromAllHelpingTables(const TaskT
PoolMode::GET_MANY,
ClusterExecutionMode::ON_EACH_NODE);
LOG_DEBUG(log, "DROP PARTITION query was successfully executed on {} nodes.", toString(num_nodes));
LOG_INFO(log, "DROP PARTITION query was successfully executed on {} nodes.", toString(num_nodes));
}
LOG_DEBUG(log, "All helping tables dropped partition {}", partition_name);
}

9
programs/copier/ClusterCopier.h
View File

@ -123,12 +123,13 @@ protected:
bool tryDropPartitionPiece(ShardPartition & task_partition, const size_t current_piece_number,
const zkutil::ZooKeeperPtr & zookeeper, const CleanStateClock & clean_state_clock);
static constexpr UInt64 max_table_tries = 1000;
static constexpr UInt64 max_shard_partition_tries = 600;
static constexpr UInt64 max_shard_partition_piece_tries_for_alter = 100;
static constexpr UInt64 max_table_tries = 3;
static constexpr UInt64 max_shard_partition_tries = 3;
static constexpr UInt64 max_shard_partition_piece_tries_for_alter = 3;
bool tryProcessTable(const ConnectionTimeouts & timeouts, TaskTable & task_table);
TaskStatus tryCreateDestinationTable(const ConnectionTimeouts & timeouts, TaskTable & task_table);
/// Job for copying partition from particular shard.
TaskStatus tryProcessPartitionTask(const ConnectionTimeouts & timeouts,
ShardPartition & task_partition,
@ -149,6 +150,8 @@ protected:
void dropHelpingTables(const TaskTable & task_table);
void dropHelpingTablesByPieceNumber(const TaskTable & task_table, size_t current_piece_number);
/// Is used for usage less disk space.
/// After all pieces were successfully moved to original destination
/// table we can get rid of partition pieces (partitions in helping tables).

1
programs/copier/TaskCluster.h
View File

@ -98,6 +98,7 @@ inline void DB::TaskCluster::reloadSettings(const Poco::Util::AbstractConfigurat
set_default_value(settings_pull.max_block_size, 8192UL);
set_default_value(settings_pull.preferred_block_size_bytes, 0);
set_default_value(settings_push.insert_distributed_timeout, 0);
set_default_value(settings_push.replication_alter_partitions_sync, 2);
}
}

16
programs/server/Server.cpp
View File

@ -97,7 +97,7 @@
#endif
#if USE_NURAFT
# include <Server/NuKeeperTCPHandlerFactory.h>
# include <Server/KeeperTCPHandlerFactory.h>
#endif
namespace CurrentMetrics
@ -867,15 +867,15 @@ int Server::main(const std::vector<std::string> & /*args*/)
listen_try = true;
}
if (config().has("test_keeper_server"))
if (config().has("keeper_server"))
{
#if USE_NURAFT
/// Initialize test keeper RAFT. Do nothing if no nu_keeper_server in config.
global_context->initializeNuKeeperStorageDispatcher();
global_context->initializeKeeperStorageDispatcher();
for (const auto & listen_host : listen_hosts)
{
/// TCP NuKeeper
const char * port_name = "test_keeper_server.tcp_port";
/// TCP Keeper
const char * port_name = "keeper_server.tcp_port";
createServer(listen_host, port_name, listen_try, [&](UInt16 port)
{
Poco::Net::ServerSocket socket;
@ -885,9 +885,9 @@ int Server::main(const std::vector<std::string> & /*args*/)
servers_to_start_before_tables->emplace_back(
port_name,
std::make_unique<Poco::Net::TCPServer>(
new NuKeeperTCPHandlerFactory(*this), server_pool, socket, new Poco::Net::TCPServerParams));
new KeeperTCPHandlerFactory(*this), server_pool, socket, new Poco::Net::TCPServerParams));
LOG_INFO(log, "Listening for connections to NuKeeper (tcp): {}", address.toString());
LOG_INFO(log, "Listening for connections to Keeper (tcp): {}", address.toString());
});
}
#else
@ -934,7 +934,7 @@ int Server::main(const std::vector<std::string> & /*args*/)
else
LOG_INFO(log, "Closed connections to servers for tables.");
global_context->shutdownNuKeeperStorageDispatcher();
global_context->shutdownKeeperStorageDispatcher();
}
/** Explicitly destroy Context. It is more convenient than in destructor of Server, because logger is still available.

1
programs/server/config.d/keeper_port.xml Symbolic link
View File

@ -0,0 +1 @@
../../../tests/config/config.d/keeper_port.xml

1
programs/server/config.d/test_keeper_port.xml
View File

@ -1 +0,0 @@
../../../tests/config/config.d/test_keeper_port.xml

1
src/AggregateFunctions/AggregateFunctionWindowFunnel.cpp
View File

@ -6,7 +6,6 @@
#include <DataTypes/DataTypeDateTime.h>
#include <ext/range.h>
#include "registerAggregateFunctions.h"
namespace DB

93
src/AggregateFunctions/AggregateFunctionWindowFunnel.h
View File

@ -19,22 +19,13 @@ namespace ErrorCodes
extern const int BAD_ARGUMENTS;
}
struct ComparePair final
{
template <typename T1, typename T2>
bool operator()(const std::pair<T1, T2> & lhs, const std::pair<T1, T2> & rhs) const
{
return lhs.first == rhs.first ? lhs.second < rhs.second : lhs.first < rhs.first;
}
};
static constexpr auto max_events = 32;
template <typename T>
struct AggregateFunctionWindowFunnelData
{
using TimestampEvent = std::pair<T, UInt8>;
using TimestampEvents = PODArrayWithStackMemory<TimestampEvent, 64>;
using Comparator = ComparePair;
bool sorted = true;
TimestampEvents events_list;
@ -46,7 +37,7 @@ struct AggregateFunctionWindowFunnelData
void add(T timestamp, UInt8 event)
{
// Since most events should have already been sorted by timestamp.
/// Since most events should have already been sorted by timestamp.
if (sorted && events_list.size() > 0)
{
if (events_list.back().first == timestamp)
@ -68,7 +59,7 @@ struct AggregateFunctionWindowFunnelData
/// either sort whole container or do so partially merging ranges afterwards
if (!sorted && !other.sorted)
std::stable_sort(std::begin(events_list), std::end(events_list), Comparator{});
std::stable_sort(std::begin(events_list), std::end(events_list));
else
{
const auto begin = std::begin(events_list);
@ -76,12 +67,12 @@ struct AggregateFunctionWindowFunnelData
const auto end = std::end(events_list);
if (!sorted)
std::stable_sort(begin, middle, Comparator{});
std::stable_sort(begin, middle);
if (!other.sorted)
std::stable_sort(middle, end, Comparator{});
std::stable_sort(middle, end);
std::inplace_merge(begin, middle, end, Comparator{});
std::inplace_merge(begin, middle, end);
}
sorted = true;
@ -91,7 +82,7 @@ struct AggregateFunctionWindowFunnelData
{
if (!sorted)
{
std::stable_sort(std::begin(events_list), std::end(events_list), Comparator{});
std::stable_sort(std::begin(events_list), std::end(events_list));
sorted = true;
}
}
@ -145,14 +136,20 @@ class AggregateFunctionWindowFunnel final
private:
UInt64 window;
UInt8 events_size;
UInt8 strict; // When the 'strict' is set, it applies conditions only for the not repeating values.
UInt8 strict_order; // When the 'strict_order' is set, it doesn't allow interventions of other events.
// In the case of 'A->B->D->C', it stops finding 'A->B->C' at the 'D' and the max event level is 2.
/// When the 'strict' is set, it applies conditions only for the not repeating values.
bool strict;
// Loop through the entire events_list, update the event timestamp value
// The level path must be 1---2---3---...---check_events_size, find the max event level that satisfied the path in the sliding window.
// If found, returns the max event level, else return 0.
// The Algorithm complexity is O(n).
/// When the 'strict_order' is set, it doesn't allow interventions of other events.
/// In the case of 'A->B->D->C', it stops finding 'A->B->C' at the 'D' and the max event level is 2.
bool strict_order;
/// Applies conditions only to events with strictly increasing timestamps
bool strict_increase;
/// Loop through the entire events_list, update the event timestamp value
/// The level path must be 1---2---3---...---check_events_size, find the max event level that satisfied the path in the sliding window.
/// If found, returns the max event level, else return 0.
/// The Algorithm complexity is O(n).
UInt8 getEventLevel(Data & data) const
{
if (data.size() == 0)
@ -162,16 +159,13 @@ private:
data.sort();
/// events_timestamp stores the timestamp that latest i-th level event happen within time window after previous level event.
/// timestamp defaults to -1, which unsigned timestamp value never meet
/// there may be some bugs when UInt64 type timstamp overflows Int64, but it works on most cases.
std::vector<Int64> events_timestamp(events_size, -1);
/// events_timestamp stores the timestamp of the first and previous i-th level event happen within time window
std::vector<std::optional<std::pair<UInt64, UInt64>>> events_timestamp(events_size);
bool first_event = false;
for (const auto & pair : data.events_list)
{
const T & timestamp = pair.first;
const auto & event_idx = pair.second - 1;
if (strict_order && event_idx == -1)
{
if (first_event)
@ -181,31 +175,39 @@ private:
}
else if (event_idx == 0)
{
events_timestamp[0] = timestamp;
events_timestamp[0] = std::make_pair(timestamp, timestamp);
first_event = true;
}
else if (strict && events_timestamp[event_idx] >= 0)
else if (strict && events_timestamp[event_idx].has_value())
{
return event_idx + 1;
}
else if (strict_order && first_event && events_timestamp[event_idx - 1] == -1)
else if (strict_order && first_event && !events_timestamp[event_idx - 1].has_value())
{
for (size_t event = 0; event < events_timestamp.size(); ++event)
{
if (events_timestamp[event] == -1)
if (!events_timestamp[event].has_value())
return event;
}
}
else if (events_timestamp[event_idx - 1] >= 0 && timestamp <= events_timestamp[event_idx - 1] + window)
else if (events_timestamp[event_idx - 1].has_value())
{
events_timestamp[event_idx] = events_timestamp[event_idx - 1];
if (event_idx + 1 == events_size)
return events_size;
auto first_timestamp = events_timestamp[event_idx - 1]->first;
bool time_matched = timestamp <= first_timestamp + window;
if (strict_increase)
time_matched = time_matched && events_timestamp[event_idx - 1]->second < timestamp;
if (time_matched)
{
events_timestamp[event_idx] = std::make_pair(first_timestamp, timestamp);
if (event_idx + 1 == events_size)
return events_size;
}
}
}
for (size_t event = events_timestamp.size(); event > 0; --event)
{
if (events_timestamp[event - 1] >= 0)
if (events_timestamp[event - 1].has_value())
return event;
}
return 0;
@ -223,15 +225,18 @@ public:
events_size = arguments.size() - 1;
window = params.at(0).safeGet<UInt64>();
strict = 0;
strict_order = 0;
strict = false;
strict_order = false;
strict_increase = false;
for (size_t i = 1; i < params.size(); ++i)
{
String option = params.at(i).safeGet<String>();
if (option.compare("strict") == 0)
strict = 1;
else if (option.compare("strict_order") == 0)
strict_order = 1;
if (option == "strict")
strict = true;
else if (option == "strict_order")
strict_order = true;
else if (option == "strict_increase")
strict_increase = true;
else
throw Exception{"Aggregate function " + getName() + " doesn't support a parameter: " + option, ErrorCodes::BAD_ARGUMENTS};
}
@ -253,7 +258,7 @@ public:
{
bool has_event = false;
const auto timestamp = assert_cast<const ColumnVector<T> *>(columns[0])->getData()[row_num];
// reverse iteration and stable sorting are needed for events that are qualified by more than one condition.
/// reverse iteration and stable sorting are needed for events that are qualified by more than one condition.
for (auto i = events_size; i > 0; --i)
{
auto event = assert_cast<const ColumnVector<UInt8> *>(columns[i])->getData()[row_num];

25
src/AggregateFunctions/ReservoirSamplerDeterministic.h
View File

@ -13,6 +13,7 @@
#include <Common/NaNUtils.h>
#include <Poco/Exception.h>
namespace DB
{
namespace ErrorCodes
@ -162,6 +163,11 @@ public:
sorted = false;
}
#if !__clang__
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wclass-memaccess"
#endif
void write(DB::WriteBuffer & buf) const
{
size_t size = samples.size();
@ -169,9 +175,26 @@ public:
DB::writeIntBinary<size_t>(total_values, buf);
for (size_t i = 0; i < size; ++i)
DB::writePODBinary(samples[i], buf);
{
/// There was a mistake in this function.
/// Instead of correctly serializing the elements,
/// it was writing them with uninitialized padding.
/// Here we ensure that padding is zero without changing the protocol.
/// TODO: After implementation of "versioning aggregate function state",
/// change the serialization format.
Element elem;
memset(&elem, 0, sizeof(elem));
elem = samples[i];
DB::writePODBinary(elem, buf);
}
}
#if !__clang__
#pragma GCC diagnostic pop
#endif
private:
/// We allocate some memory on the stack to avoid allocations when there are many objects with a small number of elements.
using Element = std::pair<T, UInt32>;

80
src/Common/IPv6ToBinary.cpp
View File

@ -13,8 +13,7 @@ namespace DB
/// Result array could be indexed with all possible uint8 values without extra check.
/// For values greater than 128 we will store same value as for 128 (all bits set).
constexpr size_t IPV6_MASKS_COUNT = 256;
using RawMaskArray = std::array<uint8_t, IPV6_BINARY_LENGTH>;
using RawMaskArrayV6 = std::array<uint8_t, IPV6_BINARY_LENGTH>;
void IPv6ToRawBinary(const Poco::Net::IPAddress & address, char * res)
{
@ -41,33 +40,86 @@ std::array<char, 16> IPv6ToBinary(const Poco::Net::IPAddress & address)
return res;
}
static constexpr RawMaskArray generateBitMask(size_t prefix)
template <typename RawMaskArrayT>
static constexpr RawMaskArrayT generateBitMask(size_t prefix)
{
if (prefix >= 128)
prefix = 128;
RawMaskArray arr{0};
RawMaskArrayT arr{0};
if (prefix >= arr.size() * 8)
prefix = arr.size() * 8;
size_t i = 0;
for (; prefix >= 8; ++i, prefix -= 8)
arr[i] = 0xff;
if (prefix > 0)
arr[i++] = ~(0xff >> prefix);
while (i < 16)
while (i < arr.size())
arr[i++] = 0x00;
return arr;
}
static constexpr std::array<RawMaskArray, IPV6_MASKS_COUNT> generateBitMasks()
template <typename RawMaskArrayT, size_t masksCount>
static constexpr std::array<RawMaskArrayT, masksCount> generateBitMasks()
{
std::array<RawMaskArray, IPV6_MASKS_COUNT> arr{};
for (size_t i = 0; i < IPV6_MASKS_COUNT; ++i)
arr[i] = generateBitMask(i);
std::array<RawMaskArrayT, masksCount> arr{};
for (size_t i = 0; i < masksCount; ++i)
arr[i] = generateBitMask<RawMaskArrayT>(i);
return arr;
}
const uint8_t * getCIDRMaskIPv6(UInt8 prefix_len)
const std::array<uint8_t, 16> & getCIDRMaskIPv6(UInt8 prefix_len)
{
static constexpr std::array<RawMaskArray, IPV6_MASKS_COUNT> IPV6_RAW_MASK_ARRAY = generateBitMasks();
return IPV6_RAW_MASK_ARRAY[prefix_len].data();
static constexpr auto IPV6_RAW_MASK_ARRAY = generateBitMasks<RawMaskArrayV6, IPV6_MASKS_COUNT>();
return IPV6_RAW_MASK_ARRAY[prefix_len];
}
bool matchIPv4Subnet(UInt32 addr, UInt32 cidr_addr, UInt8 prefix)
{
UInt32 mask = (prefix >= 32) ? 0xffffffffu : ~(0xffffffffu >> prefix);
return (addr & mask) == (cidr_addr & mask);
}
#if defined(__SSE2__)
#include <emmintrin.h>
bool matchIPv6Subnet(const uint8_t * addr, const uint8_t * cidr_addr, UInt8 prefix)
{
uint16_t mask = _mm_movemask_epi8(_mm_cmpeq_epi8(
_mm_loadu_si128(reinterpret_cast<const __m128i *>(addr)),
_mm_loadu_si128(reinterpret_cast<const __m128i *>(cidr_addr))));
mask = ~mask;
if (mask)
{
auto offset = __builtin_ctz(mask);
if (prefix / 8 != offset)
return prefix / 8 < offset;
auto cmpmask = ~(0xff >> (prefix % 8));
return (addr[offset] & cmpmask) == (cidr_addr[offset] & cmpmask);
}
return true;
}
# else
bool matchIPv6Subnet(const uint8_t * addr, const uint8_t * cidr_addr, UInt8 prefix)
{
if (prefix > IPV6_BINARY_LENGTH * 8U)
prefix = IPV6_BINARY_LENGTH * 8U;
size_t i = 0;
for (; prefix >= 8; ++i, prefix -= 8)
{
if (addr[i] != cidr_addr[i])
return false;
}
if (prefix == 0)
return true;
auto mask = ~(0xff >> prefix);
return (addr[i] & mask) == (cidr_addr[i] & mask);
}
#endif // __SSE2__
}

10
src/Common/IPv6ToBinary.h
View File

@ -14,9 +14,13 @@ void IPv6ToRawBinary(const Poco::Net::IPAddress & address, char * res);
/// Convert IP address to 16-byte array with IPv6 data (big endian). If it's an IPv4, map it to IPv6.
std::array<char, 16> IPv6ToBinary(const Poco::Net::IPAddress & address);
/// Returns pointer to 16-byte array containing mask with first `prefix_len` bits set to `1` and `128 - prefix_len` to `0`.
/// Pointer is valid during all program execution time and doesn't require freeing.
/// Returns a reference to 16-byte array containing mask with first `prefix_len` bits set to `1` and `128 - prefix_len` to `0`.
/// The reference is valid during all program execution time.
/// Values of prefix_len greater than 128 interpreted as 128 exactly.
const uint8_t * getCIDRMaskIPv6(UInt8 prefix_len);
const std::array<uint8_t, 16> & getCIDRMaskIPv6(UInt8 prefix_len);
/// Check that address contained in CIDR range
bool matchIPv4Subnet(UInt32 addr, UInt32 cidr_addr, UInt8 prefix);
bool matchIPv6Subnet(const uint8_t * addr, const uint8_t * cidr_addr, UInt8 prefix);
}

4
src/Common/formatIPv6.h
View File

@ -25,7 +25,7 @@ void formatIPv6(const unsigned char * src, char *& dst, uint8_t zeroed_tail_byte
/** Unsafe (no bounds-checking for src nor dst), optimized version of parsing IPv4 string.
*
* Parses the input string `src` and stores binary BE value into buffer pointed by `dst`,
* Parses the input string `src` and stores binary host-endian value into buffer pointed by `dst`,
* which should be long enough.
* That is "127.0.0.1" becomes 0x7f000001.
*
@ -63,7 +63,7 @@ inline bool parseIPv4(const char * src, unsigned char * dst)
/** Unsafe (no bounds-checking for src nor dst), optimized version of parsing IPv6 string.
*
* Slightly altered implementation from http://svn.apache.org/repos/asf/apr/apr/trunk/network_io/unix/inet_pton.c
* Parses the input string `src` and stores binary LE value into buffer pointed by `dst`,
* Parses the input string `src` and stores binary big-endian value into buffer pointed by `dst`,
* which should be long enough. In case of failure zeroes
* IPV6_BINARY_LENGTH bytes of buffer pointed by `dst`.
*

18
src/Common/tests/gtest_DateLUTImpl.cpp
View File

@ -470,8 +470,8 @@ INSTANTIATE_TEST_SUITE_P(AllTimezones_Year2010,
::testing::ValuesIn(allTimezones()),
::testing::ValuesIn(std::initializer_list<TimeRangeParam>{
// Values from tests/date_lut3.cpp
{YYYYMMDDToDay(20101031), YYYYMMDDToDay(20101101), 15 * 60},
{YYYYMMDDToDay(20100328), YYYYMMDDToDay(20100330), 15 * 60}
{YYYYMMDDToDay(20101031), YYYYMMDDToDay(20101101), 10 * 15 * 60},
{YYYYMMDDToDay(20100328), YYYYMMDDToDay(20100330), 10 * 15 * 60}
}))
);
@ -481,7 +481,7 @@ INSTANTIATE_TEST_SUITE_P(AllTimezones_Year1970_WHOLE,
::testing::ValuesIn(allTimezones(false)),
::testing::ValuesIn(std::initializer_list<TimeRangeParam>{
// Values from tests/date_lut3.cpp
{YYYYMMDDToDay(19700101), YYYYMMDDToDay(19701231), 3191 /*53m 11s*/},
{YYYYMMDDToDay(19700101), YYYYMMDDToDay(19701231), 10 * 3191 /*53m 11s*/},
}))
);
@ -491,7 +491,7 @@ INSTANTIATE_TEST_SUITE_P(AllTimezones_Year2010_WHOLE,
::testing::ValuesIn(allTimezones(false)),
::testing::ValuesIn(std::initializer_list<TimeRangeParam>{
// Values from tests/date_lut3.cpp
{YYYYMMDDToDay(20100101), YYYYMMDDToDay(20101231), 3191 /*53m 11s*/},
{YYYYMMDDToDay(20100101), YYYYMMDDToDay(20101231), 10 * 3191 /*53m 11s*/},
}))
);
@ -501,7 +501,7 @@ INSTANTIATE_TEST_SUITE_P(AllTimezones_Year2020_WHOLE,
::testing::ValuesIn(allTimezones()),
::testing::ValuesIn(std::initializer_list<TimeRangeParam>{
// Values from tests/date_lut3.cpp
{YYYYMMDDToDay(20200101), YYYYMMDDToDay(20201231), 3191 /*53m 11s*/},
{YYYYMMDDToDay(20200101), YYYYMMDDToDay(20201231), 10 * 3191 /*53m 11s*/},
}))
);
@ -510,8 +510,8 @@ INSTANTIATE_TEST_SUITE_P(AllTimezones_PreEpoch,
::testing::Combine(
::testing::ValuesIn(allTimezones(false)),
::testing::ValuesIn(std::initializer_list<TimeRangeParam>{
{YYYYMMDDToDay(19500101), YYYYMMDDToDay(19600101), 15 * 60},
{YYYYMMDDToDay(19300101), YYYYMMDDToDay(19350101), 11 * 15 * 60}
{YYYYMMDDToDay(19500101), YYYYMMDDToDay(19600101), 10 * 15 * 60},
{YYYYMMDDToDay(19300101), YYYYMMDDToDay(19350101), 10 * 11 * 15 * 60}
}))
);
@ -520,8 +520,8 @@ INSTANTIATE_TEST_SUITE_P(AllTimezones_Year1970,
::testing::Combine(
::testing::ValuesIn(allTimezones(false)),
::testing::ValuesIn(std::initializer_list<TimeRangeParam>{
{YYYYMMDDToDay(19700101), YYYYMMDDToDay(19700201), 15 * 60},
{YYYYMMDDToDay(19700101), YYYYMMDDToDay(19701231), 11 * 13 * 17}
{YYYYMMDDToDay(19700101), YYYYMMDDToDay(19700201), 10 * 15 * 60},
{YYYYMMDDToDay(19700101), YYYYMMDDToDay(19701231), 10 * 11 * 13 * 17}
// // 11 was chosen as a number which can't divide product of 2-combinarions of (7, 24, 60),
// // to reduce likelehood of hitting same hour/minute/second values for different days.
// // + 12 is just to make sure that last day is covered fully.

34
src/Coordination/NuKeeperLogStore.cpp → src/Coordination/KeeperLogStore.cpp
View File

@ -1,40 +1,40 @@
#include <Coordination/NuKeeperLogStore.h>
#include <Coordination/KeeperLogStore.h>
namespace DB
{
NuKeeperLogStore::NuKeeperLogStore(const std::string & changelogs_path, size_t rotate_interval_, bool force_sync_)
: log(&Poco::Logger::get("NuKeeperLogStore"))
KeeperLogStore::KeeperLogStore(const std::string & changelogs_path, size_t rotate_interval_, bool force_sync_)
: log(&Poco::Logger::get("KeeperLogStore"))
, changelog(changelogs_path, rotate_interval_, log)
, force_sync(force_sync_)
{
}
size_t NuKeeperLogStore::start_index() const
size_t KeeperLogStore::start_index() const
{
std::lock_guard lock(changelog_lock);
return changelog.getStartIndex();
}
void NuKeeperLogStore::init(size_t last_commited_log_index, size_t logs_to_keep)
void KeeperLogStore::init(size_t last_commited_log_index, size_t logs_to_keep)
{
std::lock_guard lock(changelog_lock);
changelog.readChangelogAndInitWriter(last_commited_log_index, logs_to_keep);
}
size_t NuKeeperLogStore::next_slot() const
size_t KeeperLogStore::next_slot() const
{
std::lock_guard lock(changelog_lock);
return changelog.getNextEntryIndex();
}
nuraft::ptr<nuraft::log_entry> NuKeeperLogStore::last_entry() const
nuraft::ptr<nuraft::log_entry> KeeperLogStore::last_entry() const
{
std::lock_guard lock(changelog_lock);
return changelog.getLastEntry();
}
size_t NuKeeperLogStore::append(nuraft::ptr<nuraft::log_entry> & entry)
size_t KeeperLogStore::append(nuraft::ptr<nuraft::log_entry> & entry)
{
std::lock_guard lock(changelog_lock);
size_t idx = changelog.getNextEntryIndex();
@ -43,25 +43,25 @@ size_t NuKeeperLogStore::append(nuraft::ptr<nuraft::log_entry> & entry)
}
void NuKeeperLogStore::write_at(size_t index, nuraft::ptr<nuraft::log_entry> & entry)
void KeeperLogStore::write_at(size_t index, nuraft::ptr<nuraft::log_entry> & entry)
{
std::lock_guard lock(changelog_lock);
changelog.writeAt(index, entry, force_sync);
}
nuraft::ptr<std::vector<nuraft::ptr<nuraft::log_entry>>> NuKeeperLogStore::log_entries(size_t start, size_t end)
nuraft::ptr<std::vector<nuraft::ptr<nuraft::log_entry>>> KeeperLogStore::log_entries(size_t start, size_t end)
{
std::lock_guard lock(changelog_lock);
return changelog.getLogEntriesBetween(start, end);
}
nuraft::ptr<nuraft::log_entry> NuKeeperLogStore::entry_at(size_t index)
nuraft::ptr<nuraft::log_entry> KeeperLogStore::entry_at(size_t index)
{
std::lock_guard lock(changelog_lock);
return changelog.entryAt(index);
}
size_t NuKeeperLogStore::term_at(size_t index)
size_t KeeperLogStore::term_at(size_t index)
{
std::lock_guard lock(changelog_lock);
auto entry = changelog.entryAt(index);
@ -70,33 +70,33 @@ size_t NuKeeperLogStore::term_at(size_t index)
return 0;
}
nuraft::ptr<nuraft::buffer> NuKeeperLogStore::pack(size_t index, int32_t cnt)
nuraft::ptr<nuraft::buffer> KeeperLogStore::pack(size_t index, int32_t cnt)
{
std::lock_guard lock(changelog_lock);
return changelog.serializeEntriesToBuffer(index, cnt);
}
bool NuKeeperLogStore::compact(size_t last_log_index)
bool KeeperLogStore::compact(size_t last_log_index)
{
std::lock_guard lock(changelog_lock);
changelog.compact(last_log_index);
return true;
}
bool NuKeeperLogStore::flush()
bool KeeperLogStore::flush()
{
std::lock_guard lock(changelog_lock);
changelog.flush();
return true;
}
void NuKeeperLogStore::apply_pack(size_t index, nuraft::buffer & pack)
void KeeperLogStore::apply_pack(size_t index, nuraft::buffer & pack)
{
std::lock_guard lock(changelog_lock);
changelog.applyEntriesFromBuffer(index, pack, force_sync);
}
size_t NuKeeperLogStore::size() const
size_t KeeperLogStore::size() const
{
std::lock_guard lock(changelog_lock);
return changelog.size();

4
src/Coordination/NuKeeperLogStore.h → src/Coordination/KeeperLogStore.h
View File

@ -9,10 +9,10 @@
namespace DB
{
class NuKeeperLogStore : public nuraft::log_store
class KeeperLogStore : public nuraft::log_store
{
public:
NuKeeperLogStore(const std::string & changelogs_path, size_t rotate_interval_, bool force_sync_);
KeeperLogStore(const std::string & changelogs_path, size_t rotate_interval_, bool force_sync_);
void init(size_t last_commited_log_index, size_t logs_to_keep);

45
src/Coordination/NuKeeperServer.cpp → src/Coordination/KeeperServer.cpp
View File

@ -1,7 +1,7 @@
#include <Coordination/NuKeeperServer.h>
#include <Coordination/KeeperServer.h>
#include <Coordination/LoggerWrapper.h>
#include <Coordination/NuKeeperStateMachine.h>
#include <Coordination/NuKeeperStateManager.h>
#include <Coordination/KeeperStateMachine.h>
#include <Coordination/KeeperStateManager.h>
#include <Coordination/WriteBufferFromNuraftBuffer.h>
#include <Coordination/ReadBufferFromNuraftBuffer.h>
#include <IO/ReadHelpers.h>
@ -18,7 +18,7 @@ namespace ErrorCodes
extern const int RAFT_ERROR;
}
NuKeeperServer::NuKeeperServer(
KeeperServer::KeeperServer(
int server_id_,
const CoordinationSettingsPtr & coordination_settings_,
const Poco::Util::AbstractConfiguration & config,
@ -26,15 +26,18 @@ NuKeeperServer::NuKeeperServer(
SnapshotsQueue & snapshots_queue_)
: server_id(server_id_)
, coordination_settings(coordination_settings_)
, state_machine(nuraft::cs_new<NuKeeperStateMachine>(responses_queue_, snapshots_queue_, config.getString("test_keeper_server.snapshot_storage_path", config.getString("path", DBMS_DEFAULT_PATH) + "coordination/snapshots"), coordination_settings))
, state_manager(nuraft::cs_new<NuKeeperStateManager>(server_id, "test_keeper_server", config, coordination_settings))
, state_machine(nuraft::cs_new<KeeperStateMachine>(
responses_queue_, snapshots_queue_,
config.getString("keeper_server.snapshot_storage_path", config.getString("path", DBMS_DEFAULT_PATH) + "coordination/snapshots"),
coordination_settings))
, state_manager(nuraft::cs_new<KeeperStateManager>(server_id, "keeper_server", config, coordination_settings))
, responses_queue(responses_queue_)
{
if (coordination_settings->quorum_reads)
LOG_WARNING(&Poco::Logger::get("NuKeeperServer"), "Quorum reads enabled, NuKeeper will work slower.");
LOG_WARNING(&Poco::Logger::get("KeeperServer"), "Quorum reads enabled, Keeper will work slower.");
}
void NuKeeperServer::startup()
void KeeperServer::startup()
{
state_machine->init();
@ -84,13 +87,13 @@ void NuKeeperServer::startup()
throw Exception(ErrorCodes::RAFT_ERROR, "Cannot allocate RAFT instance");
}
void NuKeeperServer::shutdown()
void KeeperServer::shutdown()
{
state_machine->shutdownStorage();
state_manager->flushLogStore();
auto timeout = coordination_settings->shutdown_timeout.totalSeconds();
if (!launcher.shutdown(timeout))
LOG_WARNING(&Poco::Logger::get("NuKeeperServer"), "Failed to shutdown RAFT server in {} seconds", timeout);
LOG_WARNING(&Poco::Logger::get("KeeperServer"), "Failed to shutdown RAFT server in {} seconds", timeout);
}
namespace
@ -106,7 +109,7 @@ nuraft::ptr<nuraft::buffer> getZooKeeperLogEntry(int64_t session_id, const Coord
}
void NuKeeperServer::putRequest(const NuKeeperStorage::RequestForSession & request_for_session)
void KeeperServer::putRequest(const KeeperStorage::RequestForSession & request_for_session)
{
auto [session_id, request] = request_for_session;
if (!coordination_settings->quorum_reads && isLeaderAlive() && request->isReadRequest())
@ -123,29 +126,29 @@ void NuKeeperServer::putRequest(const NuKeeperStorage::RequestForSession & reque
auto result = raft_instance->append_entries(entries);
if (!result->get_accepted())
{
NuKeeperStorage::ResponsesForSessions responses;
KeeperStorage::ResponsesForSessions responses;
auto response = request->makeResponse();
response->xid = request->xid;
response->zxid = 0;
response->error = Coordination::Error::ZOPERATIONTIMEOUT;
responses_queue.push(DB::NuKeeperStorage::ResponseForSession{session_id, response});
responses_queue.push(DB::KeeperStorage::ResponseForSession{session_id, response});
}
if (result->get_result_code() == nuraft::cmd_result_code::TIMEOUT)
{
NuKeeperStorage::ResponsesForSessions responses;
KeeperStorage::ResponsesForSessions responses;
auto response = request->makeResponse();
response->xid = request->xid;
response->zxid = 0;
response->error = Coordination::Error::ZOPERATIONTIMEOUT;
responses_queue.push(DB::NuKeeperStorage::ResponseForSession{session_id, response});
responses_queue.push(DB::KeeperStorage::ResponseForSession{session_id, response});
}
else if (result->get_result_code() != nuraft::cmd_result_code::OK)
throw Exception(ErrorCodes::RAFT_ERROR, "Requests result failed with code {} and message: '{}'", result->get_result_code(), result->get_result_str());
}
}
int64_t NuKeeperServer::getSessionID(int64_t session_timeout_ms)
int64_t KeeperServer::getSessionID(int64_t session_timeout_ms)
{
auto entry = nuraft::buffer::alloc(sizeof(int64_t));
/// Just special session request
@ -170,17 +173,17 @@ int64_t NuKeeperServer::getSessionID(int64_t session_timeout_ms)
return bs_resp.get_i64();
}
bool NuKeeperServer::isLeader() const
bool KeeperServer::isLeader() const
{
return raft_instance->is_leader();
}
bool NuKeeperServer::isLeaderAlive() const
bool KeeperServer::isLeaderAlive() const
{
return raft_instance->is_leader_alive();
}
nuraft::cb_func::ReturnCode NuKeeperServer::callbackFunc(nuraft::cb_func::Type type, nuraft::cb_func::Param * /* param */)
nuraft::cb_func::ReturnCode KeeperServer::callbackFunc(nuraft::cb_func::Type type, nuraft::cb_func::Param * /* param */)
{
size_t last_commited = state_machine->last_commit_index();
size_t next_index = state_manager->getLogStore()->next_slot();
@ -240,7 +243,7 @@ nuraft::cb_func::ReturnCode NuKeeperServer::callbackFunc(nuraft::cb_func::Type t
}
}
void NuKeeperServer::waitInit()
void KeeperServer::waitInit()
{
std::unique_lock lock(initialized_mutex);
int64_t timeout = coordination_settings->startup_timeout.totalMilliseconds();
@ -248,7 +251,7 @@ void NuKeeperServer::waitInit()
throw Exception(ErrorCodes::RAFT_ERROR, "Failed to wait RAFT initialization");
}
std::unordered_set<int64_t> NuKeeperServer::getDeadSessions()
std::unordered_set<int64_t> KeeperServer::getDeadSessions()
{
return state_machine->getDeadSessions();
}

16
src/Coordination/NuKeeperServer.h → src/Coordination/KeeperServer.h
View File

@ -2,25 +2,25 @@
#include <libnuraft/nuraft.hxx> // Y_IGNORE
#include <Coordination/InMemoryLogStore.h>
#include <Coordination/NuKeeperStateManager.h>
#include <Coordination/NuKeeperStateMachine.h>
#include <Coordination/NuKeeperStorage.h>
#include <Coordination/KeeperStateManager.h>
#include <Coordination/KeeperStateMachine.h>
#include <Coordination/KeeperStorage.h>
#include <Coordination/CoordinationSettings.h>
#include <unordered_map>
namespace DB
{
class NuKeeperServer
class KeeperServer
{
private:
int server_id;
CoordinationSettingsPtr coordination_settings;
nuraft::ptr<NuKeeperStateMachine> state_machine;
nuraft::ptr<KeeperStateMachine> state_machine;
nuraft::ptr<NuKeeperStateManager> state_manager;
nuraft::ptr<KeeperStateManager> state_manager;
nuraft::raft_launcher launcher;
@ -38,7 +38,7 @@ private:
nuraft::cb_func::ReturnCode callbackFunc(nuraft::cb_func::Type type, nuraft::cb_func::Param * param);
public:
NuKeeperServer(
KeeperServer(
int server_id_,
const CoordinationSettingsPtr & coordination_settings_,
const Poco::Util::AbstractConfiguration & config,
@ -47,7 +47,7 @@ public:
void startup();
void putRequest(const NuKeeperStorage::RequestForSession & request);
void putRequest(const KeeperStorage::RequestForSession & request);
int64_t getSessionID(int64_t session_timeout_ms);

44
src/Coordination/NuKeeperSnapshotManager.cpp → src/Coordination/KeeperSnapshotManager.cpp
View File

@ -1,4 +1,4 @@
#include <Coordination/NuKeeperSnapshotManager.h>
#include <Coordination/KeeperSnapshotManager.h>
#include <IO/WriteHelpers.h>
#include <Compression/CompressedReadBuffer.h>
#include <Compression/CompressedWriteBuffer.h>
@ -51,7 +51,7 @@ namespace
return "/";
}
void writeNode(const NuKeeperStorage::Node & node, WriteBuffer & out)
void writeNode(const KeeperStorage::Node & node, WriteBuffer & out)
{
writeBinary(node.data, out);
@ -81,7 +81,7 @@ namespace
writeBinary(node.seq_num, out);
}
void readNode(NuKeeperStorage::Node & node, ReadBuffer & in)
void readNode(KeeperStorage::Node & node, ReadBuffer & in)
{
readBinary(node.data, in);
@ -132,7 +132,7 @@ namespace
}
void NuKeeperStorageSnapshot::serialize(const NuKeeperStorageSnapshot & snapshot, WriteBuffer & out)
void KeeperStorageSnapshot::serialize(const KeeperStorageSnapshot & snapshot, WriteBuffer & out)
{
writeBinary(static_cast<uint8_t>(snapshot.version), out);
serializeSnapshotMetadata(snapshot.snapshot_meta, out);
@ -159,7 +159,7 @@ void NuKeeperStorageSnapshot::serialize(const NuKeeperStorageSnapshot & snapshot
}
}
SnapshotMetadataPtr NuKeeperStorageSnapshot::deserialize(NuKeeperStorage & storage, ReadBuffer & in)
SnapshotMetadataPtr KeeperStorageSnapshot::deserialize(KeeperStorage & storage, ReadBuffer & in)
{
uint8_t version;
readBinary(version, in);
@ -180,7 +180,7 @@ SnapshotMetadataPtr NuKeeperStorageSnapshot::deserialize(NuKeeperStorage & stora
{
std::string path;
readBinary(path, in);
NuKeeperStorage::Node node;
KeeperStorage::Node node;
readNode(node, in);
storage.container.insertOrReplace(path, node);
if (node.stat.ephemeralOwner != 0)
@ -194,7 +194,7 @@ SnapshotMetadataPtr NuKeeperStorageSnapshot::deserialize(NuKeeperStorage & stora
if (itr.key != "/")
{
auto parent_path = parentPath(itr.key);
storage.container.updateValue(parent_path, [&path = itr.key] (NuKeeperStorage::Node & value) { value.children.insert(getBaseName(path)); });
storage.container.updateValue(parent_path, [&path = itr.key] (KeeperStorage::Node & value) { value.children.insert(getBaseName(path)); });
}
}
@ -214,7 +214,7 @@ SnapshotMetadataPtr NuKeeperStorageSnapshot::deserialize(NuKeeperStorage & stora
return result;
}
NuKeeperStorageSnapshot::NuKeeperStorageSnapshot(NuKeeperStorage * storage_, size_t up_to_log_idx_)
KeeperStorageSnapshot::KeeperStorageSnapshot(KeeperStorage * storage_, size_t up_to_log_idx_)
: storage(storage_)
, snapshot_meta(std::make_shared<SnapshotMetadata>(up_to_log_idx_, 0, std::make_shared<nuraft::cluster_config>()))
, session_id(storage->session_id_counter)
@ -225,7 +225,7 @@ NuKeeperStorageSnapshot::NuKeeperStorageSnapshot(NuKeeperStorage * storage_, siz
session_and_timeout = storage->getActiveSessions();
}
NuKeeperStorageSnapshot::NuKeeperStorageSnapshot(NuKeeperStorage * storage_, const SnapshotMetadataPtr & snapshot_meta_)
KeeperStorageSnapshot::KeeperStorageSnapshot(KeeperStorage * storage_, const SnapshotMetadataPtr & snapshot_meta_)
: storage(storage_)
, snapshot_meta(snapshot_meta_)
, session_id(storage->session_id_counter)
@ -236,12 +236,12 @@ NuKeeperStorageSnapshot::NuKeeperStorageSnapshot(NuKeeperStorage * storage_, con
session_and_timeout = storage->getActiveSessions();
}
NuKeeperStorageSnapshot::~NuKeeperStorageSnapshot()
KeeperStorageSnapshot::~KeeperStorageSnapshot()
{
storage->disableSnapshotMode();
}
NuKeeperSnapshotManager::NuKeeperSnapshotManager(const std::string & snapshots_path_, size_t snapshots_to_keep_, size_t storage_tick_time_)
KeeperSnapshotManager::KeeperSnapshotManager(const std::string & snapshots_path_, size_t snapshots_to_keep_, size_t storage_tick_time_)
: snapshots_path(snapshots_path_)
, snapshots_to_keep(snapshots_to_keep_)
, storage_tick_time(storage_tick_time_)
@ -266,7 +266,7 @@ NuKeeperSnapshotManager::NuKeeperSnapshotManager(const std::string & snapshots_p
}
std::string NuKeeperSnapshotManager::serializeSnapshotBufferToDisk(nuraft::buffer & buffer, size_t up_to_log_idx)
std::string KeeperSnapshotManager::serializeSnapshotBufferToDisk(nuraft::buffer & buffer, size_t up_to_log_idx)
{
ReadBufferFromNuraftBuffer reader(buffer);
@ -287,7 +287,7 @@ std::string NuKeeperSnapshotManager::serializeSnapshotBufferToDisk(nuraft::buffe
return new_snapshot_path;
}
nuraft::ptr<nuraft::buffer> NuKeeperSnapshotManager::deserializeLatestSnapshotBufferFromDisk()
nuraft::ptr<nuraft::buffer> KeeperSnapshotManager::deserializeLatestSnapshotBufferFromDisk()
{
while (!existing_snapshots.empty())
{
@ -307,7 +307,7 @@ nuraft::ptr<nuraft::buffer> NuKeeperSnapshotManager::deserializeLatestSnapshotBu
return nullptr;
}
nuraft::ptr<nuraft::buffer> NuKeeperSnapshotManager::deserializeSnapshotBufferFromDisk(size_t up_to_log_idx) const
nuraft::ptr<nuraft::buffer> KeeperSnapshotManager::deserializeSnapshotBufferFromDisk(size_t up_to_log_idx) const
{
const std::string & snapshot_path = existing_snapshots.at(up_to_log_idx);
WriteBufferFromNuraftBuffer writer;
@ -316,26 +316,26 @@ nuraft::ptr<nuraft::buffer> NuKeeperSnapshotManager::deserializeSnapshotBufferFr
return writer.getBuffer();
}
nuraft::ptr<nuraft::buffer> NuKeeperSnapshotManager::serializeSnapshotToBuffer(const NuKeeperStorageSnapshot & snapshot)
nuraft::ptr<nuraft::buffer> KeeperSnapshotManager::serializeSnapshotToBuffer(const KeeperStorageSnapshot & snapshot)
{
WriteBufferFromNuraftBuffer writer;
CompressedWriteBuffer compressed_writer(writer);
NuKeeperStorageSnapshot::serialize(snapshot, compressed_writer);
KeeperStorageSnapshot::serialize(snapshot, compressed_writer);
compressed_writer.finalize();
return writer.getBuffer();
}
SnapshotMetaAndStorage NuKeeperSnapshotManager::deserializeSnapshotFromBuffer(nuraft::ptr<nuraft::buffer> buffer) const
SnapshotMetaAndStorage KeeperSnapshotManager::deserializeSnapshotFromBuffer(nuraft::ptr<nuraft::buffer> buffer) const
{
ReadBufferFromNuraftBuffer reader(buffer);
CompressedReadBuffer compressed_reader(reader);
auto storage = std::make_unique<NuKeeperStorage>(storage_tick_time);
auto snapshot_metadata = NuKeeperStorageSnapshot::deserialize(*storage, compressed_reader);
auto storage = std::make_unique<KeeperStorage>(storage_tick_time);
auto snapshot_metadata = KeeperStorageSnapshot::deserialize(*storage, compressed_reader);
return std::make_pair(snapshot_metadata, std::move(storage));
}
SnapshotMetaAndStorage NuKeeperSnapshotManager::restoreFromLatestSnapshot()
SnapshotMetaAndStorage KeeperSnapshotManager::restoreFromLatestSnapshot()
{
if (existing_snapshots.empty())
return {};
@ -346,13 +346,13 @@ SnapshotMetaAndStorage NuKeeperSnapshotManager::restoreFromLatestSnapshot()
return deserializeSnapshotFromBuffer(buffer);
}
void NuKeeperSnapshotManager::removeOutdatedSnapshotsIfNeeded()
void KeeperSnapshotManager::removeOutdatedSnapshotsIfNeeded()
{
while (existing_snapshots.size() > snapshots_to_keep)
removeSnapshot(existing_snapshots.begin()->first);
}
void NuKeeperSnapshotManager::removeSnapshot(size_t log_idx)
void KeeperSnapshotManager::removeSnapshot(size_t log_idx)
{
auto itr = existing_snapshots.find(log_idx);
if (itr == existing_snapshots.end())

32
src/Coordination/NuKeeperSnapshotManager.h → src/Coordination/KeeperSnapshotManager.h
View File

@ -1,6 +1,6 @@
#pragma once
#include <libnuraft/nuraft.hxx> // Y_IGNORE
#include <Coordination/NuKeeperStorage.h>
#include <Coordination/KeeperStorage.h>
#include <IO/WriteBuffer.h>
#include <IO/ReadBuffer.h>
@ -15,42 +15,42 @@ enum SnapshotVersion : uint8_t
V0 = 0,
};
struct NuKeeperStorageSnapshot
struct KeeperStorageSnapshot
{
public:
NuKeeperStorageSnapshot(NuKeeperStorage * storage_, size_t up_to_log_idx_);
KeeperStorageSnapshot(KeeperStorage * storage_, size_t up_to_log_idx_);
NuKeeperStorageSnapshot(NuKeeperStorage * storage_, const SnapshotMetadataPtr & snapshot_meta_);
~NuKeeperStorageSnapshot();
KeeperStorageSnapshot(KeeperStorage * storage_, const SnapshotMetadataPtr & snapshot_meta_);
~KeeperStorageSnapshot();
static void serialize(const NuKeeperStorageSnapshot & snapshot, WriteBuffer & out);
static void serialize(const KeeperStorageSnapshot & snapshot, WriteBuffer & out);
static SnapshotMetadataPtr deserialize(NuKeeperStorage & storage, ReadBuffer & in);
static SnapshotMetadataPtr deserialize(KeeperStorage & storage, ReadBuffer & in);
NuKeeperStorage * storage;
KeeperStorage * storage;
SnapshotVersion version = SnapshotVersion::V0;
SnapshotMetadataPtr snapshot_meta;
int64_t session_id;
size_t snapshot_container_size;
NuKeeperStorage::Container::const_iterator begin;
KeeperStorage::Container::const_iterator begin;
SessionAndTimeout session_and_timeout;
};
using NuKeeperStorageSnapshotPtr = std::shared_ptr<NuKeeperStorageSnapshot>;
using CreateSnapshotCallback = std::function<void(NuKeeperStorageSnapshotPtr &&)>;
using KeeperStorageSnapshotPtr = std::shared_ptr<KeeperStorageSnapshot>;
using CreateSnapshotCallback = std::function<void(KeeperStorageSnapshotPtr &&)>;
using SnapshotMetaAndStorage = std::pair<SnapshotMetadataPtr, NuKeeperStoragePtr>;
using SnapshotMetaAndStorage = std::pair<SnapshotMetadataPtr, KeeperStoragePtr>;
class NuKeeperSnapshotManager
class KeeperSnapshotManager
{
public:
NuKeeperSnapshotManager(const std::string & snapshots_path_, size_t snapshots_to_keep_, size_t storage_tick_time_ = 500);
KeeperSnapshotManager(const std::string & snapshots_path_, size_t snapshots_to_keep_, size_t storage_tick_time_ = 500);
SnapshotMetaAndStorage restoreFromLatestSnapshot();
static nuraft::ptr<nuraft::buffer> serializeSnapshotToBuffer(const NuKeeperStorageSnapshot & snapshot);
static nuraft::ptr<nuraft::buffer> serializeSnapshotToBuffer(const KeeperStorageSnapshot & snapshot);
std::string serializeSnapshotBufferToDisk(nuraft::buffer & buffer, size_t up_to_log_idx);
SnapshotMetaAndStorage deserializeSnapshotFromBuffer(nuraft::ptr<nuraft::buffer> buffer) const;
@ -82,7 +82,7 @@ private:
struct CreateSnapshotTask
{
NuKeeperStorageSnapshotPtr snapshot;
KeeperStorageSnapshotPtr snapshot;
CreateSnapshotCallback create_snapshot;
};

46
src/Coordination/NuKeeperStateMachine.cpp → src/Coordination/KeeperStateMachine.cpp
View File

@ -1,9 +1,9 @@
#include <Coordination/NuKeeperStateMachine.h>
#include <Coordination/KeeperStateMachine.h>
#include <Coordination/ReadBufferFromNuraftBuffer.h>
#include <Coordination/WriteBufferFromNuraftBuffer.h>
#include <IO/ReadHelpers.h>
#include <Common/ZooKeeper/ZooKeeperIO.h>
#include <Coordination/NuKeeperSnapshotManager.h>
#include <Coordination/KeeperSnapshotManager.h>
#include <future>
namespace DB
@ -14,10 +14,10 @@ namespace ErrorCodes
extern const int LOGICAL_ERROR;
}
NuKeeperStorage::RequestForSession parseRequest(nuraft::buffer & data)
KeeperStorage::RequestForSession parseRequest(nuraft::buffer & data)
{
ReadBufferFromNuraftBuffer buffer(data);
NuKeeperStorage::RequestForSession request_for_session;
KeeperStorage::RequestForSession request_for_session;
readIntBinary(request_for_session.session_id, buffer);
int32_t length;
@ -36,17 +36,17 @@ NuKeeperStorage::RequestForSession parseRequest(nuraft::buffer & data)
return request_for_session;
}
NuKeeperStateMachine::NuKeeperStateMachine(ResponsesQueue & responses_queue_, SnapshotsQueue & snapshots_queue_, const std::string & snapshots_path_, const CoordinationSettingsPtr & coordination_settings_)
KeeperStateMachine::KeeperStateMachine(ResponsesQueue & responses_queue_, SnapshotsQueue & snapshots_queue_, const std::string & snapshots_path_, const CoordinationSettingsPtr & coordination_settings_)
: coordination_settings(coordination_settings_)
, snapshot_manager(snapshots_path_, coordination_settings->snapshots_to_keep, coordination_settings->dead_session_check_period_ms.totalMicroseconds())
, responses_queue(responses_queue_)
, snapshots_queue(snapshots_queue_)
, last_committed_idx(0)
, log(&Poco::Logger::get("NuKeeperStateMachine"))
, log(&Poco::Logger::get("KeeperStateMachine"))
{
}
void NuKeeperStateMachine::init()
void KeeperStateMachine::init()
{
/// Do everything without mutexes, no other threads exist.
LOG_DEBUG(log, "Totally have {} snapshots", snapshot_manager.totalSnapshots());
@ -85,10 +85,10 @@ void NuKeeperStateMachine::init()
}
if (!storage)
storage = std::make_unique<NuKeeperStorage>(coordination_settings->dead_session_check_period_ms.totalMilliseconds());
storage = std::make_unique<KeeperStorage>(coordination_settings->dead_session_check_period_ms.totalMilliseconds());
}
nuraft::ptr<nuraft::buffer> NuKeeperStateMachine::commit(const size_t log_idx, nuraft::buffer & data)
nuraft::ptr<nuraft::buffer> KeeperStateMachine::commit(const size_t log_idx, nuraft::buffer & data)
{
if (data.size() == sizeof(int64_t))
{
@ -109,7 +109,7 @@ nuraft::ptr<nuraft::buffer> NuKeeperStateMachine::commit(const size_t log_idx, n
else
{
auto request_for_session = parseRequest(data);
NuKeeperStorage::ResponsesForSessions responses_for_sessions;
KeeperStorage::ResponsesForSessions responses_for_sessions;
{
std::lock_guard lock(storage_lock);
responses_for_sessions = storage->processRequest(request_for_session.request, request_for_session.session_id, log_idx);
@ -122,7 +122,7 @@ nuraft::ptr<nuraft::buffer> NuKeeperStateMachine::commit(const size_t log_idx, n
}
}
bool NuKeeperStateMachine::apply_snapshot(nuraft::snapshot & s)
bool KeeperStateMachine::apply_snapshot(nuraft::snapshot & s)
{
LOG_DEBUG(log, "Applying snapshot {}", s.get_last_log_idx());
nuraft::ptr<nuraft::buffer> latest_snapshot_ptr;
@ -142,14 +142,14 @@ bool NuKeeperStateMachine::apply_snapshot(nuraft::snapshot & s)
return true;
}
nuraft::ptr<nuraft::snapshot> NuKeeperStateMachine::last_snapshot()
nuraft::ptr<nuraft::snapshot> KeeperStateMachine::last_snapshot()
{
/// Just return the latest snapshot.
std::lock_guard<std::mutex> lock(snapshots_lock);
return latest_snapshot_meta;
}
void NuKeeperStateMachine::create_snapshot(
void KeeperStateMachine::create_snapshot(
nuraft::snapshot & s,
nuraft::async_result<bool>::handler_type & when_done)
{
@ -160,10 +160,10 @@ void NuKeeperStateMachine::create_snapshot(
CreateSnapshotTask snapshot_task;
{
std::lock_guard lock(storage_lock);
snapshot_task.snapshot = std::make_shared<NuKeeperStorageSnapshot>(storage.get(), snapshot_meta_copy);
snapshot_task.snapshot = std::make_shared<KeeperStorageSnapshot>(storage.get(), snapshot_meta_copy);
}
snapshot_task.create_snapshot = [this, when_done] (NuKeeperStorageSnapshotPtr && snapshot)
snapshot_task.create_snapshot = [this, when_done] (KeeperStorageSnapshotPtr && snapshot)
{
nuraft::ptr<std::exception> exception(nullptr);
bool ret = true;
@ -203,7 +203,7 @@ void NuKeeperStateMachine::create_snapshot(
snapshots_queue.push(std::move(snapshot_task));
}
void NuKeeperStateMachine::save_logical_snp_obj(
void KeeperStateMachine::save_logical_snp_obj(
nuraft::snapshot & s,
size_t & obj_id,
nuraft::buffer & data,
@ -217,7 +217,7 @@ void NuKeeperStateMachine::save_logical_snp_obj(
if (obj_id == 0)
{
std::lock_guard lock(storage_lock);
NuKeeperStorageSnapshot snapshot(storage.get(), s.get_last_log_idx());
KeeperStorageSnapshot snapshot(storage.get(), s.get_last_log_idx());
cloned_buffer = snapshot_manager.serializeSnapshotToBuffer(snapshot);
}
else
@ -235,7 +235,7 @@ void NuKeeperStateMachine::save_logical_snp_obj(
std::shared_ptr<std::promise<void>> waiter = std::make_shared<std::promise<void>>();
auto future = waiter->get_future();
snapshot_task.snapshot = nullptr;
snapshot_task.create_snapshot = [this, waiter, cloned_buffer, log_idx = s.get_last_log_idx()] (NuKeeperStorageSnapshotPtr &&)
snapshot_task.create_snapshot = [this, waiter, cloned_buffer, log_idx = s.get_last_log_idx()] (KeeperStorageSnapshotPtr &&)
{
try
{
@ -261,7 +261,7 @@ void NuKeeperStateMachine::save_logical_snp_obj(
obj_id++;
}
int NuKeeperStateMachine::read_logical_snp_obj(
int KeeperStateMachine::read_logical_snp_obj(
nuraft::snapshot & s,
void* & /*user_snp_ctx*/,
ulong obj_id,
@ -289,9 +289,9 @@ int NuKeeperStateMachine::read_logical_snp_obj(
return 0;
}
void NuKeeperStateMachine::processReadRequest(const NuKeeperStorage::RequestForSession & request_for_session)
void KeeperStateMachine::processReadRequest(const KeeperStorage::RequestForSession & request_for_session)
{
NuKeeperStorage::ResponsesForSessions responses;
KeeperStorage::ResponsesForSessions responses;
{
std::lock_guard lock(storage_lock);
responses = storage->processRequest(request_for_session.request, request_for_session.session_id, std::nullopt);
@ -300,13 +300,13 @@ void NuKeeperStateMachine::processReadRequest(const NuKeeperStorage::RequestForS
responses_queue.push(response);
}
std::unordered_set<int64_t> NuKeeperStateMachine::getDeadSessions()
std::unordered_set<int64_t> KeeperStateMachine::getDeadSessions()
{
std::lock_guard lock(storage_lock);
return storage->getDeadSessions();
}
void NuKeeperStateMachine::shutdownStorage()
void KeeperStateMachine::shutdownStorage()
{
std::lock_guard lock(storage_lock);
storage->finalize();

18
src/Coordination/NuKeeperStateMachine.h → src/Coordination/KeeperStateMachine.h
View File

@ -1,22 +1,22 @@
#pragma once
#include <Coordination/NuKeeperStorage.h>
#include <Coordination/KeeperStorage.h>
#include <libnuraft/nuraft.hxx> // Y_IGNORE
#include <common/logger_useful.h>
#include <Coordination/ThreadSafeQueue.h>
#include <Coordination/CoordinationSettings.h>
#include <Coordination/NuKeeperSnapshotManager.h>
#include <Coordination/KeeperSnapshotManager.h>
namespace DB
{
using ResponsesQueue = ThreadSafeQueue<NuKeeperStorage::ResponseForSession>;
using ResponsesQueue = ThreadSafeQueue<KeeperStorage::ResponseForSession>;
using SnapshotsQueue = ConcurrentBoundedQueue<CreateSnapshotTask>;
class NuKeeperStateMachine : public nuraft::state_machine
class KeeperStateMachine : public nuraft::state_machine
{
public:
NuKeeperStateMachine(ResponsesQueue & responses_queue_, SnapshotsQueue & snapshots_queue_, const std::string & snapshots_path_, const CoordinationSettingsPtr & coordination_settings_);
KeeperStateMachine(ResponsesQueue & responses_queue_, SnapshotsQueue & snapshots_queue_, const std::string & snapshots_path_, const CoordinationSettingsPtr & coordination_settings_);
void init();
@ -50,12 +50,12 @@ public:
nuraft::ptr<nuraft::buffer> & data_out,
bool & is_last_obj) override;
NuKeeperStorage & getStorage()
KeeperStorage & getStorage()
{
return *storage;
}
void processReadRequest(const NuKeeperStorage::RequestForSession & request_for_session);
void processReadRequest(const KeeperStorage::RequestForSession & request_for_session);
std::unordered_set<int64_t> getDeadSessions();
@ -68,9 +68,9 @@ private:
CoordinationSettingsPtr coordination_settings;
NuKeeperStoragePtr storage;
KeeperStoragePtr storage;
NuKeeperSnapshotManager snapshot_manager;
KeeperSnapshotManager snapshot_manager;
ResponsesQueue & responses_queue;

18
src/Coordination/NuKeeperStateManager.cpp → src/Coordination/KeeperStateManager.cpp
View File

@ -1,4 +1,4 @@
#include <Coordination/NuKeeperStateManager.h>
#include <Coordination/KeeperStateManager.h>
#include <Common/Exception.h>
namespace DB
@ -9,23 +9,23 @@ namespace ErrorCodes
extern const int RAFT_ERROR;
}
NuKeeperStateManager::NuKeeperStateManager(int server_id_, const std::string & host, int port, const std::string & logs_path)
KeeperStateManager::KeeperStateManager(int server_id_, const std::string & host, int port, const std::string & logs_path)
: my_server_id(server_id_)
, my_port(port)
, log_store(nuraft::cs_new<NuKeeperLogStore>(logs_path, 5000, false))
, log_store(nuraft::cs_new<KeeperLogStore>(logs_path, 5000, false))
, cluster_config(nuraft::cs_new<nuraft::cluster_config>())
{
auto peer_config = nuraft::cs_new<nuraft::srv_config>(my_server_id, host + ":" + std::to_string(port));
cluster_config->get_servers().push_back(peer_config);
}
NuKeeperStateManager::NuKeeperStateManager(
KeeperStateManager::KeeperStateManager(
int my_server_id_,
const std::string & config_prefix,
const Poco::Util::AbstractConfiguration & config,
const CoordinationSettingsPtr & coordination_settings)
: my_server_id(my_server_id_)
, log_store(nuraft::cs_new<NuKeeperLogStore>(
, log_store(nuraft::cs_new<KeeperLogStore>(
config.getString(config_prefix + ".log_storage_path", config.getString("path", DBMS_DEFAULT_PATH) + "coordination/logs"),
coordination_settings->rotate_log_storage_interval, coordination_settings->force_sync))
, cluster_config(nuraft::cs_new<nuraft::cluster_config>())
@ -64,17 +64,17 @@ NuKeeperStateManager::NuKeeperStateManager(
throw Exception(ErrorCodes::RAFT_ERROR, "At least one of servers should be able to start as leader (without <start_as_follower>)");
}
void NuKeeperStateManager::loadLogStore(size_t last_commited_index, size_t logs_to_keep)
void KeeperStateManager::loadLogStore(size_t last_commited_index, size_t logs_to_keep)
{
log_store->init(last_commited_index, logs_to_keep);
}
void NuKeeperStateManager::flushLogStore()
void KeeperStateManager::flushLogStore()
{
log_store->flush();
}
void NuKeeperStateManager::save_config(const nuraft::cluster_config & config)
void KeeperStateManager::save_config(const nuraft::cluster_config & config)
{
// Just keep in memory in this example.
// Need to write to disk here, if want to make it durable.
@ -82,7 +82,7 @@ void NuKeeperStateManager::save_config(const nuraft::cluster_config & config)
cluster_config = nuraft::cluster_config::deserialize(*buf);
}
void NuKeeperStateManager::save_state(const nuraft::srv_state & state)
void KeeperStateManager::save_state(const nuraft::srv_state & state)
{
// Just keep in memory in this example.
// Need to write to disk here, if want to make it durable.

12
src/Coordination/NuKeeperStateManager.h → src/Coordination/KeeperStateManager.h
View File

@ -2,7 +2,7 @@
#include <Core/Types.h>
#include <string>
#include <Coordination/NuKeeperLogStore.h>
#include <Coordination/KeeperLogStore.h>
#include <Coordination/CoordinationSettings.h>
#include <libnuraft/nuraft.hxx> // Y_IGNORE
#include <Poco/Util/AbstractConfiguration.h>
@ -10,16 +10,16 @@
namespace DB
{
class NuKeeperStateManager : public nuraft::state_mgr
class KeeperStateManager : public nuraft::state_mgr
{
public:
NuKeeperStateManager(
KeeperStateManager(
int server_id_,
const std::string & config_prefix,
const Poco::Util::AbstractConfiguration & config,
const CoordinationSettingsPtr & coordination_settings);
NuKeeperStateManager(
KeeperStateManager(
int server_id_,
const std::string & host,
int port,
@ -52,7 +52,7 @@ public:
return start_as_follower_servers.count(my_server_id);
}
nuraft::ptr<NuKeeperLogStore> getLogStore() const { return log_store; }
nuraft::ptr<KeeperLogStore> getLogStore() const { return log_store; }
size_t getTotalServers() const { return total_servers; }
@ -61,7 +61,7 @@ private:
int my_port;
size_t total_servers{0};
std::unordered_set<int> start_as_follower_servers;
nuraft::ptr<NuKeeperLogStore> log_store;
nuraft::ptr<KeeperLogStore> log_store;
nuraft::ptr<nuraft::srv_config> my_server_config;
nuraft::ptr<nuraft::cluster_config> cluster_config;
nuraft::ptr<nuraft::srv_state> server_state;

184
src/Coordination/NuKeeperStorage.cpp → src/Coordination/KeeperStorage.cpp
View File

@ -1,4 +1,4 @@
#include <Coordination/NuKeeperStorage.h>
#include <Coordination/KeeperStorage.h>
#include <Common/ZooKeeper/IKeeper.h>
#include <Common/setThreadName.h>
#include <mutex>
@ -31,9 +31,9 @@ static std::string getBaseName(const String & path)
return std::string{&path[basename_start + 1], path.length() - basename_start - 1};
}
static NuKeeperStorage::ResponsesForSessions processWatchesImpl(const String & path, NuKeeperStorage::Watches & watches, NuKeeperStorage::Watches & list_watches, Coordination::Event event_type)
static KeeperStorage::ResponsesForSessions processWatchesImpl(const String & path, KeeperStorage::Watches & watches, KeeperStorage::Watches & list_watches, Coordination::Event event_type)
{
NuKeeperStorage::ResponsesForSessions result;
KeeperStorage::ResponsesForSessions result;
auto it = watches.find(path);
if (it != watches.end())
{
@ -44,7 +44,7 @@ static NuKeeperStorage::ResponsesForSessions processWatchesImpl(const String & p
watch_response->type = event_type;
watch_response->state = Coordination::State::CONNECTED;
for (auto watcher_session : it->second)
result.push_back(NuKeeperStorage::ResponseForSession{watcher_session, watch_response});
result.push_back(KeeperStorage::ResponseForSession{watcher_session, watch_response});
watches.erase(it);
}
@ -60,14 +60,14 @@ static NuKeeperStorage::ResponsesForSessions processWatchesImpl(const String & p
watch_list_response->type = Coordination::Event::CHILD;
watch_list_response->state = Coordination::State::CONNECTED;
for (auto watcher_session : it->second)
result.push_back(NuKeeperStorage::ResponseForSession{watcher_session, watch_list_response});
result.push_back(KeeperStorage::ResponseForSession{watcher_session, watch_list_response});
list_watches.erase(it);
}
return result;
}
NuKeeperStorage::NuKeeperStorage(int64_t tick_time_ms)
KeeperStorage::KeeperStorage(int64_t tick_time_ms)
: session_expiry_queue(tick_time_ms)
{
container.insert("/", Node());
@ -75,32 +75,32 @@ NuKeeperStorage::NuKeeperStorage(int64_t tick_time_ms)
using Undo = std::function<void()>;
struct NuKeeperStorageRequest
struct KeeperStorageRequest
{
Coordination::ZooKeeperRequestPtr zk_request;
explicit NuKeeperStorageRequest(const Coordination::ZooKeeperRequestPtr & zk_request_)
explicit KeeperStorageRequest(const Coordination::ZooKeeperRequestPtr & zk_request_)
: zk_request(zk_request_)
{}
virtual std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(NuKeeperStorage::Container & container, NuKeeperStorage::Ephemerals & ephemerals, int64_t zxid, int64_t session_id) const = 0;
virtual NuKeeperStorage::ResponsesForSessions processWatches(NuKeeperStorage::Watches & /*watches*/, NuKeeperStorage::Watches & /*list_watches*/) const { return {}; }
virtual std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(KeeperStorage::Container & container, KeeperStorage::Ephemerals & ephemerals, int64_t zxid, int64_t session_id) const = 0;
virtual KeeperStorage::ResponsesForSessions processWatches(KeeperStorage::Watches & /*watches*/, KeeperStorage::Watches & /*list_watches*/) const { return {}; }
virtual ~NuKeeperStorageRequest() = default;
virtual ~KeeperStorageRequest() = default;
};
struct NuKeeperStorageHeartbeatRequest final : public NuKeeperStorageRequest
struct KeeperStorageHeartbeatRequest final : public KeeperStorageRequest
{
using NuKeeperStorageRequest::NuKeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(NuKeeperStorage::Container & /* container */, NuKeeperStorage::Ephemerals & /* ephemerals */, int64_t /* zxid */, int64_t /* session_id */) const override
using KeeperStorageRequest::KeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(KeeperStorage::Container & /* container */, KeeperStorage::Ephemerals & /* ephemerals */, int64_t /* zxid */, int64_t /* session_id */) const override
{
return {zk_request->makeResponse(), {}};
}
};
struct NuKeeperStorageSyncRequest final : public NuKeeperStorageRequest
struct KeeperStorageSyncRequest final : public KeeperStorageRequest
{
using NuKeeperStorageRequest::NuKeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(NuKeeperStorage::Container & /* container */, NuKeeperStorage::Ephemerals & /* ephemerals */, int64_t /* zxid */, int64_t /* session_id */) const override
using KeeperStorageRequest::KeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(KeeperStorage::Container & /* container */, KeeperStorage::Ephemerals & /* ephemerals */, int64_t /* zxid */, int64_t /* session_id */) const override
{
auto response = zk_request->makeResponse();
dynamic_cast<Coordination::ZooKeeperSyncResponse *>(response.get())->path = dynamic_cast<Coordination::ZooKeeperSyncRequest *>(zk_request.get())->path;
@ -108,16 +108,16 @@ struct NuKeeperStorageSyncRequest final : public NuKeeperStorageRequest
}
};
struct NuKeeperStorageCreateRequest final : public NuKeeperStorageRequest
struct KeeperStorageCreateRequest final : public KeeperStorageRequest
{
using NuKeeperStorageRequest::NuKeeperStorageRequest;
using KeeperStorageRequest::KeeperStorageRequest;
NuKeeperStorage::ResponsesForSessions processWatches(NuKeeperStorage::Watches & watches, NuKeeperStorage::Watches & list_watches) const override
KeeperStorage::ResponsesForSessions processWatches(KeeperStorage::Watches & watches, KeeperStorage::Watches & list_watches) const override
{
return processWatchesImpl(zk_request->getPath(), watches, list_watches, Coordination::Event::CREATED);
}
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(NuKeeperStorage::Container & container, NuKeeperStorage::Ephemerals & ephemerals, int64_t zxid, int64_t session_id) const override
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(KeeperStorage::Container & container, KeeperStorage::Ephemerals & ephemerals, int64_t zxid, int64_t session_id) const override
{
Coordination::ZooKeeperResponsePtr response_ptr = zk_request->makeResponse();
Undo undo;
@ -143,7 +143,7 @@ struct NuKeeperStorageCreateRequest final : public NuKeeperStorageRequest
}
else
{
NuKeeperStorage::Node created_node;
KeeperStorage::Node created_node;
created_node.stat.czxid = zxid;
created_node.stat.mzxid = zxid;
created_node.stat.ctime = std::chrono::system_clock::now().time_since_epoch() / std::chrono::milliseconds(1);
@ -167,7 +167,7 @@ struct NuKeeperStorageCreateRequest final : public NuKeeperStorageRequest
}
auto child_path = getBaseName(path_created);
container.updateValue(parent_path, [child_path] (NuKeeperStorage::Node & parent)
container.updateValue(parent_path, [child_path] (KeeperStorage::Node & parent)
{
/// Increment sequential number even if node is not sequential
++parent.seq_num;
@ -188,7 +188,7 @@ struct NuKeeperStorageCreateRequest final : public NuKeeperStorageRequest
if (is_ephemeral)
ephemerals[session_id].erase(path_created);
container.updateValue(parent_path, [child_path] (NuKeeperStorage::Node & undo_parent)
container.updateValue(parent_path, [child_path] (KeeperStorage::Node & undo_parent)
{
--undo_parent.stat.cversion;
--undo_parent.stat.numChildren;
@ -205,10 +205,10 @@ struct NuKeeperStorageCreateRequest final : public NuKeeperStorageRequest
}
};
struct NuKeeperStorageGetRequest final : public NuKeeperStorageRequest
struct KeeperStorageGetRequest final : public KeeperStorageRequest
{
using NuKeeperStorageRequest::NuKeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(NuKeeperStorage::Container & container, NuKeeperStorage::Ephemerals & /* ephemerals */, int64_t /* zxid */, int64_t /* session_id */) const override
using KeeperStorageRequest::KeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(KeeperStorage::Container & container, KeeperStorage::Ephemerals & /* ephemerals */, int64_t /* zxid */, int64_t /* session_id */) const override
{
Coordination::ZooKeeperResponsePtr response_ptr = zk_request->makeResponse();
Coordination::ZooKeeperGetResponse & response = dynamic_cast<Coordination::ZooKeeperGetResponse &>(*response_ptr);
@ -230,10 +230,10 @@ struct NuKeeperStorageGetRequest final : public NuKeeperStorageRequest
}
};
struct NuKeeperStorageRemoveRequest final : public NuKeeperStorageRequest
struct KeeperStorageRemoveRequest final : public KeeperStorageRequest
{
using NuKeeperStorageRequest::NuKeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(NuKeeperStorage::Container & container, NuKeeperStorage::Ephemerals & ephemerals, int64_t /*zxid*/, int64_t /*session_id*/) const override
using KeeperStorageRequest::KeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(KeeperStorage::Container & container, KeeperStorage::Ephemerals & ephemerals, int64_t /*zxid*/, int64_t /*session_id*/) const override
{
Coordination::ZooKeeperResponsePtr response_ptr = zk_request->makeResponse();
Coordination::ZooKeeperRemoveResponse & response = dynamic_cast<Coordination::ZooKeeperRemoveResponse &>(*response_ptr);
@ -265,7 +265,7 @@ struct NuKeeperStorageRemoveRequest final : public NuKeeperStorageRequest
}
auto child_basename = getBaseName(it->key);
container.updateValue(parentPath(request.path), [&child_basename] (NuKeeperStorage::Node & parent)
container.updateValue(parentPath(request.path), [&child_basename] (KeeperStorage::Node & parent)
{
--parent.stat.numChildren;
++parent.stat.cversion;
@ -282,7 +282,7 @@ struct NuKeeperStorageRemoveRequest final : public NuKeeperStorageRequest
ephemerals[prev_node.stat.ephemeralOwner].emplace(path);
container.insert(path, prev_node);
container.updateValue(parentPath(path), [&child_basename] (NuKeeperStorage::Node & parent)
container.updateValue(parentPath(path), [&child_basename] (KeeperStorage::Node & parent)
{
++parent.stat.numChildren;
--parent.stat.cversion;
@ -294,16 +294,16 @@ struct NuKeeperStorageRemoveRequest final : public NuKeeperStorageRequest
return { response_ptr, undo };
}
NuKeeperStorage::ResponsesForSessions processWatches(NuKeeperStorage::Watches & watches, NuKeeperStorage::Watches & list_watches) const override
KeeperStorage::ResponsesForSessions processWatches(KeeperStorage::Watches & watches, KeeperStorage::Watches & list_watches) const override
{
return processWatchesImpl(zk_request->getPath(), watches, list_watches, Coordination::Event::DELETED);
}
};
struct NuKeeperStorageExistsRequest final : public NuKeeperStorageRequest
struct KeeperStorageExistsRequest final : public KeeperStorageRequest
{
using NuKeeperStorageRequest::NuKeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(NuKeeperStorage::Container & container, NuKeeperStorage::Ephemerals & /* ephemerals */, int64_t /*zxid*/, int64_t /* session_id */) const override
using KeeperStorageRequest::KeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(KeeperStorage::Container & container, KeeperStorage::Ephemerals & /* ephemerals */, int64_t /*zxid*/, int64_t /* session_id */) const override
{
Coordination::ZooKeeperResponsePtr response_ptr = zk_request->makeResponse();
Coordination::ZooKeeperExistsResponse & response = dynamic_cast<Coordination::ZooKeeperExistsResponse &>(*response_ptr);
@ -324,10 +324,10 @@ struct NuKeeperStorageExistsRequest final : public NuKeeperStorageRequest
}
};
struct NuKeeperStorageSetRequest final : public NuKeeperStorageRequest
struct KeeperStorageSetRequest final : public KeeperStorageRequest
{
using NuKeeperStorageRequest::NuKeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(NuKeeperStorage::Container & container, NuKeeperStorage::Ephemerals & /* ephemerals */, int64_t zxid, int64_t /* session_id */) const override
using KeeperStorageRequest::KeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(KeeperStorage::Container & container, KeeperStorage::Ephemerals & /* ephemerals */, int64_t zxid, int64_t /* session_id */) const override
{
Coordination::ZooKeeperResponsePtr response_ptr = zk_request->makeResponse();
Coordination::ZooKeeperSetResponse & response = dynamic_cast<Coordination::ZooKeeperSetResponse &>(*response_ptr);
@ -343,7 +343,7 @@ struct NuKeeperStorageSetRequest final : public NuKeeperStorageRequest
{
auto prev_node = it->value;
auto itr = container.updateValue(request.path, [zxid, request] (NuKeeperStorage::Node & value)
auto itr = container.updateValue(request.path, [zxid, request] (KeeperStorage::Node & value)
{
value.data = request.data;
value.stat.version++;
@ -353,7 +353,7 @@ struct NuKeeperStorageSetRequest final : public NuKeeperStorageRequest
value.data = request.data;
});
container.updateValue(parentPath(request.path), [] (NuKeeperStorage::Node & parent)
container.updateValue(parentPath(request.path), [] (KeeperStorage::Node & parent)
{
parent.stat.cversion++;
});
@ -363,8 +363,8 @@ struct NuKeeperStorageSetRequest final : public NuKeeperStorageRequest
undo = [prev_node, &container, path = request.path]
{
container.updateValue(path, [&prev_node] (NuKeeperStorage::Node & value) { value = prev_node; });
container.updateValue(parentPath(path), [] (NuKeeperStorage::Node & parent)
container.updateValue(path, [&prev_node] (KeeperStorage::Node & value) { value = prev_node; });
container.updateValue(parentPath(path), [] (KeeperStorage::Node & parent)
{
parent.stat.cversion--;
});
@ -378,16 +378,16 @@ struct NuKeeperStorageSetRequest final : public NuKeeperStorageRequest
return { response_ptr, undo };
}
NuKeeperStorage::ResponsesForSessions processWatches(NuKeeperStorage::Watches & watches, NuKeeperStorage::Watches & list_watches) const override
KeeperStorage::ResponsesForSessions processWatches(KeeperStorage::Watches & watches, KeeperStorage::Watches & list_watches) const override
{
return processWatchesImpl(zk_request->getPath(), watches, list_watches, Coordination::Event::CHANGED);
}
};
struct NuKeeperStorageListRequest final : public NuKeeperStorageRequest
struct KeeperStorageListRequest final : public KeeperStorageRequest
{
using NuKeeperStorageRequest::NuKeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(NuKeeperStorage::Container & container, NuKeeperStorage::Ephemerals & /* ephemerals */, int64_t /*zxid*/, int64_t /*session_id*/) const override
using KeeperStorageRequest::KeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(KeeperStorage::Container & container, KeeperStorage::Ephemerals & /* ephemerals */, int64_t /*zxid*/, int64_t /*session_id*/) const override
{
Coordination::ZooKeeperResponsePtr response_ptr = zk_request->makeResponse();
Coordination::ZooKeeperListResponse & response = dynamic_cast<Coordination::ZooKeeperListResponse &>(*response_ptr);
@ -415,10 +415,10 @@ struct NuKeeperStorageListRequest final : public NuKeeperStorageRequest
}
};
struct NuKeeperStorageCheckRequest final : public NuKeeperStorageRequest
struct KeeperStorageCheckRequest final : public KeeperStorageRequest
{
using NuKeeperStorageRequest::NuKeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(NuKeeperStorage::Container & container, NuKeeperStorage::Ephemerals & /* ephemerals */, int64_t /*zxid*/, int64_t /*session_id*/) const override
using KeeperStorageRequest::KeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(KeeperStorage::Container & container, KeeperStorage::Ephemerals & /* ephemerals */, int64_t /*zxid*/, int64_t /*session_id*/) const override
{
Coordination::ZooKeeperResponsePtr response_ptr = zk_request->makeResponse();
Coordination::ZooKeeperCheckResponse & response = dynamic_cast<Coordination::ZooKeeperCheckResponse &>(*response_ptr);
@ -441,11 +441,11 @@ struct NuKeeperStorageCheckRequest final : public NuKeeperStorageRequest
}
};
struct NuKeeperStorageMultiRequest final : public NuKeeperStorageRequest
struct KeeperStorageMultiRequest final : public KeeperStorageRequest
{
std::vector<NuKeeperStorageRequestPtr> concrete_requests;
explicit NuKeeperStorageMultiRequest(const Coordination::ZooKeeperRequestPtr & zk_request_)
: NuKeeperStorageRequest(zk_request_)
std::vector<KeeperStorageRequestPtr> concrete_requests;
explicit KeeperStorageMultiRequest(const Coordination::ZooKeeperRequestPtr & zk_request_)
: KeeperStorageRequest(zk_request_)
{
Coordination::ZooKeeperMultiRequest & request = dynamic_cast<Coordination::ZooKeeperMultiRequest &>(*zk_request);
concrete_requests.reserve(request.requests.size());
@ -455,26 +455,26 @@ struct NuKeeperStorageMultiRequest final : public NuKeeperStorageRequest
auto sub_zk_request = std::dynamic_pointer_cast<Coordination::ZooKeeperRequest>(sub_request);
if (sub_zk_request->getOpNum() == Coordination::OpNum::Create)
{
concrete_requests.push_back(std::make_shared<NuKeeperStorageCreateRequest>(sub_zk_request));
concrete_requests.push_back(std::make_shared<KeeperStorageCreateRequest>(sub_zk_request));
}
else if (sub_zk_request->getOpNum() == Coordination::OpNum::Remove)
{
concrete_requests.push_back(std::make_shared<NuKeeperStorageRemoveRequest>(sub_zk_request));
concrete_requests.push_back(std::make_shared<KeeperStorageRemoveRequest>(sub_zk_request));
}
else if (sub_zk_request->getOpNum() == Coordination::OpNum::Set)
{
concrete_requests.push_back(std::make_shared<NuKeeperStorageSetRequest>(sub_zk_request));
concrete_requests.push_back(std::make_shared<KeeperStorageSetRequest>(sub_zk_request));
}
else if (sub_zk_request->getOpNum() == Coordination::OpNum::Check)
{
concrete_requests.push_back(std::make_shared<NuKeeperStorageCheckRequest>(sub_zk_request));
concrete_requests.push_back(std::make_shared<KeeperStorageCheckRequest>(sub_zk_request));
}
else
throw DB::Exception(ErrorCodes::BAD_ARGUMENTS, "Illegal command as part of multi ZooKeeper request {}", sub_zk_request->getOpNum());
}
}
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(NuKeeperStorage::Container & container, NuKeeperStorage::Ephemerals & ephemerals, int64_t zxid, int64_t session_id) const override
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(KeeperStorage::Container & container, KeeperStorage::Ephemerals & ephemerals, int64_t zxid, int64_t session_id) const override
{
Coordination::ZooKeeperResponsePtr response_ptr = zk_request->makeResponse();
Coordination::ZooKeeperMultiResponse & response = dynamic_cast<Coordination::ZooKeeperMultiResponse &>(*response_ptr);
@ -527,9 +527,9 @@ struct NuKeeperStorageMultiRequest final : public NuKeeperStorageRequest
}
}
NuKeeperStorage::ResponsesForSessions processWatches(NuKeeperStorage::Watches & watches, NuKeeperStorage::Watches & list_watches) const override
KeeperStorage::ResponsesForSessions processWatches(KeeperStorage::Watches & watches, KeeperStorage::Watches & list_watches) const override
{
NuKeeperStorage::ResponsesForSessions result;
KeeperStorage::ResponsesForSessions result;
for (const auto & generic_request : concrete_requests)
{
auto responses = generic_request->processWatches(watches, list_watches);
@ -539,16 +539,16 @@ struct NuKeeperStorageMultiRequest final : public NuKeeperStorageRequest
}
};
struct NuKeeperStorageCloseRequest final : public NuKeeperStorageRequest
struct KeeperStorageCloseRequest final : public KeeperStorageRequest
{
using NuKeeperStorageRequest::NuKeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(NuKeeperStorage::Container &, NuKeeperStorage::Ephemerals &, int64_t, int64_t) const override
using KeeperStorageRequest::KeeperStorageRequest;
std::pair<Coordination::ZooKeeperResponsePtr, Undo> process(KeeperStorage::Container &, KeeperStorage::Ephemerals &, int64_t, int64_t) const override
{
throw DB::Exception("Called process on close request", ErrorCodes::LOGICAL_ERROR);
}
};
void NuKeeperStorage::finalize()
void KeeperStorage::finalize()
{
if (finalized)
throw DB::Exception("Testkeeper storage already finalized", ErrorCodes::LOGICAL_ERROR);
@ -568,20 +568,20 @@ void NuKeeperStorage::finalize()
}
class NuKeeperWrapperFactory final : private boost::noncopyable
class KeeperWrapperFactory final : private boost::noncopyable
{
public:
using Creator = std::function<NuKeeperStorageRequestPtr(const Coordination::ZooKeeperRequestPtr &)>;
using Creator = std::function<KeeperStorageRequestPtr(const Coordination::ZooKeeperRequestPtr &)>;
using OpNumToRequest = std::unordered_map<Coordination::OpNum, Creator>;
static NuKeeperWrapperFactory & instance()
static KeeperWrapperFactory & instance()
{
static NuKeeperWrapperFactory factory;
static KeeperWrapperFactory factory;
return factory;
}
NuKeeperStorageRequestPtr get(const Coordination::ZooKeeperRequestPtr & zk_request) const
KeeperStorageRequestPtr get(const Coordination::ZooKeeperRequestPtr & zk_request) const
{
auto it = op_num_to_request.find(zk_request->getOpNum());
if (it == op_num_to_request.end())
@ -598,37 +598,37 @@ public:
private:
OpNumToRequest op_num_to_request;
NuKeeperWrapperFactory();
KeeperWrapperFactory();
};
template<Coordination::OpNum num, typename RequestT>
void registerNuKeeperRequestWrapper(NuKeeperWrapperFactory & factory)
void registerKeeperRequestWrapper(KeeperWrapperFactory & factory)
{
factory.registerRequest(num, [] (const Coordination::ZooKeeperRequestPtr & zk_request) { return std::make_shared<RequestT>(zk_request); });
}
NuKeeperWrapperFactory::NuKeeperWrapperFactory()
KeeperWrapperFactory::KeeperWrapperFactory()
{
registerNuKeeperRequestWrapper<Coordination::OpNum::Heartbeat, NuKeeperStorageHeartbeatRequest>(*this);
registerNuKeeperRequestWrapper<Coordination::OpNum::Sync, NuKeeperStorageSyncRequest>(*this);
//registerNuKeeperRequestWrapper<Coordination::OpNum::Auth, NuKeeperStorageAuthRequest>(*this);
registerNuKeeperRequestWrapper<Coordination::OpNum::Close, NuKeeperStorageCloseRequest>(*this);
registerNuKeeperRequestWrapper<Coordination::OpNum::Create, NuKeeperStorageCreateRequest>(*this);
registerNuKeeperRequestWrapper<Coordination::OpNum::Remove, NuKeeperStorageRemoveRequest>(*this);
registerNuKeeperRequestWrapper<Coordination::OpNum::Exists, NuKeeperStorageExistsRequest>(*this);
registerNuKeeperRequestWrapper<Coordination::OpNum::Get, NuKeeperStorageGetRequest>(*this);
registerNuKeeperRequestWrapper<Coordination::OpNum::Set, NuKeeperStorageSetRequest>(*this);
registerNuKeeperRequestWrapper<Coordination::OpNum::List, NuKeeperStorageListRequest>(*this);
registerNuKeeperRequestWrapper<Coordination::OpNum::SimpleList, NuKeeperStorageListRequest>(*this);
registerNuKeeperRequestWrapper<Coordination::OpNum::Check, NuKeeperStorageCheckRequest>(*this);
registerNuKeeperRequestWrapper<Coordination::OpNum::Multi, NuKeeperStorageMultiRequest>(*this);
registerKeeperRequestWrapper<Coordination::OpNum::Heartbeat, KeeperStorageHeartbeatRequest>(*this);
registerKeeperRequestWrapper<Coordination::OpNum::Sync, KeeperStorageSyncRequest>(*this);
//registerKeeperRequestWrapper<Coordination::OpNum::Auth, KeeperStorageAuthRequest>(*this);
registerKeeperRequestWrapper<Coordination::OpNum::Close, KeeperStorageCloseRequest>(*this);
registerKeeperRequestWrapper<Coordination::OpNum::Create, KeeperStorageCreateRequest>(*this);
registerKeeperRequestWrapper<Coordination::OpNum::Remove, KeeperStorageRemoveRequest>(*this);
registerKeeperRequestWrapper<Coordination::OpNum::Exists, KeeperStorageExistsRequest>(*this);
registerKeeperRequestWrapper<Coordination::OpNum::Get, KeeperStorageGetRequest>(*this);
registerKeeperRequestWrapper<Coordination::OpNum::Set, KeeperStorageSetRequest>(*this);
registerKeeperRequestWrapper<Coordination::OpNum::List, KeeperStorageListRequest>(*this);
registerKeeperRequestWrapper<Coordination::OpNum::SimpleList, KeeperStorageListRequest>(*this);
registerKeeperRequestWrapper<Coordination::OpNum::Check, KeeperStorageCheckRequest>(*this);
registerKeeperRequestWrapper<Coordination::OpNum::Multi, KeeperStorageMultiRequest>(*this);
}
NuKeeperStorage::ResponsesForSessions NuKeeperStorage::processRequest(const Coordination::ZooKeeperRequestPtr & zk_request, int64_t session_id, std::optional<int64_t> new_last_zxid)
KeeperStorage::ResponsesForSessions KeeperStorage::processRequest(const Coordination::ZooKeeperRequestPtr & zk_request, int64_t session_id, std::optional<int64_t> new_last_zxid)
{
NuKeeperStorage::ResponsesForSessions results;
KeeperStorage::ResponsesForSessions results;
if (new_last_zxid)
{
if (zxid >= *new_last_zxid)
@ -645,7 +645,7 @@ NuKeeperStorage::ResponsesForSessions NuKeeperStorage::processRequest(const Coor
for (const auto & ephemeral_path : it->second)
{
container.erase(ephemeral_path);
container.updateValue(parentPath(ephemeral_path), [&ephemeral_path] (NuKeeperStorage::Node & parent)
container.updateValue(parentPath(ephemeral_path), [&ephemeral_path] (KeeperStorage::Node & parent)
{
--parent.stat.numChildren;
++parent.stat.cversion;
@ -669,7 +669,7 @@ NuKeeperStorage::ResponsesForSessions NuKeeperStorage::processRequest(const Coor
}
else if (zk_request->getOpNum() == Coordination::OpNum::Heartbeat)
{
NuKeeperStorageRequestPtr storage_request = NuKeeperWrapperFactory::instance().get(zk_request);
KeeperStorageRequestPtr storage_request = KeeperWrapperFactory::instance().get(zk_request);
auto [response, _] = storage_request->process(container, ephemerals, zxid, session_id);
response->xid = zk_request->xid;
response->zxid = getZXID();
@ -678,7 +678,7 @@ NuKeeperStorage::ResponsesForSessions NuKeeperStorage::processRequest(const Coor
}
else
{
NuKeeperStorageRequestPtr storage_request = NuKeeperWrapperFactory::instance().get(zk_request);
KeeperStorageRequestPtr storage_request = KeeperWrapperFactory::instance().get(zk_request);
auto [response, _] = storage_request->process(container, ephemerals, zxid, session_id);
if (zk_request->has_watch)
@ -715,7 +715,7 @@ NuKeeperStorage::ResponsesForSessions NuKeeperStorage::processRequest(const Coor
}
void NuKeeperStorage::clearDeadWatches(int64_t session_id)
void KeeperStorage::clearDeadWatches(int64_t session_id)
{
auto watches_it = sessions_and_watchers.find(session_id);
if (watches_it != sessions_and_watchers.end())

12
src/Coordination/NuKeeperStorage.h → src/Coordination/KeeperStorage.h
View File

@ -14,15 +14,15 @@ namespace DB
{
using namespace DB;
struct NuKeeperStorageRequest;
using NuKeeperStorageRequestPtr = std::shared_ptr<NuKeeperStorageRequest>;
struct KeeperStorageRequest;
using KeeperStorageRequestPtr = std::shared_ptr<KeeperStorageRequest>;
using ResponseCallback = std::function<void(const Coordination::ZooKeeperResponsePtr &)>;
using ChildrenSet = std::unordered_set<std::string>;
using SessionAndTimeout = std::unordered_map<int64_t, int64_t>;
struct NuKeeperStorageSnapshot;
struct KeeperStorageSnapshot;
class NuKeeperStorage
class KeeperStorage
{
public:
int64_t session_id_counter{1};
@ -80,7 +80,7 @@ public:
}
public:
NuKeeperStorage(int64_t tick_time_ms);
KeeperStorage(int64_t tick_time_ms);
int64_t getSessionID(int64_t session_timeout_ms)
{
@ -131,6 +131,6 @@ public:
}
};
using NuKeeperStoragePtr = std::unique_ptr<NuKeeperStorage>;
using KeeperStoragePtr = std::unique_ptr<KeeperStorage>;
}

50
src/Coordination/NuKeeperStorageDispatcher.cpp → src/Coordination/KeeperStorageDispatcher.cpp
View File

@ -1,4 +1,4 @@
#include <Coordination/NuKeeperStorageDispatcher.h>
#include <Coordination/KeeperStorageDispatcher.h>
#include <Common/setThreadName.h>
namespace DB
@ -11,18 +11,18 @@ namespace ErrorCodes
extern const int TIMEOUT_EXCEEDED;
}
NuKeeperStorageDispatcher::NuKeeperStorageDispatcher()
KeeperStorageDispatcher::KeeperStorageDispatcher()
: coordination_settings(std::make_shared<CoordinationSettings>())
, log(&Poco::Logger::get("NuKeeperDispatcher"))
, log(&Poco::Logger::get("KeeperDispatcher"))
{
}
void NuKeeperStorageDispatcher::requestThread()
void KeeperStorageDispatcher::requestThread()
{
setThreadName("NuKeeperReqT");
setThreadName("KeeperReqT");
while (!shutdown_called)
{
NuKeeperStorage::RequestForSession request;
KeeperStorage::RequestForSession request;
UInt64 max_wait = UInt64(coordination_settings->operation_timeout_ms.totalMilliseconds());
@ -43,12 +43,12 @@ void NuKeeperStorageDispatcher::requestThread()
}
}
void NuKeeperStorageDispatcher::responseThread()
void KeeperStorageDispatcher::responseThread()
{
setThreadName("NuKeeperRspT");
setThreadName("KeeperRspT");
while (!shutdown_called)
{
NuKeeperStorage::ResponseForSession response_for_session;
KeeperStorage::ResponseForSession response_for_session;
UInt64 max_wait = UInt64(coordination_settings->operation_timeout_ms.totalMilliseconds());
@ -69,9 +69,9 @@ void NuKeeperStorageDispatcher::responseThread()
}
}
void NuKeeperStorageDispatcher::snapshotThread()
void KeeperStorageDispatcher::snapshotThread()
{
setThreadName("NuKeeperSnpT");
setThreadName("KeeperSnpT");
while (!shutdown_called)
{
CreateSnapshotTask task;
@ -91,7 +91,7 @@ void NuKeeperStorageDispatcher::snapshotThread()
}
}
void NuKeeperStorageDispatcher::setResponse(int64_t session_id, const Coordination::ZooKeeperResponsePtr & response)
void KeeperStorageDispatcher::setResponse(int64_t session_id, const Coordination::ZooKeeperResponsePtr & response)
{
std::lock_guard lock(session_to_response_callback_mutex);
auto session_writer = session_to_response_callback.find(session_id);
@ -104,7 +104,7 @@ void NuKeeperStorageDispatcher::setResponse(int64_t session_id, const Coordinati
session_to_response_callback.erase(session_writer);
}
bool NuKeeperStorageDispatcher::putRequest(const Coordination::ZooKeeperRequestPtr & request, int64_t session_id)
bool KeeperStorageDispatcher::putRequest(const Coordination::ZooKeeperRequestPtr & request, int64_t session_id)
{
{
std::lock_guard lock(session_to_response_callback_mutex);
@ -112,7 +112,7 @@ bool NuKeeperStorageDispatcher::putRequest(const Coordination::ZooKeeperRequestP
return false;
}
NuKeeperStorage::RequestForSession request_info;
KeeperStorage::RequestForSession request_info;
request_info.request = request;
request_info.session_id = session_id;
@ -125,18 +125,18 @@ bool NuKeeperStorageDispatcher::putRequest(const Coordination::ZooKeeperRequestP
return true;
}
void NuKeeperStorageDispatcher::initialize(const Poco::Util::AbstractConfiguration & config)
void KeeperStorageDispatcher::initialize(const Poco::Util::AbstractConfiguration & config)
{
LOG_DEBUG(log, "Initializing storage dispatcher");
int myid = config.getInt("test_keeper_server.server_id");
int myid = config.getInt("keeper_server.server_id");
coordination_settings->loadFromConfig("test_keeper_server.coordination_settings", config);
coordination_settings->loadFromConfig("keeper_server.coordination_settings", config);
request_thread = ThreadFromGlobalPool([this] { requestThread(); });
responses_thread = ThreadFromGlobalPool([this] { responseThread(); });
snapshot_thread = ThreadFromGlobalPool([this] { snapshotThread(); });
server = std::make_unique<NuKeeperServer>(myid, coordination_settings, config, responses_queue, snapshots_queue);
server = std::make_unique<KeeperServer>(myid, coordination_settings, config, responses_queue, snapshots_queue);
try
{
LOG_DEBUG(log, "Waiting server to initialize");
@ -158,7 +158,7 @@ void NuKeeperStorageDispatcher::initialize(const Poco::Util::AbstractConfigurati
LOG_DEBUG(log, "Dispatcher initialized");
}
void NuKeeperStorageDispatcher::shutdown()
void KeeperStorageDispatcher::shutdown()
{
try
{
@ -191,7 +191,7 @@ void NuKeeperStorageDispatcher::shutdown()
if (server)
server->shutdown();
NuKeeperStorage::RequestForSession request_for_session;
KeeperStorage::RequestForSession request_for_session;
while (requests_queue.tryPop(request_for_session))
{
if (request_for_session.request)
@ -215,19 +215,19 @@ void NuKeeperStorageDispatcher::shutdown()
LOG_DEBUG(log, "Dispatcher shut down");
}
NuKeeperStorageDispatcher::~NuKeeperStorageDispatcher()
KeeperStorageDispatcher::~KeeperStorageDispatcher()
{
shutdown();
}
void NuKeeperStorageDispatcher::registerSession(int64_t session_id, ZooKeeperResponseCallback callback)
void KeeperStorageDispatcher::registerSession(int64_t session_id, ZooKeeperResponseCallback callback)
{
std::lock_guard lock(session_to_response_callback_mutex);
if (!session_to_response_callback.try_emplace(session_id, callback).second)
throw Exception(DB::ErrorCodes::LOGICAL_ERROR, "Session with id {} already registered in dispatcher", session_id);
}
void NuKeeperStorageDispatcher::sessionCleanerTask()
void KeeperStorageDispatcher::sessionCleanerTask()
{
while (true)
{
@ -244,7 +244,7 @@ void NuKeeperStorageDispatcher::sessionCleanerTask()
LOG_INFO(log, "Found dead session {}, will try to close it", dead_session);
Coordination::ZooKeeperRequestPtr request = Coordination::ZooKeeperRequestFactory::instance().get(Coordination::OpNum::Close);
request->xid = Coordination::CLOSE_XID;
NuKeeperStorage::RequestForSession request_info;
KeeperStorage::RequestForSession request_info;
request_info.request = request;
request_info.session_id = dead_session;
{
@ -265,7 +265,7 @@ void NuKeeperStorageDispatcher::sessionCleanerTask()
}
}
void NuKeeperStorageDispatcher::finishSession(int64_t session_id)
void KeeperStorageDispatcher::finishSession(int64_t session_id)
{
std::lock_guard lock(session_to_response_callback_mutex);
auto session_it = session_to_response_callback.find(session_id);

12
src/Coordination/NuKeeperStorageDispatcher.h → src/Coordination/KeeperStorageDispatcher.h
View File

@ -13,7 +13,7 @@
#include <Common/Exception.h>
#include <common/logger_useful.h>
#include <functional>
#include <Coordination/NuKeeperServer.h>
#include <Coordination/KeeperServer.h>
#include <Coordination/CoordinationSettings.h>
@ -22,14 +22,14 @@ namespace DB
using ZooKeeperResponseCallback = std::function<void(const Coordination::ZooKeeperResponsePtr & response)>;
class NuKeeperStorageDispatcher
class KeeperStorageDispatcher
{
private:
std::mutex push_request_mutex;
CoordinationSettingsPtr coordination_settings;
using RequestsQueue = ConcurrentBoundedQueue<NuKeeperStorage::RequestForSession>;
using RequestsQueue = ConcurrentBoundedQueue<KeeperStorage::RequestForSession>;
using SessionToResponseCallback = std::unordered_map<int64_t, ZooKeeperResponseCallback>;
RequestsQueue requests_queue{1};
@ -46,7 +46,7 @@ private:
ThreadFromGlobalPool session_cleaner_thread;
ThreadFromGlobalPool snapshot_thread;
std::unique_ptr<NuKeeperServer> server;
std::unique_ptr<KeeperServer> server;
Poco::Logger * log;
@ -58,13 +58,13 @@ private:
void setResponse(int64_t session_id, const Coordination::ZooKeeperResponsePtr & response);
public:
NuKeeperStorageDispatcher();
KeeperStorageDispatcher();
void initialize(const Poco::Util::AbstractConfiguration & config);
void shutdown();
~NuKeeperStorageDispatcher();
~KeeperStorageDispatcher();
bool putRequest(const Coordination::ZooKeeperRequestPtr & request, int64_t session_id);

24
src/Coordination/NuKeeperCommon.h
View File

@ -1,24 +0,0 @@
#pragma once
#include <Common/ZooKeeper/ZooKeeperCommon.h>
namespace DB
{
struct NuKeeperRequest
{
int64_t session_id;
Coordination::ZooKeeperRequestPtr request;
};
using NuKeeperRequests = std::vector<NuKeeperRequest>;
struct NuKeeperResponse
{
int64_t session_id;
Coordination::ZooKeeperRequestPtr response;
};
using NuKeeperResponses = std::vector<NuKeeperResponse>;
}

116
src/Coordination/tests/gtest_for_build.cpp
View File

@ -9,10 +9,10 @@
#include <Poco/ConsoleChannel.h>
#include <Poco/Logger.h>
#include <Coordination/InMemoryLogStore.h>
#include <Coordination/NuKeeperStateManager.h>
#include <Coordination/NuKeeperSnapshotManager.h>
#include <Coordination/KeeperStateManager.h>
#include <Coordination/KeeperSnapshotManager.h>
#include <Coordination/SummingStateMachine.h>
#include <Coordination/NuKeeperStateMachine.h>
#include <Coordination/KeeperStateMachine.h>
#include <Coordination/LoggerWrapper.h>
#include <Coordination/WriteBufferFromNuraftBuffer.h>
#include <Coordination/ReadBufferFromNuraftBuffer.h>
@ -24,7 +24,7 @@
#include <common/logger_useful.h>
#include <libnuraft/nuraft.hxx> // Y_IGNORE
#include <thread>
#include <Coordination/NuKeeperLogStore.h>
#include <Coordination/KeeperLogStore.h>
#include <Coordination/Changelog.h>
#include <filesystem>
@ -102,7 +102,7 @@ struct SimpliestRaftServer
, port(port_)
, endpoint(hostname + ":" + std::to_string(port))
, state_machine(nuraft::cs_new<StateMachine>())
, state_manager(nuraft::cs_new<DB::NuKeeperStateManager>(server_id, hostname, port, logs_path))
, state_manager(nuraft::cs_new<DB::KeeperStateManager>(server_id, hostname, port, logs_path))
{
state_manager->loadLogStore(1, 0);
nuraft::raft_params params;
@ -153,7 +153,7 @@ struct SimpliestRaftServer
nuraft::ptr<StateMachine> state_machine;
// State manager.
nuraft::ptr<DB::NuKeeperStateManager> state_manager;
nuraft::ptr<DB::KeeperStateManager> state_manager;
// Raft launcher.
nuraft::raft_launcher launcher;
@ -207,7 +207,7 @@ DB::LogEntryPtr getLogEntry(const std::string & s, size_t term)
TEST(CoordinationTest, ChangelogTestSimple)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 5, true);
DB::KeeperLogStore changelog("./logs", 5, true);
changelog.init(1, 0);
auto entry = getLogEntry("hello world", 77);
changelog.append(entry);
@ -221,7 +221,7 @@ TEST(CoordinationTest, ChangelogTestSimple)
TEST(CoordinationTest, ChangelogTestFile)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 5, true);
DB::KeeperLogStore changelog("./logs", 5, true);
changelog.init(1, 0);
auto entry = getLogEntry("hello world", 77);
changelog.append(entry);
@ -242,7 +242,7 @@ TEST(CoordinationTest, ChangelogTestFile)
TEST(CoordinationTest, ChangelogReadWrite)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 1000, true);
DB::KeeperLogStore changelog("./logs", 1000, true);
changelog.init(1, 0);
for (size_t i = 0; i < 10; ++i)
{
@ -250,7 +250,7 @@ TEST(CoordinationTest, ChangelogReadWrite)
changelog.append(entry);
}
EXPECT_EQ(changelog.size(), 10);
DB::NuKeeperLogStore changelog_reader("./logs", 1000, true);
DB::KeeperLogStore changelog_reader("./logs", 1000, true);
changelog_reader.init(1, 0);
EXPECT_EQ(changelog_reader.size(), 10);
EXPECT_EQ(changelog_reader.last_entry()->get_term(), changelog.last_entry()->get_term());
@ -269,7 +269,7 @@ TEST(CoordinationTest, ChangelogReadWrite)
TEST(CoordinationTest, ChangelogWriteAt)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 1000, true);
DB::KeeperLogStore changelog("./logs", 1000, true);
changelog.init(1, 0);
for (size_t i = 0; i < 10; ++i)
{
@ -285,7 +285,7 @@ TEST(CoordinationTest, ChangelogWriteAt)
EXPECT_EQ(changelog.entry_at(7)->get_term(), 77);
EXPECT_EQ(changelog.next_slot(), 8);
DB::NuKeeperLogStore changelog_reader("./logs", 1000, true);
DB::KeeperLogStore changelog_reader("./logs", 1000, true);
changelog_reader.init(1, 0);
EXPECT_EQ(changelog_reader.size(), changelog.size());
@ -298,7 +298,7 @@ TEST(CoordinationTest, ChangelogWriteAt)
TEST(CoordinationTest, ChangelogTestAppendAfterRead)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 5, true);
DB::KeeperLogStore changelog("./logs", 5, true);
changelog.init(1, 0);
for (size_t i = 0; i < 7; ++i)
{
@ -310,7 +310,7 @@ TEST(CoordinationTest, ChangelogTestAppendAfterRead)
EXPECT_TRUE(fs::exists("./logs/changelog_1_5.bin"));
EXPECT_TRUE(fs::exists("./logs/changelog_6_10.bin"));
DB::NuKeeperLogStore changelog_reader("./logs", 5, true);
DB::KeeperLogStore changelog_reader("./logs", 5, true);
changelog_reader.init(1, 0);
EXPECT_EQ(changelog_reader.size(), 7);
@ -346,7 +346,7 @@ TEST(CoordinationTest, ChangelogTestAppendAfterRead)
TEST(CoordinationTest, ChangelogTestCompaction)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 5, true);
DB::KeeperLogStore changelog("./logs", 5, true);
changelog.init(1, 0);
for (size_t i = 0; i < 3; ++i)
@ -387,7 +387,7 @@ TEST(CoordinationTest, ChangelogTestCompaction)
EXPECT_EQ(changelog.next_slot(), 8);
EXPECT_EQ(changelog.last_entry()->get_term(), 60);
/// And we able to read it
DB::NuKeeperLogStore changelog_reader("./logs", 5, true);
DB::KeeperLogStore changelog_reader("./logs", 5, true);
changelog_reader.init(7, 0);
EXPECT_EQ(changelog_reader.size(), 1);
EXPECT_EQ(changelog_reader.start_index(), 7);
@ -398,7 +398,7 @@ TEST(CoordinationTest, ChangelogTestCompaction)
TEST(CoordinationTest, ChangelogTestBatchOperations)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 100, true);
DB::KeeperLogStore changelog("./logs", 100, true);
changelog.init(1, 0);
for (size_t i = 0; i < 10; ++i)
{
@ -410,7 +410,7 @@ TEST(CoordinationTest, ChangelogTestBatchOperations)
auto entries = changelog.pack(1, 5);
DB::NuKeeperLogStore apply_changelog("./logs", 100, true);
DB::KeeperLogStore apply_changelog("./logs", 100, true);
apply_changelog.init(1, 0);
for (size_t i = 0; i < 10; ++i)
@ -440,7 +440,7 @@ TEST(CoordinationTest, ChangelogTestBatchOperations)
TEST(CoordinationTest, ChangelogTestBatchOperationsEmpty)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 100, true);
DB::KeeperLogStore changelog("./logs", 100, true);
changelog.init(1, 0);
for (size_t i = 0; i < 10; ++i)
{
@ -453,7 +453,7 @@ TEST(CoordinationTest, ChangelogTestBatchOperationsEmpty)
auto entries = changelog.pack(5, 5);
ChangelogDirTest test1("./logs1");
DB::NuKeeperLogStore changelog_new("./logs1", 100, true);
DB::KeeperLogStore changelog_new("./logs1", 100, true);
changelog_new.init(1, 0);
EXPECT_EQ(changelog_new.size(), 0);
@ -472,7 +472,7 @@ TEST(CoordinationTest, ChangelogTestBatchOperationsEmpty)
EXPECT_EQ(changelog_new.start_index(), 5);
EXPECT_EQ(changelog_new.next_slot(), 11);
DB::NuKeeperLogStore changelog_reader("./logs1", 100, true);
DB::KeeperLogStore changelog_reader("./logs1", 100, true);
changelog_reader.init(5, 0);
}
@ -480,7 +480,7 @@ TEST(CoordinationTest, ChangelogTestBatchOperationsEmpty)
TEST(CoordinationTest, ChangelogTestWriteAtPreviousFile)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 5, true);
DB::KeeperLogStore changelog("./logs", 5, true);
changelog.init(1, 0);
for (size_t i = 0; i < 33; ++i)
@ -515,7 +515,7 @@ TEST(CoordinationTest, ChangelogTestWriteAtPreviousFile)
EXPECT_FALSE(fs::exists("./logs/changelog_26_30.bin"));
EXPECT_FALSE(fs::exists("./logs/changelog_31_35.bin"));
DB::NuKeeperLogStore changelog_read("./logs", 5, true);
DB::KeeperLogStore changelog_read("./logs", 5, true);
changelog_read.init(1, 0);
EXPECT_EQ(changelog_read.size(), 7);
EXPECT_EQ(changelog_read.start_index(), 1);
@ -526,7 +526,7 @@ TEST(CoordinationTest, ChangelogTestWriteAtPreviousFile)
TEST(CoordinationTest, ChangelogTestWriteAtFileBorder)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 5, true);
DB::KeeperLogStore changelog("./logs", 5, true);
changelog.init(1, 0);
for (size_t i = 0; i < 33; ++i)
@ -561,7 +561,7 @@ TEST(CoordinationTest, ChangelogTestWriteAtFileBorder)
EXPECT_FALSE(fs::exists("./logs/changelog_26_30.bin"));
EXPECT_FALSE(fs::exists("./logs/changelog_31_35.bin"));
DB::NuKeeperLogStore changelog_read("./logs", 5, true);
DB::KeeperLogStore changelog_read("./logs", 5, true);
changelog_read.init(1, 0);
EXPECT_EQ(changelog_read.size(), 11);
EXPECT_EQ(changelog_read.start_index(), 1);
@ -572,7 +572,7 @@ TEST(CoordinationTest, ChangelogTestWriteAtFileBorder)
TEST(CoordinationTest, ChangelogTestWriteAtAllFiles)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 5, true);
DB::KeeperLogStore changelog("./logs", 5, true);
changelog.init(1, 0);
for (size_t i = 0; i < 33; ++i)
@ -611,7 +611,7 @@ TEST(CoordinationTest, ChangelogTestWriteAtAllFiles)
TEST(CoordinationTest, ChangelogTestStartNewLogAfterRead)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 5, true);
DB::KeeperLogStore changelog("./logs", 5, true);
changelog.init(1, 0);
for (size_t i = 0; i < 35; ++i)
@ -630,7 +630,7 @@ TEST(CoordinationTest, ChangelogTestStartNewLogAfterRead)
EXPECT_FALSE(fs::exists("./logs/changelog_36_40.bin"));
DB::NuKeeperLogStore changelog_reader("./logs", 5, true);
DB::KeeperLogStore changelog_reader("./logs", 5, true);
changelog_reader.init(1, 0);
auto entry = getLogEntry("36_hello_world", 360);
@ -652,7 +652,7 @@ TEST(CoordinationTest, ChangelogTestReadAfterBrokenTruncate)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 5, true);
DB::KeeperLogStore changelog("./logs", 5, true);
changelog.init(1, 0);
for (size_t i = 0; i < 35; ++i)
@ -672,7 +672,7 @@ TEST(CoordinationTest, ChangelogTestReadAfterBrokenTruncate)
DB::WriteBufferFromFile plain_buf("./logs/changelog_11_15.bin", DBMS_DEFAULT_BUFFER_SIZE, O_APPEND | O_CREAT | O_WRONLY);
plain_buf.truncate(0);
DB::NuKeeperLogStore changelog_reader("./logs", 5, true);
DB::KeeperLogStore changelog_reader("./logs", 5, true);
changelog_reader.init(1, 0);
EXPECT_EQ(changelog_reader.size(), 10);
@ -701,7 +701,7 @@ TEST(CoordinationTest, ChangelogTestReadAfterBrokenTruncate)
EXPECT_FALSE(fs::exists("./logs/changelog_26_30.bin"));
EXPECT_FALSE(fs::exists("./logs/changelog_31_35.bin"));
DB::NuKeeperLogStore changelog_reader2("./logs", 5, true);
DB::KeeperLogStore changelog_reader2("./logs", 5, true);
changelog_reader2.init(1, 0);
EXPECT_EQ(changelog_reader2.size(), 11);
EXPECT_EQ(changelog_reader2.last_entry()->get_term(), 7777);
@ -711,7 +711,7 @@ TEST(CoordinationTest, ChangelogTestReadAfterBrokenTruncate2)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 20, true);
DB::KeeperLogStore changelog("./logs", 20, true);
changelog.init(1, 0);
for (size_t i = 0; i < 35; ++i)
@ -726,7 +726,7 @@ TEST(CoordinationTest, ChangelogTestReadAfterBrokenTruncate2)
DB::WriteBufferFromFile plain_buf("./logs/changelog_1_20.bin", DBMS_DEFAULT_BUFFER_SIZE, O_APPEND | O_CREAT | O_WRONLY);
plain_buf.truncate(140);
DB::NuKeeperLogStore changelog_reader("./logs", 20, true);
DB::KeeperLogStore changelog_reader("./logs", 20, true);
changelog_reader.init(1, 0);
EXPECT_EQ(changelog_reader.size(), 2);
@ -739,7 +739,7 @@ TEST(CoordinationTest, ChangelogTestReadAfterBrokenTruncate2)
EXPECT_EQ(changelog_reader.last_entry()->get_term(), 7777);
DB::NuKeeperLogStore changelog_reader2("./logs", 20, true);
DB::KeeperLogStore changelog_reader2("./logs", 20, true);
changelog_reader2.init(1, 0);
EXPECT_EQ(changelog_reader2.size(), 3);
EXPECT_EQ(changelog_reader2.last_entry()->get_term(), 7777);
@ -749,7 +749,7 @@ TEST(CoordinationTest, ChangelogTestLostFiles)
{
ChangelogDirTest test("./logs");
DB::NuKeeperLogStore changelog("./logs", 20, true);
DB::KeeperLogStore changelog("./logs", 20, true);
changelog.init(1, 0);
for (size_t i = 0; i < 35; ++i)
@ -763,7 +763,7 @@ TEST(CoordinationTest, ChangelogTestLostFiles)
fs::remove("./logs/changelog_1_20.bin");
DB::NuKeeperLogStore changelog_reader("./logs", 20, true);
DB::KeeperLogStore changelog_reader("./logs", 20, true);
/// It should print error message, but still able to start
changelog_reader.init(5, 0);
EXPECT_FALSE(fs::exists("./logs/changelog_1_20.bin"));
@ -862,9 +862,9 @@ TEST(CoordinationTest, SnapshotableHashMapTrySnapshot)
map_snp.disableSnapshotMode();
}
void addNode(DB::NuKeeperStorage & storage, const std::string & path, const std::string & data, int64_t ephemeral_owner=0)
void addNode(DB::KeeperStorage & storage, const std::string & path, const std::string & data, int64_t ephemeral_owner=0)
{
using Node = DB::NuKeeperStorage::Node;
using Node = DB::KeeperStorage::Node;
Node node{};
node.data = data;
node.stat.ephemeralOwner = ephemeral_owner;
@ -874,9 +874,9 @@ void addNode(DB::NuKeeperStorage & storage, const std::string & path, const std:
TEST(CoordinationTest, TestStorageSnapshotSimple)
{
ChangelogDirTest test("./snapshots");
DB::NuKeeperSnapshotManager manager("./snapshots", 3);
DB::KeeperSnapshotManager manager("./snapshots", 3);
DB::NuKeeperStorage storage(500);
DB::KeeperStorage storage(500);
addNode(storage, "/hello", "world", 1);
addNode(storage, "/hello/somepath", "somedata", 3);
storage.session_id_counter = 5;
@ -886,7 +886,7 @@ TEST(CoordinationTest, TestStorageSnapshotSimple)
storage.getSessionID(130);
storage.getSessionID(130);
DB::NuKeeperStorageSnapshot snapshot(&storage, 2);
DB::KeeperStorageSnapshot snapshot(&storage, 2);
EXPECT_EQ(snapshot.snapshot_meta->get_last_log_idx(), 2);
EXPECT_EQ(snapshot.session_id, 7);
@ -921,9 +921,9 @@ TEST(CoordinationTest, TestStorageSnapshotSimple)
TEST(CoordinationTest, TestStorageSnapshotMoreWrites)
{
ChangelogDirTest test("./snapshots");
DB::NuKeeperSnapshotManager manager("./snapshots", 3);
DB::KeeperSnapshotManager manager("./snapshots", 3);
DB::NuKeeperStorage storage(500);
DB::KeeperStorage storage(500);
storage.getSessionID(130);
for (size_t i = 0; i < 50; ++i)
@ -931,7 +931,7 @@ TEST(CoordinationTest, TestStorageSnapshotMoreWrites)
addNode(storage, "/hello_" + std::to_string(i), "world_" + std::to_string(i));
}
DB::NuKeeperStorageSnapshot snapshot(&storage, 50);
DB::KeeperStorageSnapshot snapshot(&storage, 50);
EXPECT_EQ(snapshot.snapshot_meta->get_last_log_idx(), 50);
EXPECT_EQ(snapshot.snapshot_container_size, 51);
@ -961,9 +961,9 @@ TEST(CoordinationTest, TestStorageSnapshotMoreWrites)
TEST(CoordinationTest, TestStorageSnapshotManySnapshots)
{
ChangelogDirTest test("./snapshots");
DB::NuKeeperSnapshotManager manager("./snapshots", 3);
DB::KeeperSnapshotManager manager("./snapshots", 3);
DB::NuKeeperStorage storage(500);
DB::KeeperStorage storage(500);
storage.getSessionID(130);
for (size_t j = 1; j <= 5; ++j)
@ -973,7 +973,7 @@ TEST(CoordinationTest, TestStorageSnapshotManySnapshots)
addNode(storage, "/hello_" + std::to_string(i), "world_" + std::to_string(i));
}
DB::NuKeeperStorageSnapshot snapshot(&storage, j * 50);
DB::KeeperStorageSnapshot snapshot(&storage, j * 50);
auto buf = manager.serializeSnapshotToBuffer(snapshot);
manager.serializeSnapshotBufferToDisk(*buf, j * 50);
EXPECT_TRUE(fs::exists(std::string{"./snapshots/snapshot_"} + std::to_string(j * 50) + ".bin"));
@ -999,15 +999,15 @@ TEST(CoordinationTest, TestStorageSnapshotManySnapshots)
TEST(CoordinationTest, TestStorageSnapshotMode)
{
ChangelogDirTest test("./snapshots");
DB::NuKeeperSnapshotManager manager("./snapshots", 3);
DB::NuKeeperStorage storage(500);
DB::KeeperSnapshotManager manager("./snapshots", 3);
DB::KeeperStorage storage(500);
for (size_t i = 0; i < 50; ++i)
{
addNode(storage, "/hello_" + std::to_string(i), "world_" + std::to_string(i));
}
{
DB::NuKeeperStorageSnapshot snapshot(&storage, 50);
DB::KeeperStorageSnapshot snapshot(&storage, 50);
for (size_t i = 0; i < 50; ++i)
{
addNode(storage, "/hello_" + std::to_string(i), "wlrd_" + std::to_string(i));
@ -1050,14 +1050,14 @@ TEST(CoordinationTest, TestStorageSnapshotMode)
TEST(CoordinationTest, TestStorageSnapshotBroken)
{
ChangelogDirTest test("./snapshots");
DB::NuKeeperSnapshotManager manager("./snapshots", 3);
DB::NuKeeperStorage storage(500);
DB::KeeperSnapshotManager manager("./snapshots", 3);
DB::KeeperStorage storage(500);
for (size_t i = 0; i < 50; ++i)
{
addNode(storage, "/hello_" + std::to_string(i), "world_" + std::to_string(i));
}
{
DB::NuKeeperStorageSnapshot snapshot(&storage, 50);
DB::KeeperStorageSnapshot snapshot(&storage, 50);
auto buf = manager.serializeSnapshotToBuffer(snapshot);
manager.serializeSnapshotBufferToDisk(*buf, 50);
}
@ -1095,9 +1095,9 @@ void testLogAndStateMachine(Coordination::CoordinationSettingsPtr settings, size
ResponsesQueue queue;
SnapshotsQueue snapshots_queue{1};
auto state_machine = std::make_shared<NuKeeperStateMachine>(queue, snapshots_queue, "./snapshots", settings);
auto state_machine = std::make_shared<KeeperStateMachine>(queue, snapshots_queue, "./snapshots", settings);
state_machine->init();
DB::NuKeeperLogStore changelog("./logs", settings->rotate_log_storage_interval, true);
DB::KeeperLogStore changelog("./logs", settings->rotate_log_storage_interval, true);
changelog.init(state_machine->last_commit_index() + 1, settings->reserved_log_items);
for (size_t i = 1; i < total_logs + 1; ++i)
{
@ -1132,11 +1132,11 @@ void testLogAndStateMachine(Coordination::CoordinationSettingsPtr settings, size
}
SnapshotsQueue snapshots_queue1{1};
auto restore_machine = std::make_shared<NuKeeperStateMachine>(queue, snapshots_queue1, "./snapshots", settings);
auto restore_machine = std::make_shared<KeeperStateMachine>(queue, snapshots_queue1, "./snapshots", settings);
restore_machine->init();
EXPECT_EQ(restore_machine->last_commit_index(), total_logs - total_logs % settings->snapshot_distance);
DB::NuKeeperLogStore restore_changelog("./logs", settings->rotate_log_storage_interval, true);
DB::KeeperLogStore restore_changelog("./logs", settings->rotate_log_storage_interval, true);
restore_changelog.init(restore_machine->last_commit_index() + 1, settings->reserved_log_items);
EXPECT_EQ(restore_changelog.size(), std::min(settings->reserved_log_items + total_logs % settings->snapshot_distance, total_logs));
@ -1242,7 +1242,7 @@ TEST(CoordinationTest, TestEphemeralNodeRemove)
ResponsesQueue queue;
SnapshotsQueue snapshots_queue{1};
auto state_machine = std::make_shared<NuKeeperStateMachine>(queue, snapshots_queue, "./snapshots", settings);
auto state_machine = std::make_shared<KeeperStateMachine>(queue, snapshots_queue, "./snapshots", settings);
state_machine->init();
std::shared_ptr<ZooKeeperCreateRequest> request_c = std::make_shared<ZooKeeperCreateRequest>();

1
src/Core/Settings.h
View File

@ -240,6 +240,7 @@ class IColumn;
M(Bool, metrics_perf_events_enabled, false, "If enabled, some of the perf events will be measured throughout queries' execution.", 0) \
M(String, metrics_perf_events_list, "", "Comma separated list of perf metrics that will be measured throughout queries' execution. Empty means all events. See PerfEventInfo in sources for the available events.", 0) \
M(Float, opentelemetry_start_trace_probability, 0., "Probability to start an OpenTelemetry trace for an incoming query.", 0) \
M(Bool, prefer_column_name_to_alias, false, "Prefer using column names instead of aliases if possible.", 0) \
\
\
/** Limits during query execution are part of the settings. \

2
src/Dictionaries/CacheDictionaryStorage.h
View File

@ -188,8 +188,6 @@ private:
size_t fetched_columns_index = 0;
size_t keys_size = keys.size();
std::chrono::seconds max_lifetime_seconds(configuration.strict_max_lifetime_seconds);
PaddedPODArray<FetchedKey> fetched_keys;
fetched_keys.resize_fill(keys_size);

57
src/Dictionaries/IPAddressDictionary.cpp
View File

@ -4,19 +4,17 @@
#include <Common/assert_cast.h>
#include <Common/IPv6ToBinary.h>
#include <Common/memcmpSmall.h>
#include <Common/memcpySmall.h>
#include <Common/typeid_cast.h>
#include <DataTypes/DataTypeFixedString.h>
#include <DataTypes/DataTypeString.h>
#include <DataTypes/DataTypesDecimal.h>
#include <IO/WriteIntText.h>
#include <Poco/ByteOrder.h>
#include <Common/formatIPv6.h>
#include <common/itoa.h>
#include <ext/map.h>
#include <ext/range.h>
#include "DictionaryBlockInputStream.h"
#include "DictionaryFactory.h"
#include <Dictionaries/DictionaryBlockInputStream.h>
#include <Dictionaries/DictionaryFactory.h>
#include <Functions/FunctionHelpers.h>
namespace DB
@ -191,57 +189,6 @@ inline static void mapIPv4ToIPv6(UInt32 addr, uint8_t * buf)
memcpy(&buf[12], &addr, 4);
}
static bool matchIPv4Subnet(UInt32 target, UInt32 addr, UInt8 prefix)
{
UInt32 mask = (prefix >= 32) ? 0xffffffffu : ~(0xffffffffu >> prefix);
return (target & mask) == addr;
}
#if defined(__SSE2__)
#include <emmintrin.h>
static bool matchIPv6Subnet(const uint8_t * target, const uint8_t * addr, UInt8 prefix)
{
uint16_t mask = _mm_movemask_epi8(_mm_cmpeq_epi8(
_mm_loadu_si128(reinterpret_cast<const __m128i *>(target)),
_mm_loadu_si128(reinterpret_cast<const __m128i *>(addr))));
mask = ~mask;
if (mask)
{
auto offset = __builtin_ctz(mask);
if (prefix / 8 != offset)
return prefix / 8 < offset;
auto cmpmask = ~(0xff >> (prefix % 8));
return (target[offset] & cmpmask) == addr[offset];
}
return true;
}
# else
static bool matchIPv6Subnet(const uint8_t * target, const uint8_t * addr, UInt8 prefix)
{
if (prefix > IPV6_BINARY_LENGTH * 8U)
prefix = IPV6_BINARY_LENGTH * 8U;
size_t i = 0;
for (; prefix >= 8; ++i, prefix -= 8)
{
if (target[i] != addr[i])
return false;
}
if (prefix == 0)
return true;
auto mask = ~(0xff >> prefix);
return (target[i] & mask) == addr[i];
}
#endif // __SSE2__
IPAddressDictionary::IPAddressDictionary(
const StorageID & dict_id_,
const DictionaryStructure & dict_struct_,

2
src/Dictionaries/PolygonDictionary.cpp
View File

@ -271,7 +271,7 @@ std::vector<IPolygonDictionary::Point> IPolygonDictionary::extractPoints(const C
throw Exception(ErrorCodes::BAD_ARGUMENTS,
"PolygonDictionary input point component must not be NaN");
if (isinf(x) || isinf(y))
if (std::isinf(x) || std::isinf(y))
throw Exception(ErrorCodes::BAD_ARGUMENTS,
"PolygonDictionary input point component must not be infinite");

4
src/Functions/FunctionsCoding.h
View File

@ -1645,7 +1645,7 @@ private:
static inline void applyCIDRMask(const UInt8 * __restrict src, UInt8 * __restrict dst_lower, UInt8 * __restrict dst_upper, UInt8 bits_to_keep)
{
__m128i mask = _mm_loadu_si128(reinterpret_cast<const __m128i *>(getCIDRMaskIPv6(bits_to_keep)));
__m128i mask = _mm_loadu_si128(reinterpret_cast<const __m128i *>(getCIDRMaskIPv6(bits_to_keep).data()));
__m128i lower = _mm_and_si128(_mm_loadu_si128(reinterpret_cast<const __m128i *>(src)), mask);
_mm_storeu_si128(reinterpret_cast<__m128i *>(dst_lower), lower);
@ -1659,7 +1659,7 @@ private:
/// NOTE IPv6 is stored in memory in big endian format that makes some difficulties.
static void applyCIDRMask(const UInt8 * __restrict src, UInt8 * __restrict dst_lower, UInt8 * __restrict dst_upper, UInt8 bits_to_keep)
{
const auto * mask = getCIDRMaskIPv6(bits_to_keep);
const auto & mask = getCIDRMaskIPv6(bits_to_keep);
for (size_t i = 0; i < 16; ++i)
{

2
src/Functions/geometryConverters.h
View File

@ -76,7 +76,7 @@ struct ColumnToPointsConverter
if (isNaN(first) || isNaN(second))
throw Exception("Point's component must not be NaN", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
if (isinf(first) || isinf(second))
if (std::isinf(first) || std::isinf(second))
throw Exception("Point's component must not be infinite", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
answer[i] = Point(first, second);

259
src/Functions/isIPAddressContainedIn.cpp Normal file
View File

@ -0,0 +1,259 @@
#include <Columns/ColumnConst.h>
#include <Columns/ColumnString.h>
#include <Columns/ColumnsNumber.h>
#include <Common/IPv6ToBinary.h>
#include <Common/formatIPv6.h>
#include <DataTypes/DataTypeNullable.h>
#include <DataTypes/DataTypesNumber.h>
#include <Functions/IFunctionImpl.h>
#include <Functions/FunctionFactory.h>
#include <Functions/FunctionHelpers.h>
#include <variant>
#include <charconv>
#include <common/logger_useful.h>
namespace DB::ErrorCodes
{
extern const int CANNOT_PARSE_TEXT;
extern const int ILLEGAL_TYPE_OF_ARGUMENT;
extern const int NUMBER_OF_ARGUMENTS_DOESNT_MATCH;
}
namespace
{
class IPAddressVariant
{
public:
explicit IPAddressVariant(const StringRef & address_str)
{
/// IP address parser functions require that the input is
/// NULL-terminated so we need to copy it.
const auto address_str_copy = std::string(address_str);
UInt32 v4;
if (DB::parseIPv4(address_str_copy.c_str(), reinterpret_cast<unsigned char *>(&v4)))
{
addr = v4;
}
else
{
addr = IPv6AddrType();
bool success = DB::parseIPv6(address_str_copy.c_str(), std::get<IPv6AddrType>(addr).data());
if (!success)
throw DB::Exception("Neither IPv4 nor IPv6 address: '" + address_str_copy + "'",
DB::ErrorCodes::CANNOT_PARSE_TEXT);
}
}
UInt32 asV4() const
{
if (const auto * val = std::get_if<IPv4AddrType>(&addr))
return *val;
return 0;
}
const uint8_t * asV6() const
{
if (const auto * val = std::get_if<IPv6AddrType>(&addr))
return val->data();
return nullptr;
}
private:
using IPv4AddrType = UInt32;
using IPv6AddrType = std::array<uint8_t, IPV6_BINARY_LENGTH>;
std::variant<IPv4AddrType, IPv6AddrType> addr;
};
struct IPAddressCIDR
{
IPAddressVariant address;
UInt8 prefix;
};
IPAddressCIDR parseIPWithCIDR(const StringRef cidr_str)
{
std::string_view cidr_str_view(cidr_str);
size_t pos_slash = cidr_str_view.find('/');
if (pos_slash == 0)
throw DB::Exception("Error parsing IP address with prefix: " + std::string(cidr_str), DB::ErrorCodes::CANNOT_PARSE_TEXT);
if (pos_slash == std::string_view::npos)
throw DB::Exception("The text does not contain '/': " + std::string(cidr_str), DB::ErrorCodes::CANNOT_PARSE_TEXT);
std::string_view addr_str = cidr_str_view.substr(0, pos_slash);
IPAddressVariant addr(StringRef{addr_str.data(), addr_str.size()});
uint8_t prefix = 0;
auto prefix_str = cidr_str_view.substr(pos_slash+1);
const auto * prefix_str_end = prefix_str.data() + prefix_str.size();
auto [parse_end, parse_error] = std::from_chars(prefix_str.data(), prefix_str_end, prefix);
uint8_t max_prefix = (addr.asV6() ? IPV6_BINARY_LENGTH : IPV4_BINARY_LENGTH) * 8;
bool has_error = parse_error != std::errc() || parse_end != prefix_str_end || prefix > max_prefix;
if (has_error)
throw DB::Exception("The CIDR has a malformed prefix bits: " + std::string(cidr_str), DB::ErrorCodes::CANNOT_PARSE_TEXT);
return {addr, static_cast<UInt8>(prefix)};
}
inline bool isAddressInRange(const IPAddressVariant & address, const IPAddressCIDR & cidr)
{
if (const auto * cidr_v6 = cidr.address.asV6())
{
if (const auto * addr_v6 = address.asV6())
return DB::matchIPv6Subnet(addr_v6, cidr_v6, cidr.prefix);
}
else
{
if (!address.asV6())
return DB::matchIPv4Subnet(address.asV4(), cidr.address.asV4(), cidr.prefix);
}
return false;
}
}
namespace DB
{
class FunctionIsIPAddressContainedIn : public IFunction
{
public:
static constexpr auto name = "isIPAddressInRange";
String getName() const override { return name; }
static FunctionPtr create(const Context &) { return std::make_shared<FunctionIsIPAddressContainedIn>(); }
ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr & /* return_type */, size_t input_rows_count) const override
{
const IColumn * col_addr = arguments[0].column.get();
const IColumn * col_cidr = arguments[1].column.get();
if (const auto * col_addr_const = checkAndGetAnyColumnConst(col_addr))
{
if (const auto * col_cidr_const = checkAndGetAnyColumnConst(col_cidr))
return executeImpl(*col_addr_const, *col_cidr_const, input_rows_count);
else
return executeImpl(*col_addr_const, *col_cidr, input_rows_count);
}
else
{
if (const auto * col_cidr_const = checkAndGetAnyColumnConst(col_cidr))
return executeImpl(*col_addr, *col_cidr_const, input_rows_count);
else
return executeImpl(*col_addr, *col_cidr, input_rows_count);
}
}
virtual DataTypePtr getReturnTypeImpl(const DataTypes & arguments) const override
{
if (arguments.size() != 2)
throw Exception(
"Number of arguments for function " + getName() + " doesn't match: passed " + toString(arguments.size()) + ", should be 2",
ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH);
const DataTypePtr & addr_type = arguments[0];
const DataTypePtr & prefix_type = arguments[1];
if (!isString(addr_type) || !isString(prefix_type))
throw Exception("The arguments of function " + getName() + " must be String",
ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
return std::make_shared<DataTypeUInt8>();
}
virtual size_t getNumberOfArguments() const override { return 2; }
bool useDefaultImplementationForNulls() const override { return false; }
private:
/// Like checkAndGetColumnConst() but this function doesn't
/// care about the type of data column.
static const ColumnConst * checkAndGetAnyColumnConst(const IColumn * column)
{
if (!column || !isColumnConst(*column))
return nullptr;
return assert_cast<const ColumnConst *>(column);
}
/// Both columns are constant.
static ColumnPtr executeImpl(
const ColumnConst & col_addr_const,
const ColumnConst & col_cidr_const,
size_t input_rows_count)
{
const auto & col_addr = col_addr_const.getDataColumn();
const auto & col_cidr = col_cidr_const.getDataColumn();
const auto addr = IPAddressVariant(col_addr.getDataAt(0));
const auto cidr = parseIPWithCIDR(col_cidr.getDataAt(0));
ColumnUInt8::MutablePtr col_res = ColumnUInt8::create(1);
ColumnUInt8::Container & vec_res = col_res->getData();
vec_res[0] = isAddressInRange(addr, cidr) ? 1 : 0;
return ColumnConst::create(std::move(col_res), input_rows_count);
}
/// Address is constant.
static ColumnPtr executeImpl(const ColumnConst & col_addr_const, const IColumn & col_cidr, size_t input_rows_count)
{
const auto & col_addr = col_addr_const.getDataColumn();
const auto addr = IPAddressVariant(col_addr.getDataAt (0));
ColumnUInt8::MutablePtr col_res = ColumnUInt8::create(input_rows_count);
ColumnUInt8::Container & vec_res = col_res->getData();
for (size_t i = 0; i < input_rows_count; i++)
{
const auto cidr = parseIPWithCIDR(col_cidr.getDataAt(i));
vec_res[i] = isAddressInRange(addr, cidr) ? 1 : 0;
}
return col_res;
}
/// CIDR is constant.
static ColumnPtr executeImpl(const IColumn & col_addr, const ColumnConst & col_cidr_const, size_t input_rows_count)
{
const auto & col_cidr = col_cidr_const.getDataColumn();
const auto cidr = parseIPWithCIDR(col_cidr.getDataAt(0));
ColumnUInt8::MutablePtr col_res = ColumnUInt8::create(input_rows_count);
ColumnUInt8::Container & vec_res = col_res->getData();
for (size_t i = 0; i < input_rows_count; i++)
{
const auto addr = IPAddressVariant(col_addr.getDataAt(i));
vec_res[i] = isAddressInRange(addr, cidr) ? 1 : 0;
}
return col_res;
}
/// Neither are constant.
static ColumnPtr executeImpl(const IColumn & col_addr, const IColumn & col_cidr, size_t input_rows_count)
{
ColumnUInt8::MutablePtr col_res = ColumnUInt8::create(input_rows_count);
ColumnUInt8::Container & vec_res = col_res->getData();
for (size_t i = 0; i < input_rows_count; i++)
{
const auto addr = IPAddressVariant(col_addr.getDataAt(i));
const auto cidr = parseIPWithCIDR(col_cidr.getDataAt(i));
vec_res[i] = isAddressInRange(addr, cidr) ? 1 : 0;
}
return col_res;
}
};
void registerFunctionIsIPAddressContainedIn(FunctionFactory & factory)
{
factory.registerFunction<FunctionIsIPAddressContainedIn>();
}
}

2
src/Functions/registerFunctionsMiscellaneous.cpp
View File

@ -72,6 +72,7 @@ void registerFunctionByteSize(FunctionFactory &);
void registerFunctionFile(FunctionFactory & factory);
void registerFunctionConnectionId(FunctionFactory & factory);
void registerFunctionPartitionId(FunctionFactory & factory);
void registerFunctionIsIPAddressContainedIn(FunctionFactory &);
#if USE_ICU
void registerFunctionConvertCharset(FunctionFactory &);
@ -144,6 +145,7 @@ void registerFunctionsMiscellaneous(FunctionFactory & factory)
registerFunctionFile(factory);
registerFunctionConnectionId(factory);
registerFunctionPartitionId(factory);
registerFunctionIsIPAddressContainedIn(factory);
#if USE_ICU
registerFunctionConvertCharset(factory);

1
src/Functions/ya.make
View File

@ -309,6 +309,7 @@ SRCS(
isConstant.cpp
isDecimalOverflow.cpp
isFinite.cpp
isIPAddressContainedIn.cpp
isInfinite.cpp
isNaN.cpp
isNotNull.cpp

7
src/IO/ReadBufferFromPocoSocket.cpp
View File

@ -37,13 +37,6 @@ bool ReadBufferFromPocoSocket::nextImpl()
while (async_callback && !socket.poll(0, Poco::Net::Socket::SELECT_READ))
async_callback(socket.impl()->sockfd(), socket.getReceiveTimeout(), socket_description);
/// receiveBytes in SecureStreamSocket throws TimeoutException after max(receive_timeout, send_timeout),
/// but we want to get this exception exactly after receive_timeout. So, set send_timeout = receive_timeout
/// before receiveBytes.
std::unique_ptr<TimeoutSetter> timeout_setter = nullptr;
if (socket.secure())
timeout_setter = std::make_unique<TimeoutSetter>(dynamic_cast<Poco::Net::StreamSocket &>(socket), socket.getReceiveTimeout(), socket.getReceiveTimeout());
bytes_read = socket.impl()->receiveBytes(internal_buffer.begin(), internal_buffer.size());
}
catch (const Poco::Net::NetException & e)

7
src/IO/WriteBufferFromPocoSocket.cpp
View File

@ -41,13 +41,6 @@ void WriteBufferFromPocoSocket::nextImpl()
/// Add more details to exceptions.
try
{
/// sendBytes in SecureStreamSocket throws TimeoutException after max(receive_timeout, send_timeout),
/// but we want to get this exception exactly after send_timeout. So, set receive_timeout = send_timeout
/// before sendBytes.
std::unique_ptr<TimeoutSetter> timeout_setter = nullptr;
if (socket.secure())
timeout_setter = std::make_unique<TimeoutSetter>(dynamic_cast<Poco::Net::StreamSocket &>(socket), socket.getSendTimeout(), socket.getSendTimeout());
res = socket.impl()->sendBytes(working_buffer.begin() + bytes_written, offset() - bytes_written);
}
catch (const Poco::Net::NetException & e)

18
src/Interpreters/Context.cpp
View File

@ -12,7 +12,7 @@
#include <Common/Stopwatch.h>
#include <Common/formatReadable.h>
#include <Common/thread_local_rng.h>
#include <Coordination/NuKeeperStorageDispatcher.h>
#include <Coordination/KeeperStorageDispatcher.h>
#include <Compression/ICompressionCodec.h>
#include <Core/BackgroundSchedulePool.h>
#include <Formats/FormatFactory.h>
@ -314,7 +314,7 @@ struct ContextShared
#if USE_NURAFT
mutable std::mutex nu_keeper_storage_dispatcher_mutex;
mutable std::shared_ptr<NuKeeperStorageDispatcher> nu_keeper_storage_dispatcher;
mutable std::shared_ptr<KeeperStorageDispatcher> nu_keeper_storage_dispatcher;
#endif
mutable std::mutex auxiliary_zookeepers_mutex;
mutable std::map<String, zkutil::ZooKeeperPtr> auxiliary_zookeepers; /// Map for auxiliary ZooKeeper clients.
@ -1616,35 +1616,35 @@ zkutil::ZooKeeperPtr Context::getZooKeeper() const
}
void Context::initializeNuKeeperStorageDispatcher() const
void Context::initializeKeeperStorageDispatcher() const
{
#if USE_NURAFT
std::lock_guard lock(shared->nu_keeper_storage_dispatcher_mutex);
if (shared->nu_keeper_storage_dispatcher)
throw Exception(ErrorCodes::LOGICAL_ERROR, "Trying to initialize NuKeeper multiple times");
throw Exception(ErrorCodes::LOGICAL_ERROR, "Trying to initialize Keeper multiple times");
const auto & config = getConfigRef();
if (config.has("test_keeper_server"))
if (config.has("keeper_server"))
{
shared->nu_keeper_storage_dispatcher = std::make_shared<NuKeeperStorageDispatcher>();
shared->nu_keeper_storage_dispatcher = std::make_shared<KeeperStorageDispatcher>();
shared->nu_keeper_storage_dispatcher->initialize(config);
}
#endif
}
#if USE_NURAFT
std::shared_ptr<NuKeeperStorageDispatcher> & Context::getNuKeeperStorageDispatcher() const
std::shared_ptr<KeeperStorageDispatcher> & Context::getKeeperStorageDispatcher() const
{
std::lock_guard lock(shared->nu_keeper_storage_dispatcher_mutex);
if (!shared->nu_keeper_storage_dispatcher)
throw Exception(ErrorCodes::LOGICAL_ERROR, "NuKeeper must be initialized before requests");
throw Exception(ErrorCodes::LOGICAL_ERROR, "Keeper must be initialized before requests");
return shared->nu_keeper_storage_dispatcher;
}
#endif
void Context::shutdownNuKeeperStorageDispatcher() const
void Context::shutdownKeeperStorageDispatcher() const
{
#if USE_NURAFT
std::lock_guard lock(shared->nu_keeper_storage_dispatcher_mutex);

8
src/Interpreters/Context.h
View File

@ -111,7 +111,7 @@ class StoragePolicySelector;
using StoragePolicySelectorPtr = std::shared_ptr<const StoragePolicySelector>;
struct PartUUIDs;
using PartUUIDsPtr = std::shared_ptr<PartUUIDs>;
class NuKeeperStorageDispatcher;
class KeeperStorageDispatcher;
class IOutputFormat;
using OutputFormatPtr = std::shared_ptr<IOutputFormat>;
@ -598,10 +598,10 @@ public:
std::shared_ptr<zkutil::ZooKeeper> getAuxiliaryZooKeeper(const String & name) const;
#if USE_NURAFT
std::shared_ptr<NuKeeperStorageDispatcher> & getNuKeeperStorageDispatcher() const;
std::shared_ptr<KeeperStorageDispatcher> & getKeeperStorageDispatcher() const;
#endif
void initializeNuKeeperStorageDispatcher() const;
void shutdownNuKeeperStorageDispatcher() const;
void initializeKeeperStorageDispatcher() const;
void shutdownKeeperStorageDispatcher() const;
/// Set auxiliary zookeepers configuration at server starting or configuration reloading.
void reloadAuxiliaryZooKeepersConfigIfChanged(const ConfigurationPtr & config);

16
src/Interpreters/MySQL/InterpretersMySQLDDLQuery.cpp
View File

@ -7,6 +7,7 @@
#include <Parsers/ASTAlterQuery.h>
#include <Parsers/ASTCreateQuery.h>
#include <Parsers/ASTColumnDeclaration.h>
#include <Parsers/ASTIndexDeclaration.h>
#include <Parsers/MySQL/ASTCreateQuery.h>
#include <Parsers/MySQL/ASTAlterCommand.h>
#include <Parsers/MySQL/ASTDeclareColumn.h>
@ -411,13 +412,26 @@ ASTs InterpreterCreateImpl::getRewrittenQueries(
return column_declaration;
};
/// Add _sign and _version column.
/// Add _sign and _version columns.
String sign_column_name = getUniqueColumnName(columns_name_and_type, "_sign");
String version_column_name = getUniqueColumnName(columns_name_and_type, "_version");
columns->set(columns->columns, InterpreterCreateQuery::formatColumns(columns_name_and_type));
columns->columns->children.emplace_back(create_materialized_column_declaration(sign_column_name, "Int8", UInt64(1)));
columns->columns->children.emplace_back(create_materialized_column_declaration(version_column_name, "UInt64", UInt64(1)));
/// Add minmax skipping index for _version column.
auto version_index = std::make_shared<ASTIndexDeclaration>();
version_index->name = version_column_name;
auto index_expr = std::make_shared<ASTIdentifier>(version_column_name);
auto index_type = makeASTFunction("minmax");
index_type->no_empty_args = true;
version_index->set(version_index->expr, index_expr);
version_index->set(version_index->type, index_type);
version_index->granularity = 1;
ASTPtr indices = std::make_shared<ASTExpressionList>();
indices->children.push_back(version_index);
columns->set(columns->indices, indices);
auto storage = std::make_shared<ASTStorage>();
/// The `partition by` expression must use primary keys, otherwise the primary keys will not be merge.

82
src/Interpreters/MySQL/tests/gtest_create_rewritten.cpp
View File

@ -28,6 +28,10 @@ static inline ASTPtr tryRewrittenCreateQuery(const String & query, const Context
context, "test_database", "test_database")[0];
}
static const char MATERIALIZEMYSQL_TABLE_COLUMNS[] = ", `_sign` Int8() MATERIALIZED 1"
", `_version` UInt64() MATERIALIZED 1"
", INDEX _version _version TYPE minmax GRANULARITY 1";
TEST(MySQLCreateRewritten, ColumnsDataType)
{
tryRegisterFunctions();
@ -45,46 +49,46 @@ TEST(MySQLCreateRewritten, ColumnsDataType)
{
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1`(`key` INT NOT NULL PRIMARY KEY, test " + test_type + ")", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` Int32, `test` Nullable(" + mapped_type + ")"
", `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1) ENGINE = "
"CREATE TABLE test_database.test_table_1 (`key` Int32, `test` Nullable(" + mapped_type + ")" +
MATERIALIZEMYSQL_TABLE_COLUMNS + ") ENGINE = "
"ReplacingMergeTree(_version) PARTITION BY intDiv(key, 4294967) ORDER BY tuple(key)");
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1`(`key` INT NOT NULL PRIMARY KEY, test " + test_type + " NOT NULL)", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` Int32, `test` " + mapped_type +
", `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1) ENGINE = "
MATERIALIZEMYSQL_TABLE_COLUMNS + ") ENGINE = "
"ReplacingMergeTree(_version) PARTITION BY intDiv(key, 4294967) ORDER BY tuple(key)");
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1`(`key` INT NOT NULL PRIMARY KEY, test " + test_type + " COMMENT 'test_comment' NOT NULL)", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` Int32, `test` " + mapped_type +
", `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1) ENGINE = "
MATERIALIZEMYSQL_TABLE_COLUMNS + ") ENGINE = "
"ReplacingMergeTree(_version) PARTITION BY intDiv(key, 4294967) ORDER BY tuple(key)");
if (Poco::toUpper(test_type).find("INT") != std::string::npos)
{
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1`(`key` INT NOT NULL PRIMARY KEY, test " + test_type + " UNSIGNED)", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` Int32, `test` Nullable(U" + mapped_type + ")"
", `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1) ENGINE = "
"CREATE TABLE test_database.test_table_1 (`key` Int32, `test` Nullable(U" + mapped_type + ")" +
MATERIALIZEMYSQL_TABLE_COLUMNS + ") ENGINE = "
"ReplacingMergeTree(_version) PARTITION BY intDiv(key, 4294967) ORDER BY tuple(key)");
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1`(`key` INT NOT NULL PRIMARY KEY, test " + test_type + " COMMENT 'test_comment' UNSIGNED)", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` Int32, `test` Nullable(U" + mapped_type + ")"
", `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1) ENGINE = "
"CREATE TABLE test_database.test_table_1 (`key` Int32, `test` Nullable(U" + mapped_type + ")" +
MATERIALIZEMYSQL_TABLE_COLUMNS + ") ENGINE = "
"ReplacingMergeTree(_version) PARTITION BY intDiv(key, 4294967) ORDER BY tuple(key)");
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1`(`key` INT NOT NULL PRIMARY KEY, test " + test_type + " NOT NULL UNSIGNED)", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` Int32, `test` U" + mapped_type +
", `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1) ENGINE = "
MATERIALIZEMYSQL_TABLE_COLUMNS + ") ENGINE = "
"ReplacingMergeTree(_version) PARTITION BY intDiv(key, 4294967) ORDER BY tuple(key)");
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1`(`key` INT NOT NULL PRIMARY KEY, test " + test_type + " COMMENT 'test_comment' UNSIGNED NOT NULL)", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` Int32, `test` U" + mapped_type +
", `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1) ENGINE = "
MATERIALIZEMYSQL_TABLE_COLUMNS + ") ENGINE = "
"ReplacingMergeTree(_version) PARTITION BY intDiv(key, 4294967) ORDER BY tuple(key)");
}
}
@ -109,13 +113,15 @@ TEST(MySQLCreateRewritten, PartitionPolicy)
{
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1` (`key` " + test_type + " PRIMARY KEY)", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` " + mapped_type + ", `_sign` Int8() MATERIALIZED 1, "
"`_version` UInt64() MATERIALIZED 1) ENGINE = ReplacingMergeTree(_version)" + partition_policy + " ORDER BY tuple(key)");
"CREATE TABLE test_database.test_table_1 (`key` " + mapped_type +
MATERIALIZEMYSQL_TABLE_COLUMNS +
") ENGINE = ReplacingMergeTree(_version)" + partition_policy + " ORDER BY tuple(key)");
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1` (`key` " + test_type + " NOT NULL PRIMARY KEY)", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` " + mapped_type + ", `_sign` Int8() MATERIALIZED 1, "
"`_version` UInt64() MATERIALIZED 1) ENGINE = ReplacingMergeTree(_version)" + partition_policy + " ORDER BY tuple(key)");
"CREATE TABLE test_database.test_table_1 (`key` " + mapped_type +
MATERIALIZEMYSQL_TABLE_COLUMNS +
") ENGINE = ReplacingMergeTree(_version)" + partition_policy + " ORDER BY tuple(key)");
}
}
@ -138,23 +144,27 @@ TEST(MySQLCreateRewritten, OrderbyPolicy)
{
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1` (`key` " + test_type + " PRIMARY KEY, `key2` " + test_type + " UNIQUE KEY)", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` " + mapped_type + ", `key2` Nullable(" + mapped_type + "), `_sign` Int8() MATERIALIZED 1, "
"`_version` UInt64() MATERIALIZED 1) ENGINE = ReplacingMergeTree(_version)" + partition_policy + " ORDER BY (key, assumeNotNull(key2))");
"CREATE TABLE test_database.test_table_1 (`key` " + mapped_type + ", `key2` Nullable(" + mapped_type + ")" +
MATERIALIZEMYSQL_TABLE_COLUMNS +
") ENGINE = ReplacingMergeTree(_version)" + partition_policy + " ORDER BY (key, assumeNotNull(key2))");
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1` (`key` " + test_type + " NOT NULL PRIMARY KEY, `key2` " + test_type + " NOT NULL UNIQUE KEY)", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` " + mapped_type + ", `key2` " + mapped_type + ", `_sign` Int8() MATERIALIZED 1, "
"`_version` UInt64() MATERIALIZED 1) ENGINE = ReplacingMergeTree(_version)" + partition_policy + " ORDER BY (key, key2)");
"CREATE TABLE test_database.test_table_1 (`key` " + mapped_type + ", `key2` " + mapped_type +
MATERIALIZEMYSQL_TABLE_COLUMNS +
") ENGINE = ReplacingMergeTree(_version)" + partition_policy + " ORDER BY (key, key2)");
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1` (`key` " + test_type + " KEY UNIQUE KEY)", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` " + mapped_type + ", `_sign` Int8() MATERIALIZED 1, "
"`_version` UInt64() MATERIALIZED 1) ENGINE = ReplacingMergeTree(_version)" + partition_policy + " ORDER BY tuple(key)");
"CREATE TABLE test_database.test_table_1 (`key` " + mapped_type +
MATERIALIZEMYSQL_TABLE_COLUMNS +
") ENGINE = ReplacingMergeTree(_version)" + partition_policy + " ORDER BY tuple(key)");
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1` (`key` " + test_type + ", `key2` " + test_type + " UNIQUE KEY, PRIMARY KEY(`key`, `key2`))", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` " + mapped_type + ", `key2` " + mapped_type + ", `_sign` Int8() MATERIALIZED 1, "
"`_version` UInt64() MATERIALIZED 1) ENGINE = ReplacingMergeTree(_version)" + partition_policy + " ORDER BY (key, key2)");
"CREATE TABLE test_database.test_table_1 (`key` " + mapped_type + ", `key2` " + mapped_type +
MATERIALIZEMYSQL_TABLE_COLUMNS +
") ENGINE = ReplacingMergeTree(_version)" + partition_policy + " ORDER BY (key, key2)");
}
}
@ -165,23 +175,27 @@ TEST(MySQLCreateRewritten, RewrittenQueryWithPrimaryKey)
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1` (`key` int NOT NULL PRIMARY KEY) ENGINE=InnoDB DEFAULT CHARSET=utf8", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` Int32, `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1) ENGINE = ReplacingMergeTree(_version) "
"PARTITION BY intDiv(key, 4294967) ORDER BY tuple(key)");
"CREATE TABLE test_database.test_table_1 (`key` Int32" +
std::string(MATERIALIZEMYSQL_TABLE_COLUMNS) +
") ENGINE = ReplacingMergeTree(_version) PARTITION BY intDiv(key, 4294967) ORDER BY tuple(key)");
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1` (`key` int NOT NULL, PRIMARY KEY (`key`)) ENGINE=InnoDB DEFAULT CHARSET=utf8", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` Int32, `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1) ENGINE = ReplacingMergeTree(_version) "
"PARTITION BY intDiv(key, 4294967) ORDER BY tuple(key)");
"CREATE TABLE test_database.test_table_1 (`key` Int32" +
std::string(MATERIALIZEMYSQL_TABLE_COLUMNS) +
") ENGINE = ReplacingMergeTree(_version) PARTITION BY intDiv(key, 4294967) ORDER BY tuple(key)");
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1` (`key_1` int NOT NULL, key_2 INT NOT NULL, PRIMARY KEY (`key_1`, `key_2`)) ENGINE=InnoDB DEFAULT CHARSET=utf8", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key_1` Int32, `key_2` Int32, `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1) ENGINE = "
"ReplacingMergeTree(_version) PARTITION BY intDiv(key_1, 4294967) ORDER BY (key_1, key_2)");
"CREATE TABLE test_database.test_table_1 (`key_1` Int32, `key_2` Int32" +
std::string(MATERIALIZEMYSQL_TABLE_COLUMNS) +
") ENGINE = ReplacingMergeTree(_version) PARTITION BY intDiv(key_1, 4294967) ORDER BY (key_1, key_2)");
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1` (`key_1` BIGINT NOT NULL, key_2 INT NOT NULL, PRIMARY KEY (`key_1`, `key_2`)) ENGINE=InnoDB DEFAULT CHARSET=utf8", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key_1` Int64, `key_2` Int32, `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1) ENGINE = "
"ReplacingMergeTree(_version) PARTITION BY intDiv(key_2, 4294967) ORDER BY (key_1, key_2)");
"CREATE TABLE test_database.test_table_1 (`key_1` Int64, `key_2` Int32" +
std::string(MATERIALIZEMYSQL_TABLE_COLUMNS) +
") ENGINE = ReplacingMergeTree(_version) PARTITION BY intDiv(key_2, 4294967) ORDER BY (key_1, key_2)");
}
TEST(MySQLCreateRewritten, RewrittenQueryWithPrefixKey)
@ -191,7 +205,8 @@ TEST(MySQLCreateRewritten, RewrittenQueryWithPrefixKey)
EXPECT_EQ(queryToString(tryRewrittenCreateQuery(
"CREATE TABLE `test_database`.`test_table_1` (`key` int NOT NULL PRIMARY KEY, `prefix_key` varchar(200) NOT NULL, KEY prefix_key_index(prefix_key(2))) ENGINE=InnoDB DEFAULT CHARSET=utf8", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`key` Int32, `prefix_key` String, `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1) ENGINE = "
"CREATE TABLE test_database.test_table_1 (`key` Int32, `prefix_key` String" +
std::string(MATERIALIZEMYSQL_TABLE_COLUMNS) + ") ENGINE = "
"ReplacingMergeTree(_version) PARTITION BY intDiv(key, 4294967) ORDER BY (key, prefix_key)");
}
@ -204,6 +219,7 @@ TEST(MySQLCreateRewritten, UniqueKeysConvert)
"CREATE TABLE `test_database`.`test_table_1` (code varchar(255) CHARACTER SET utf8 COLLATE utf8_bin NOT NULL,name varchar(255) CHARACTER SET utf8 COLLATE utf8_bin NOT NULL,"
" id bigint NOT NULL AUTO_INCREMENT, tenant_id bigint NOT NULL, PRIMARY KEY (id), UNIQUE KEY code_id (code, tenant_id), UNIQUE KEY name_id (name, tenant_id))"
" ENGINE=InnoDB AUTO_INCREMENT=100 DEFAULT CHARSET=utf8 COLLATE=utf8_bin;", context_holder.context)),
"CREATE TABLE test_database.test_table_1 (`code` String, `name` String, `id` Int64, `tenant_id` Int64, `_sign` Int8() MATERIALIZED 1, `_version` UInt64() MATERIALIZED 1)"
" ENGINE = ReplacingMergeTree(_version) PARTITION BY intDiv(id, 18446744073709551) ORDER BY (code, name, tenant_id, id)");
"CREATE TABLE test_database.test_table_1 (`code` String, `name` String, `id` Int64, `tenant_id` Int64" +
std::string(MATERIALIZEMYSQL_TABLE_COLUMNS) +
") ENGINE = ReplacingMergeTree(_version) PARTITION BY intDiv(id, 18446744073709551) ORDER BY (code, name, tenant_id, id)");
}

20
src/Interpreters/QueryNormalizer.cpp
View File

@ -72,6 +72,12 @@ void QueryNormalizer::visit(ASTIdentifier & node, ASTPtr & ast, Data & data)
if (!IdentifierSemantic::getColumnName(node))
return;
if (data.settings.prefer_column_name_to_alias)
{
if (data.source_columns_set.find(node.name()) != data.source_columns_set.end())
return;
}
/// If it is an alias, but not a parent alias (for constructs like "SELECT column + 1 AS column").
auto it_alias = data.aliases.find(node.name());
if (it_alias != data.aliases.end() && current_alias != node.name())
@ -131,8 +137,20 @@ static bool needVisitChild(const ASTPtr & child)
void QueryNormalizer::visit(ASTSelectQuery & select, const ASTPtr &, Data & data)
{
for (auto & child : select.children)
if (needVisitChild(child))
{
if (child == select.groupBy() || child == select.orderBy() || child == select.having())
{
bool old_setting = data.settings.prefer_column_name_to_alias;
data.settings.prefer_column_name_to_alias = false;
visit(child, data);
data.settings.prefer_column_name_to_alias = old_setting;
}
else
{
if (needVisitChild(child))
visit(child, data);
}
}
/// If the WHERE clause or HAVING consists of a single alias, the reference must be replaced not only in children,
/// but also in where_expression and having_expression.

16
src/Interpreters/QueryNormalizer.h
View File

@ -4,6 +4,7 @@
#include <Parsers/IAST.h>
#include <Interpreters/Aliases.h>
#include <Core/Names.h>
namespace DB
{
@ -21,12 +22,15 @@ class QueryNormalizer
{
const UInt64 max_ast_depth;
const UInt64 max_expanded_ast_elements;
bool prefer_column_name_to_alias;
template <typename T>
ExtractedSettings(const T & settings)
: max_ast_depth(settings.max_ast_depth),
max_expanded_ast_elements(settings.max_expanded_ast_elements)
{}
: max_ast_depth(settings.max_ast_depth)
, max_expanded_ast_elements(settings.max_expanded_ast_elements)
, prefer_column_name_to_alias(settings.prefer_column_name_to_alias)
{
}
};
public:
@ -36,7 +40,8 @@ public:
using MapOfASTs = std::map<ASTPtr, ASTPtr>;
const Aliases & aliases;
const ExtractedSettings settings;
const NameSet & source_columns_set;
ExtractedSettings settings;
/// tmp data
size_t level;
@ -44,8 +49,9 @@ public:
SetOfASTs current_asts; /// vertices in the current call stack of this method
std::string current_alias; /// the alias referencing to the ancestor of ast (the deepest ancestor with aliases)
Data(const Aliases & aliases_, ExtractedSettings && settings_)
Data(const Aliases & aliases_, const NameSet & source_columns_set_, ExtractedSettings && settings_)
: aliases(aliases_)
, source_columns_set(source_columns_set_)
, settings(settings_)
, level(0)
{}

15
src/Interpreters/TreeRewriter.cpp
View File

@ -816,7 +816,14 @@ TreeRewriterResultPtr TreeRewriter::analyzeSelect(
/// Optimizes logical expressions.
LogicalExpressionsOptimizer(select_query, settings.optimize_min_equality_disjunction_chain_length.value).perform();
normalize(query, result.aliases, settings);
NameSet all_source_columns_set = source_columns_set;
if (table_join)
{
for (const auto & [name, _] : table_join->columns_from_joined_table)
all_source_columns_set.insert(name);
}
normalize(query, result.aliases, all_source_columns_set, settings);
/// Remove unneeded columns according to 'required_result_columns'.
/// Leave all selected columns in case of DISTINCT; columns that contain arrayJoin function inside.
@ -874,7 +881,7 @@ TreeRewriterResultPtr TreeRewriter::analyze(
TreeRewriterResult result(source_columns, storage, metadata_snapshot, false);
normalize(query, result.aliases, settings);
normalize(query, result.aliases, result.source_columns_set, settings);
/// Executing scalar subqueries. Column defaults could be a scalar subquery.
executeScalarSubqueries(query, context, 0, result.scalars, false);
@ -899,7 +906,7 @@ TreeRewriterResultPtr TreeRewriter::analyze(
return std::make_shared<const TreeRewriterResult>(result);
}
void TreeRewriter::normalize(ASTPtr & query, Aliases & aliases, const Settings & settings)
void TreeRewriter::normalize(ASTPtr & query, Aliases & aliases, const NameSet & source_columns_set, const Settings & settings)
{
CustomizeCountDistinctVisitor::Data data_count_distinct{settings.count_distinct_implementation};
CustomizeCountDistinctVisitor(data_count_distinct).visit(query);
@ -948,7 +955,7 @@ void TreeRewriter::normalize(ASTPtr & query, Aliases & aliases, const Settings &
FunctionNameNormalizer().visit(query.get());
/// Common subexpression elimination. Rewrite rules.
QueryNormalizer::Data normalizer_data(aliases, settings);
QueryNormalizer::Data normalizer_data(aliases, source_columns_set, settings);
QueryNormalizer(normalizer_data).visit(query);
}

2
src/Interpreters/TreeRewriter.h
View File

@ -119,7 +119,7 @@ public:
private:
const Context & context;
static void normalize(ASTPtr & query, Aliases & aliases, const Settings & settings);
static void normalize(ASTPtr & query, Aliases & aliases, const NameSet & source_columns_set, const Settings & settings);
};
}

2
src/Interpreters/tests/gtest_cycle_aliases.cpp
View File

@ -20,6 +20,6 @@ TEST(QueryNormalizer, SimpleCycleAlias)
aliases["b"] = parseQuery(parser, "a as b", 0, 0)->children[0];
Settings settings;
QueryNormalizer::Data normalizer_data(aliases, settings);
QueryNormalizer::Data normalizer_data(aliases, {}, settings);
EXPECT_THROW(QueryNormalizer(normalizer_data).visit(ast), Exception);
}

22
src/Server/NuKeeperTCPHandler.cpp → src/Server/KeeperTCPHandler.cpp
View File

@ -1,4 +1,4 @@
#include <Server/NuKeeperTCPHandler.h>
#include <Server/KeeperTCPHandler.h>
#if USE_NURAFT
@ -189,20 +189,20 @@ struct SocketInterruptablePollWrapper
#endif
};
NuKeeperTCPHandler::NuKeeperTCPHandler(IServer & server_, const Poco::Net::StreamSocket & socket_)
KeeperTCPHandler::KeeperTCPHandler(IServer & server_, const Poco::Net::StreamSocket & socket_)
: Poco::Net::TCPServerConnection(socket_)
, server(server_)
, log(&Poco::Logger::get("NuKeeperTCPHandler"))
, log(&Poco::Logger::get("KeeperTCPHandler"))
, global_context(server.context())
, nu_keeper_storage_dispatcher(global_context.getNuKeeperStorageDispatcher())
, operation_timeout(0, global_context.getConfigRef().getUInt("test_keeper_server.operation_timeout_ms", Coordination::DEFAULT_OPERATION_TIMEOUT_MS) * 1000)
, session_timeout(0, global_context.getConfigRef().getUInt("test_keeper_server.session_timeout_ms", Coordination::DEFAULT_SESSION_TIMEOUT_MS) * 1000)
, nu_keeper_storage_dispatcher(global_context.getKeeperStorageDispatcher())
, operation_timeout(0, global_context.getConfigRef().getUInt("keeper_server.operation_timeout_ms", Coordination::DEFAULT_OPERATION_TIMEOUT_MS) * 1000)
, session_timeout(0, global_context.getConfigRef().getUInt("keeper_server.session_timeout_ms", Coordination::DEFAULT_SESSION_TIMEOUT_MS) * 1000)
, poll_wrapper(std::make_unique<SocketInterruptablePollWrapper>(socket_))
, responses(std::make_unique<ThreadSafeResponseQueue>())
{
}
void NuKeeperTCPHandler::sendHandshake(bool has_leader)
void KeeperTCPHandler::sendHandshake(bool has_leader)
{
Coordination::write(Coordination::SERVER_HANDSHAKE_LENGTH, *out);
if (has_leader)
@ -217,12 +217,12 @@ void NuKeeperTCPHandler::sendHandshake(bool has_leader)
out->next();
}
void NuKeeperTCPHandler::run()
void KeeperTCPHandler::run()
{
runImpl();
}
Poco::Timespan NuKeeperTCPHandler::receiveHandshake()
Poco::Timespan KeeperTCPHandler::receiveHandshake()
{
int32_t handshake_length;
int32_t protocol_version;
@ -254,7 +254,7 @@ Poco::Timespan NuKeeperTCPHandler::receiveHandshake()
}
void NuKeeperTCPHandler::runImpl()
void KeeperTCPHandler::runImpl()
{
setThreadName("TstKprHandler");
ThreadStatus thread_status;
@ -393,7 +393,7 @@ void NuKeeperTCPHandler::runImpl()
}
}
std::pair<Coordination::OpNum, Coordination::XID> NuKeeperTCPHandler::receiveRequest()
std::pair<Coordination::OpNum, Coordination::XID> KeeperTCPHandler::receiveRequest()
{
int32_t length;
Coordination::read(length, *in);

8
src/Server/NuKeeperTCPHandler.h → src/Server/KeeperTCPHandler.h
View File

@ -13,7 +13,7 @@
#include <Interpreters/Context.h>
#include <Common/ZooKeeper/ZooKeeperCommon.h>
#include <Common/ZooKeeper/ZooKeeperConstants.h>
#include <Coordination/NuKeeperStorageDispatcher.h>
#include <Coordination/KeeperStorageDispatcher.h>
#include <IO/WriteBufferFromPocoSocket.h>
#include <IO/ReadBufferFromPocoSocket.h>
#include <Coordination/ThreadSafeQueue.h>
@ -29,16 +29,16 @@ using ThreadSafeResponseQueue = ThreadSafeQueue<Coordination::ZooKeeperResponseP
using ThreadSafeResponseQueuePtr = std::unique_ptr<ThreadSafeResponseQueue>;
class NuKeeperTCPHandler : public Poco::Net::TCPServerConnection
class KeeperTCPHandler : public Poco::Net::TCPServerConnection
{
public:
NuKeeperTCPHandler(IServer & server_, const Poco::Net::StreamSocket & socket_);
KeeperTCPHandler(IServer & server_, const Poco::Net::StreamSocket & socket_);
void run() override;
private:
IServer & server;
Poco::Logger * log;
Context global_context;
std::shared_ptr<NuKeeperStorageDispatcher> nu_keeper_storage_dispatcher;
std::shared_ptr<KeeperStorageDispatcher> nu_keeper_storage_dispatcher;
Poco::Timespan operation_timeout;
Poco::Timespan session_timeout;
int64_t session_id{-1};

12
src/Server/NuKeeperTCPHandlerFactory.h → src/Server/KeeperTCPHandlerFactory.h
View File

@ -1,6 +1,6 @@
#pragma once
#include <Server/NuKeeperTCPHandler.h>
#include <Server/KeeperTCPHandler.h>
#include <Poco/Net/TCPServerConnectionFactory.h>
#include <Poco/Net/NetException.h>
#include <common/logger_useful.h>
@ -9,7 +9,7 @@
namespace DB
{
class NuKeeperTCPHandlerFactory : public Poco::Net::TCPServerConnectionFactory
class KeeperTCPHandlerFactory : public Poco::Net::TCPServerConnectionFactory
{
private:
IServer & server;
@ -21,9 +21,9 @@ private:
void run() override {}
};
public:
NuKeeperTCPHandlerFactory(IServer & server_)
KeeperTCPHandlerFactory(IServer & server_)
: server(server_)
, log(&Poco::Logger::get("NuKeeperTCPHandlerFactory"))
, log(&Poco::Logger::get("KeeperTCPHandlerFactory"))
{
}
@ -31,8 +31,8 @@ public:
{
try
{
LOG_TRACE(log, "NuKeeper request. Address: {}", socket.peerAddress().toString());
return new NuKeeperTCPHandler(server, socket);
LOG_TRACE(log, "Keeper request. Address: {}", socket.peerAddress().toString());
return new KeeperTCPHandler(server, socket);
}
catch (const Poco::Net::NetException &)
{

2
src/Server/ya.make
View File

@ -22,10 +22,10 @@ SRCS(
HTTPHandler.cpp
HTTPHandlerFactory.cpp
InterserverIOHTTPHandler.cpp
KeeperTCPHandler.cpp
MySQLHandler.cpp
MySQLHandlerFactory.cpp
NotFoundHandler.cpp
NuKeeperTCPHandler.cpp
PostgreSQLHandler.cpp
PostgreSQLHandlerFactory.cpp
PrometheusMetricsWriter.cpp

3
src/Storages/MergeTree/MergeTreeSettings.h
View File

@ -83,6 +83,9 @@ struct Settings;
M(UInt64, replicated_max_parallel_fetches_for_host, DEFAULT_COUNT_OF_HTTP_CONNECTIONS_PER_ENDPOINT, "Limit parallel fetches from endpoint (actually pool size).", 0) \
M(UInt64, replicated_max_parallel_sends, 0, "Limit parallel sends.", 0) \
M(UInt64, replicated_max_parallel_sends_for_table, 0, "Limit parallel sends for one table.", 0) \
M(Seconds, replicated_fetches_http_connection_timeout, 0, "HTTP connection timeout for part fetch requests. Inherited from default profile `http_connection_timeout` if not set explicitly.", 0) \
M(Seconds, replicated_fetches_http_send_timeout, 0, "HTTP send timeout for part fetch requests. Inherited from default profile `http_send_timeout` if not set explicitly.", 0) \
M(Seconds, replicated_fetches_http_receive_timeout, 0, "HTTP receive timeout for fetch part requests. Inherited from default profile `http_receive_timeout` if not set explicitly.", 0) \
M(Bool, replicated_can_become_leader, true, "If true, Replicated tables replicas on this node will try to acquire leadership.", 0) \
M(Seconds, zookeeper_session_expiration_check_period, 60, "ZooKeeper session expiration check period, in seconds.", 0) \
M(Bool, detach_old_local_parts_when_cloning_replica, 1, "Do not remove old local parts when repairing lost replica.", 0) \

2
src/Storages/MergeTree/MergeTreeWhereOptimizer.cpp
View File

@ -39,7 +39,7 @@ MergeTreeWhereOptimizer::MergeTreeWhereOptimizer(
, queried_columns{queried_columns_}
, sorting_key_names{NameSet(
metadata_snapshot->getSortingKey().column_names.begin(), metadata_snapshot->getSortingKey().column_names.end())}
, block_with_constants{KeyCondition::getBlockWithConstants(query_info.query, query_info.syntax_analyzer_result, context)}
, block_with_constants{KeyCondition::getBlockWithConstants(query_info.query->clone(), query_info.syntax_analyzer_result, context)}
, log{log_}
, column_sizes{std::move(column_sizes_)}
{

1
src/Storages/PostgreSQL/PostgreSQLReplicaConnection.h
View File

@ -3,6 +3,7 @@
#include <Core/Types.h>
#include <Poco/Util/AbstractConfiguration.h>
#include "PostgreSQLConnectionPool.h"
#include <mutex>
namespace DB

30
src/Storages/StorageReplicatedMergeTree.cpp
View File

@ -1663,9 +1663,10 @@ bool StorageReplicatedMergeTree::tryExecutePartMutation(const StorageReplicatedM
if (source_part->name != source_part_name)
{
throw Exception("Part " + source_part_name + " is covered by " + source_part->name
+ " but should be mutated to " + entry.new_part_name + ". This is a bug.",
ErrorCodes::LOGICAL_ERROR);
LOG_WARNING(log, "Part " + source_part_name + " is covered by " + source_part->name
+ " but should be mutated to " + entry.new_part_name + ". "
+ "Possibly the mutation of this part is not needed and will be skipped. This shouldn't happen often.");
return false;
}
/// TODO - some better heuristic?
@ -2313,7 +2314,8 @@ bool StorageReplicatedMergeTree::executeReplaceRange(const LogEntry & entry)
{
String source_replica_path = zookeeper_path + "/replicas/" + part_desc->replica;
ReplicatedMergeTreeAddress address(getZooKeeper()->get(source_replica_path + "/host"));
auto timeouts = ConnectionTimeouts::getHTTPTimeouts(global_context);
auto timeouts = getFetchPartHTTPTimeouts(global_context);
auto [user, password] = global_context.getInterserverCredentials();
String interserver_scheme = global_context.getInterserverScheme();
@ -3246,6 +3248,23 @@ void StorageReplicatedMergeTree::exitLeaderElection()
leader_election = nullptr;
}
ConnectionTimeouts StorageReplicatedMergeTree::getFetchPartHTTPTimeouts(const Context & context)
{
auto timeouts = ConnectionTimeouts::getHTTPTimeouts(context);
auto settings = getSettings();
if (settings->replicated_fetches_http_connection_timeout.changed)
timeouts.connection_timeout = settings->replicated_fetches_http_connection_timeout;
if (settings->replicated_fetches_http_send_timeout.changed)
timeouts.send_timeout = settings->replicated_fetches_http_send_timeout;
if (settings->replicated_fetches_http_receive_timeout.changed)
timeouts.receive_timeout = settings->replicated_fetches_http_receive_timeout;
return timeouts;
}
bool StorageReplicatedMergeTree::checkReplicaHavePart(const String & replica, const String & part_name)
{
auto zookeeper = getZooKeeper();
@ -3661,7 +3680,8 @@ bool StorageReplicatedMergeTree::fetchPart(const String & part_name, const Stora
else
{
address.fromString(zookeeper->get(source_replica_path + "/host"));
timeouts = ConnectionTimeouts::getHTTPTimeouts(global_context);
timeouts = getFetchPartHTTPTimeouts(global_context);
user_password = global_context.getInterserverCredentials();
interserver_scheme = global_context.getInterserverScheme();

2
src/Storages/StorageReplicatedMergeTree.h
View File

@ -507,6 +507,8 @@ private:
/// Exchange parts.
ConnectionTimeouts getFetchPartHTTPTimeouts(const Context & context);
/** Returns an empty string if no one has a part.
*/
String findReplicaHavingPart(const String & part_name, bool active);

25
tests/clickhouse-test
View File

@ -1,5 +1,6 @@
#!/usr/bin/env python3
import shutil
import sys
import os
import os.path
@ -112,13 +113,14 @@ def get_db_engine(args, database_name):
return " ENGINE=" + args.db_engine
return "" # Will use default engine
def run_single_test(args, ext, server_logs_level, client_options, case_file, stdout_file, stderr_file):
def run_single_test(args, ext, server_logs_level, client_options, case_file, stdout_file, stderr_file, suite_tmp_dir):
# print(client_options)
start_time = datetime.now()
if args.database:
database = args.database
os.environ.setdefault("CLICKHOUSE_DATABASE", database)
os.environ.setdefault("CLICKHOUSE_TMP", suite_tmp_dir)
else:
# If --database is not specified, we will create temporary database with unique name
@ -136,6 +138,12 @@ def run_single_test(args, ext, server_logs_level, client_options, case_file, std
return clickhouse_proc_create, "", "Timeout creating database {} before test".format(database), total_time
os.environ["CLICKHOUSE_DATABASE"] = database
# Set temporary directory to match the randomly generated database,
# because .sh tests also use it for temporary files and we want to avoid
# collisions.
test_tmp_dir = os.path.join(suite_tmp_dir, database)
os.mkdir(test_tmp_dir)
os.environ.setdefault("CLICKHOUSE_TMP", test_tmp_dir)
# This is for .sh tests
os.environ["CLICKHOUSE_LOG_COMMENT"] = case_file
@ -170,7 +178,7 @@ def run_single_test(args, ext, server_logs_level, client_options, case_file, std
if need_drop_database:
clickhouse_proc_create = Popen(shlex.split(args.client), stdin=PIPE, stdout=PIPE, stderr=PIPE, universal_newlines=True)
seconds_left = max(args.timeout - (datetime.now() - start_time).total_seconds(), 10)
seconds_left = max(args.timeout - (datetime.now() - start_time).total_seconds(), 20)
try:
clickhouse_proc_create.communicate(("DROP DATABASE " + database), timeout=seconds_left)
except TimeoutExpired:
@ -185,6 +193,8 @@ def run_single_test(args, ext, server_logs_level, client_options, case_file, std
total_time = (datetime.now() - start_time).total_seconds()
return clickhouse_proc_create, "", "Timeout dropping database {} after test".format(database), total_time
shutil.rmtree(test_tmp_dir)
total_time = (datetime.now() - start_time).total_seconds()
# Normalize randomized database names in stdout, stderr files.
@ -207,7 +217,7 @@ def need_retry(stderr):
def get_processlist(args):
try:
clickhouse_proc = Popen(shlex.split(args.client), stdin=PIPE, stdout=PIPE, stderr=PIPE)
(stdout, _) = clickhouse_proc.communicate((b"SHOW PROCESSLIST FORMAT Vertical"), timeout=10)
(stdout, _) = clickhouse_proc.communicate((b"SHOW PROCESSLIST FORMAT Vertical"), timeout=20)
return False, stdout.decode('utf-8')
except Exception as ex:
print("Exception", ex)
@ -352,7 +362,7 @@ def run_tests_array(all_tests_with_params):
clickhouse_proc = Popen(shlex.split(args.client), stdin=PIPE, stdout=PIPE, stderr=PIPE, universal_newlines=True)
failed_to_check = False
try:
clickhouse_proc.communicate(("SELECT 'Running test {suite}/{case} from pid={pid}';".format(pid = os.getpid(), case = case, suite = suite)), timeout=10)
clickhouse_proc.communicate(("SELECT 'Running test {suite}/{case} from pid={pid}';".format(pid = os.getpid(), case = case, suite = suite)), timeout=20)
except:
failed_to_check = True
@ -367,7 +377,7 @@ def run_tests_array(all_tests_with_params):
stdout_file = os.path.join(suite_tmp_dir, name) + file_suffix + '.stdout'
stderr_file = os.path.join(suite_tmp_dir, name) + file_suffix + '.stderr'
proc, stdout, stderr, total_time = run_single_test(args, ext, server_logs_level, client_options, case_file, stdout_file, stderr_file)
proc, stdout, stderr, total_time = run_single_test(args, ext, server_logs_level, client_options, case_file, stdout_file, stderr_file, suite_tmp_dir)
if proc.returncode is None:
try:
@ -385,7 +395,7 @@ def run_tests_array(all_tests_with_params):
else:
counter = 1
while proc.returncode != 0 and need_retry(stderr):
proc, stdout, stderr, total_time = run_single_test(args, ext, server_logs_level, client_options, case_file, stdout_file, stderr_file)
proc, stdout, stderr, total_time = run_single_test(args, ext, server_logs_level, client_options, case_file, stdout_file, stderr_file, suite_tmp_dir)
sleep(2**counter)
counter += 1
if counter > 6:
@ -448,7 +458,7 @@ def run_tests_array(all_tests_with_params):
failures_chain += 1
status += MSG_FAIL
status += print_test_time(total_time)
status += " - Long test not marked as 'long'"
status += " - Test runs too long (> 30s). Make it faster."
else:
passed_total += 1
failures_chain = 0
@ -649,7 +659,6 @@ def main(args):
os.environ.setdefault("CLICKHOUSE_CONFIG", args.configserver)
if args.configclient:
os.environ.setdefault("CLICKHOUSE_CONFIG_CLIENT", args.configclient)
os.environ.setdefault("CLICKHOUSE_TMP", tmp_dir)
# Force to print server warnings in stderr
# Shell scripts could change logging level

4
tests/config/config.d/test_keeper_port.xml → tests/config/config.d/keeper_port.xml
View File

@ -1,5 +1,5 @@
<yandex>
<test_keeper_server>
<keeper_server>
<tcp_port>9181</tcp_port>
<server_id>1</server_id>
@ -17,5 +17,5 @@
<port>44444</port>
</server>
</raft_configuration>
</test_keeper_server>
</keeper_server>
</yandex>

2
tests/config/install.sh
View File

@ -29,7 +29,7 @@ ln -sf $SRC_PATH/config.d/graphite.xml $DEST_SERVER_PATH/config.d/
ln -sf $SRC_PATH/config.d/database_atomic.xml $DEST_SERVER_PATH/config.d/
ln -sf $SRC_PATH/config.d/max_concurrent_queries.xml $DEST_SERVER_PATH/config.d/
ln -sf $SRC_PATH/config.d/test_cluster_with_incorrect_pw.xml $DEST_SERVER_PATH/config.d/
ln -sf $SRC_PATH/config.d/test_keeper_port.xml $DEST_SERVER_PATH/config.d/
ln -sf $SRC_PATH/config.d/keeper_port.xml $DEST_SERVER_PATH/config.d/
ln -sf $SRC_PATH/config.d/logging_no_rotate.xml $DEST_SERVER_PATH/config.d/
ln -sf $SRC_PATH/config.d/tcp_with_proxy.xml $DEST_SERVER_PATH/config.d/
ln -sf $SRC_PATH/config.d/top_level_domains_lists.xml $DEST_SERVER_PATH/config.d/

0
tests/integration/test_testkeeper_back_to_back/__init__.py → tests/integration/test_keeper_back_to_back/__init__.py
View File

4
tests/integration/test_testkeeper_back_to_back/configs/enable_test_keeper.xml → tests/integration/test_keeper_back_to_back/configs/enable_keeper.xml
View File

@ -1,5 +1,5 @@
<yandex>
<test_keeper_server>
<keeper_server>
<tcp_port>9181</tcp_port>
<server_id>1</server_id>
<log_storage_path>/var/lib/clickhouse/coordination/log</log_storage_path>
@ -19,5 +19,5 @@
<port>44444</port>
</server>
</raft_configuration>
</test_keeper_server>
</keeper_server>
</yandex>

0
tests/integration/test_testkeeper_back_to_back/configs/logs_conf.xml → tests/integration/test_keeper_back_to_back/configs/logs_conf.xml
View File

0
tests/integration/test_testkeeper_back_to_back/configs/use_test_keeper.xml → tests/integration/test_keeper_back_to_back/configs/use_keeper.xml
View File

2
tests/integration/test_testkeeper_back_to_back/test.py → tests/integration/test_keeper_back_to_back/test.py
View File

@ -7,7 +7,7 @@ import time
from multiprocessing.dummy import Pool
cluster = ClickHouseCluster(__file__)
node = cluster.add_instance('node', main_configs=['configs/enable_test_keeper.xml', 'configs/logs_conf.xml'], with_zookeeper=True)
node = cluster.add_instance('node', main_configs=['configs/enable_keeper.xml', 'configs/logs_conf.xml'], with_zookeeper=True)
from kazoo.client import KazooClient, KazooState, KeeperState
def get_genuine_zk():

0
tests/integration/test_testkeeper_multinode_blocade_leader/__init__.py → tests/integration/test_keeper_multinode_blocade_leader/__init__.py
View File

4
tests/integration/test_testkeeper_multinode_blocade_leader/configs/enable_test_keeper1.xml → tests/integration/test_keeper_multinode_blocade_leader/configs/enable_keeper1.xml
View File

@ -1,5 +1,5 @@
<yandex>
<test_keeper_server>
<keeper_server>
<tcp_port>9181</tcp_port>
<server_id>1</server_id>
<log_storage_path>/var/lib/clickhouse/coordination/log</log_storage_path>
@ -37,5 +37,5 @@
<priority>1</priority>
</server>
</raft_configuration>
</test_keeper_server>
</keeper_server>
</yandex>

4
tests/integration/test_testkeeper_multinode_simple/configs/enable_test_keeper2.xml → tests/integration/test_keeper_multinode_blocade_leader/configs/enable_keeper2.xml
View File

@ -1,5 +1,5 @@
<yandex>
<test_keeper_server>
<keeper_server>
<tcp_port>9181</tcp_port>
<server_id>2</server_id>
<log_storage_path>/var/lib/clickhouse/coordination/log</log_storage_path>
@ -37,5 +37,5 @@
<priority>1</priority>
</server>
</raft_configuration>
</test_keeper_server>
</keeper_server>
</yandex>

4
tests/integration/test_testkeeper_multinode_simple/configs/enable_test_keeper3.xml → tests/integration/test_keeper_multinode_blocade_leader/configs/enable_keeper3.xml
View File

@ -1,5 +1,5 @@
<yandex>
<test_keeper_server>
<keeper_server>
<tcp_port>9181</tcp_port>
<server_id>3</server_id>
<log_storage_path>/var/lib/clickhouse/coordination/log</log_storage_path>
@ -37,5 +37,5 @@
<priority>1</priority>
</server>
</raft_configuration>
</test_keeper_server>
</keeper_server>
</yandex>

0
tests/integration/test_testkeeper_multinode_blocade_leader/configs/log_conf.xml → tests/integration/test_keeper_multinode_blocade_leader/configs/log_conf.xml
View File

0
tests/integration/test_testkeeper_multinode_blocade_leader/configs/use_test_keeper.xml → tests/integration/test_keeper_multinode_blocade_leader/configs/use_keeper.xml
View File

6
tests/integration/test_testkeeper_multinode_blocade_leader/test.py → tests/integration/test_keeper_multinode_blocade_leader/test.py
View File

@ -9,9 +9,9 @@ from helpers.network import PartitionManager
from helpers.test_tools import assert_eq_with_retry
cluster = ClickHouseCluster(__file__)
node1 = cluster.add_instance('node1', main_configs=['configs/enable_test_keeper1.xml', 'configs/log_conf.xml', 'configs/use_test_keeper.xml'], stay_alive=True)
node2 = cluster.add_instance('node2', main_configs=['configs/enable_test_keeper2.xml', 'configs/log_conf.xml', 'configs/use_test_keeper.xml'], stay_alive=True)
node3 = cluster.add_instance('node3', main_configs=['configs/enable_test_keeper3.xml', 'configs/log_conf.xml', 'configs/use_test_keeper.xml'], stay_alive=True)
node1 = cluster.add_instance('node1', main_configs=['configs/enable_keeper1.xml', 'configs/log_conf.xml', 'configs/use_keeper.xml'], stay_alive=True)
node2 = cluster.add_instance('node2', main_configs=['configs/enable_keeper2.xml', 'configs/log_conf.xml', 'configs/use_keeper.xml'], stay_alive=True)
node3 = cluster.add_instance('node3', main_configs=['configs/enable_keeper3.xml', 'configs/log_conf.xml', 'configs/use_keeper.xml'], stay_alive=True)
from kazoo.client import KazooClient, KazooState

0
tests/integration/test_testkeeper_multinode_simple/__init__.py → tests/integration/test_keeper_multinode_simple/__init__.py
View File

4
tests/integration/test_testkeeper_multinode_simple/configs/enable_test_keeper1.xml → tests/integration/test_keeper_multinode_simple/configs/enable_keeper1.xml
View File

@ -1,5 +1,5 @@
<yandex>
<test_keeper_server>
<keeper_server>
<tcp_port>9181</tcp_port>
<server_id>1</server_id>
<log_storage_path>/var/lib/clickhouse/coordination/log</log_storage_path>
@ -37,5 +37,5 @@
<priority>1</priority>
</server>
</raft_configuration>
</test_keeper_server>
</keeper_server>
</yandex>

4
tests/integration/test_testkeeper_multinode_blocade_leader/configs/enable_test_keeper2.xml → tests/integration/test_keeper_multinode_simple/configs/enable_keeper2.xml
View File

@ -1,5 +1,5 @@
<yandex>
<test_keeper_server>
<keeper_server>
<tcp_port>9181</tcp_port>
<server_id>2</server_id>
<log_storage_path>/var/lib/clickhouse/coordination/log</log_storage_path>
@ -37,5 +37,5 @@
<priority>1</priority>
</server>
</raft_configuration>
</test_keeper_server>
</keeper_server>
</yandex>

4
tests/integration/test_testkeeper_multinode_blocade_leader/configs/enable_test_keeper3.xml → tests/integration/test_keeper_multinode_simple/configs/enable_keeper3.xml
View File

@ -1,5 +1,5 @@
<yandex>
<test_keeper_server>
<keeper_server>
<tcp_port>9181</tcp_port>
<server_id>3</server_id>
<log_storage_path>/var/lib/clickhouse/coordination/log</log_storage_path>
@ -37,5 +37,5 @@
<priority>1</priority>
</server>
</raft_configuration>
</test_keeper_server>
</keeper_server>
</yandex>

0
tests/integration/test_testkeeper_multinode_simple/configs/log_conf.xml → tests/integration/test_keeper_multinode_simple/configs/log_conf.xml
View File

0
tests/integration/test_testkeeper_multinode_simple/configs/use_test_keeper.xml → tests/integration/test_keeper_multinode_simple/configs/use_keeper.xml
View File

6
tests/integration/test_testkeeper_multinode_simple/test.py → tests/integration/test_keeper_multinode_simple/test.py
View File

@ -9,9 +9,9 @@ from helpers.network import PartitionManager
from helpers.test_tools import assert_eq_with_retry
cluster = ClickHouseCluster(__file__)
node1 = cluster.add_instance('node1', main_configs=['configs/enable_test_keeper1.xml', 'configs/log_conf.xml', 'configs/use_test_keeper.xml'], stay_alive=True)
node2 = cluster.add_instance('node2', main_configs=['configs/enable_test_keeper2.xml', 'configs/log_conf.xml', 'configs/use_test_keeper.xml'], stay_alive=True)
node3 = cluster.add_instance('node3', main_configs=['configs/enable_test_keeper3.xml', 'configs/log_conf.xml', 'configs/use_test_keeper.xml'], stay_alive=True)
node1 = cluster.add_instance('node1', main_configs=['configs/enable_keeper1.xml', 'configs/log_conf.xml', 'configs/use_keeper.xml'], stay_alive=True)
node2 = cluster.add_instance('node2', main_configs=['configs/enable_keeper2.xml', 'configs/log_conf.xml', 'configs/use_keeper.xml'], stay_alive=True)
node3 = cluster.add_instance('node3', main_configs=['configs/enable_keeper3.xml', 'configs/log_conf.xml', 'configs/use_keeper.xml'], stay_alive=True)
from kazoo.client import KazooClient, KazooState

0
tests/integration/test_testkeeper_persistent_log/__init__.py → tests/integration/test_keeper_persistent_log/__init__.py
View File

4
tests/integration/test_testkeeper_persistent_log/configs/enable_test_keeper.xml → tests/integration/test_keeper_persistent_log/configs/enable_keeper.xml
View File

@ -1,5 +1,5 @@
<yandex>
<test_keeper_server>
<keeper_server>
<tcp_port>9181</tcp_port>
<server_id>1</server_id>
<log_storage_path>/var/lib/clickhouse/coordination/log</log_storage_path>
@ -18,5 +18,5 @@
<port>44444</port>
</server>
</raft_configuration>
</test_keeper_server>
</keeper_server>
</yandex>

0
tests/integration/test_testkeeper_persistent_log/configs/logs_conf.xml → tests/integration/test_keeper_persistent_log/configs/logs_conf.xml
View File

0
tests/integration/test_testkeeper_persistent_log/configs/use_test_keeper.xml → tests/integration/test_keeper_persistent_log/configs/use_keeper.xml
View File

2
tests/integration/test_testkeeper_persistent_log/test.py → tests/integration/test_keeper_persistent_log/test.py
View File

@ -10,7 +10,7 @@ from kazoo.client import KazooClient, KazooState
cluster = ClickHouseCluster(__file__)
node = cluster.add_instance('node', main_configs=['configs/enable_test_keeper.xml', 'configs/logs_conf.xml', 'configs/use_test_keeper.xml'], stay_alive=True)
node = cluster.add_instance('node', main_configs=['configs/enable_keeper.xml', 'configs/logs_conf.xml', 'configs/use_keeper.xml'], stay_alive=True)
def random_string(length):

Some files were not shown because too many files have changed in this diff Show More