PS-9422 Release tasks ticket for PS 8.0.39 #5446
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…rade When the Router is bootstrapped to the fresh directory it tries to create a new account in the `routers` metadata table. This table has a unique constraint UNIQUE KEY `address` (`address`,`router_name`). If there already is a record with that address and router_name pair the insert fails due to the contraint mentioned before. If no --force parameter is given the Router fails to bootstrap with an error as expected. However if the --force parameter is provided and there already is a record matching (`address`,`router_name`) pair, the Router queries for the router_id to re-use it. The problem is that the query that does it is invalid, it only has "WHERE router_name=?" clause, ignoring the address part. This means that it can return multiple rows for any router with a given name for various hosts/addresses. Then the id of the random (first from the list of the rows) is taken. This can lead to multiple Routers on various host using the same account, updating the attibutes etc. This patch fixes the query retrieving the router_id so that it matches the unique constrain (address, router_name). Change-Id: I8f8f7227e9597cbd5a917d8ef068c7b6214c5288
The router does not allow the router_id higher than 999999 in order to be able to create an user name containing mysql_routerXX_YY where XX is the router_id and YY is random 12-character suffix. This patch changes that so that the highest possible router_id is supported (2^32-1). To allow that the YY part is reduced to 7 characters. Change-Id: I7db486d8849b01a790e07f265941c9fd89d27309
Post push fix. Also handle rmdir in PosixAsyncFile::rmrfReq. Change-Id: I56ff2274b0ccd34e0e538ab499834d09288e9e84
Followup patch: The upgrade to the latest protobuf library has made some changes that are incompatible with the MSVC PGO compiler and linker options on Windows. These incompatibilities are addressed by excluding the protoc executable and Abseil/protobuf DLLs from PGO. Change-Id: I587e2c8729bd07afb6508c748ddf73eca881d2a7
Followup patch: Suppress linker warnings generated as a consequence of combining CMAKE_WINDOWS_EXPORT_ALL_SYMBOLS with __declspec(dllexport) Change-Id: I7c6b63e6db5c327b582eb8994ee6ef3a64356abc
… error Loading of some of the libraries may fail: ELF load command alignment not page-aligned Linux aarch64 platform seems to be two flavours: 4k and 64k page size. 4k systems can load binaries with 4k and 64k page size, while 64k systems can load 64k binaries only. Hence, hard code argument for patchelf to --page-size=65536 to produce most portable binary. Change-Id: Ifbde485f61e0ce3814393b21f027b8be2f48d547
Problem: mysqldump not sanitizing the version string obtained from server which may lead to injecting malicious commands to the output. Fix: added function sanitizing the version string by cutting off illegal part and issuing warning. Test: check the server version in the output with and without injected payload. Change-Id: I1f19e1c90bdb8d444285e427092face3bb16da01
Problem: mysqldump_bugs.test failing on release builds Fix: replace --source include/have_debug.inc by mysql_have_debug.inc Change-Id: I151b69dc0da1c5d36714b453ba51349e5057564b
…e failing Symptom: Test failing on Linux when run with --mem and --no-skip. Cause: Warning in error log due to lack of O_DIRECT support in tmpfs; --no-skip causes have_odirect.inc to have no effect. Solution: have_odirect.inc added to excludenoskip.list will cause the test (and all other tests including it) to be skipped when O_DIRECT is not supported, regardless of --no-skip. Change-Id: I2ef0e9939567f03db93dabd264c4590ed7c6c759
X Plugin documentation https://dev.mysql.com/doc/dev/mysql-server/latest/mysqlx_protocol_notices.html mentions "SLAVE" in a few places. Change ====== - s/slave/replica/ - s/master/replication source/ Change-Id: I672d2ecda2556daac0b3a61b903ab484ae9a8c50
Fixed to use strtoul() instead of std::stoul. Change-Id: Ic108b8e559c20cb88d4db3963cf3ca147c17189e
Problem: SET offline_mode=on could hang and make system unresponsive Root cause: With many connections SET offline_mode could end up writing so much to the TP notify descriptor that send() would block. If this happens when the TP waiting_thread is waiting for a mutex held by SET offline_mode a deadlock would result. Solution: Add a flag which is set when the notify descriptor is written to, and cleared by the waiting thread when it drains (reads all available bytes) the notify descriptor. As long as the flag is set new kill notifcations do not actually write to the notify descriptor. That way the capacity of the notify descriptor is not exceeded. Additional writes before the notify descriptor has been drained are redundant since the waiting_thread iterates over all connections in the thread group to identify those that are killed. Increase the number of connections allowed by mysqltest, to be able to create the required number of connections. Change-Id: I8c84c0b9c9e5a4e2312fd6758e07eebc18e7c7a4
Problem: SET offline_mode=on could hang and make system unresponsive Root cause: With many connections SET offline_mode could end up writing so much to the TP notify descriptor that send() would block. If this happens when the TP waiting_thread is waiting for a mutex held by SET offline_mode a deadlock would result. Solution: Add a flag which is set when the notify descriptor is written to, and cleared by the waiting thread when it drains (reads all available bytes) the notify descriptor. As long as the flag is set new kill notifcations do not actually write to the notify descriptor. That way the capacity of the notify descriptor is not exceeded. Additional writes before the notify descriptor has been drained are redundant since the waiting_thread iterates over all connections in the thread group to identify those that are killed. Increase the number of connections allowed by mysqltest, to be able to create the required number of connections. Change-Id: I8c84c0b9c9e5a4e2312fd6758e07eebc18e7c7a4
Change-Id: I733cc835b1617920e8700d7ce0fc3fb979069963
Change-Id: I0baedae7b84b8dd4a5800c14f3e715052071e65e
Change-Id: Ica35d9293a67a51c803f67381a933124d7470564
Post push fix. Remove compiler warnings when building for MySQL 5.7, MySQL Cluster 7.5 and 7.6. Change-Id: I5e8decbf776f8a1f4171dd3beae257235418756f
Change-Id: Iab626a487ddf444979d812f50a6ff9e8405c2882
Change-Id: I3adc25f4e53201b02dac7b68f7740472f6f7119b
Post push fix. Remove compiler warnings when building for MySQL 5.7, MySQL Cluster 7.5 and 7.6. Change-Id: I5e8decbf776f8a1f4171dd3beae257235418756f
…n_cols thread 1406284440 Issue: An assert in storage/innobase/include/dict0dict.ic triggers, indicating an invalid column index, while the column index is in fact correct. Background: The column index provided to dict_table_get_index_on_first_col is an index among all columns, while the n_cols field of dict_table_t only countains the number of non-virtual columns. Thus, the assert could be triggered with correct arguments. Fix: Changing the assert to compare against the n_t_cols field, the total number of columns. Change-Id: I90cc8c07d428c9117ca7873e9f53752d50f800e6
Approved-by: Balasubramanian Kandasamy <balasubramanian.kandasamy@oracle.com>
Approved-by: Balasubramanian Kandasamy <balasubramanian.kandasamy@oracle.com>
Approved-by: Balasubramanian Kandasamy <balasubramanian.kandasamy@oracle.com>
Approved-by: Balasubramanian Kandasamy <balasubramanian.kandasamy@oracle.com>
Approved-by: Balasubramanian Kandasamy <balasubramanian.kandasamy@oracle.com>
Change-Id: I2dfd8d9b84724dc820a098fcbd65a4d9f2d49a1c
Change-Id: I948602390fd4848b45dc99c528eab87b912db706
https://perconadev.atlassian.net/browse/PS-9293 *** New release just added SQL echo of prepared statement. Test results re-recorded
Merge fixes for PS-9144 and PS-9214 into 8.0 release tree.
After Upstream's fix for bug#35952353 "SELECT COUNT(*) degraded performance on 8.0 compared to 5.7" (mysql/mysql-server@ffad6cf) MySQL stopped using parallel read for COUNT(*) evaluation when secondary index scan is used. As result innodb.parallel_read_kill test started to fail, as it employed COUNT(*) that used secondary index scan to test InnoDB parallel read killability. The parallel read code is simply no longer invoked causing DEBUG_SYNC time out. This patch fixes the failure by forcing the test to use COUNT(*) that uses primary key scan, which is still processed using parallel read, instead.
PS-9293, PS-9328: Fix innodb.parallel_read_kill test failures (8.0 version).
https://perconadev.atlassian.net/browse/PS-9286 extra/libkmip head is at baf6832 with activate operation implementation. keyring_kmip/backend/backend.cc updateed to execute op_activate after key registration.
…lect Sys_var_tz::session_value_ptr) Reverting Percona Server's version of the fix which is no longer necessary. Upstream fixed this issue in a different, more complex, but also probably more clean way in their fix for bug#31168097 "SIG 6 ASSERTION `0' IN TIME_ZONE_UTC::GET_NAME AT TZTIME.CC:1182". See: mysql/mysql-server@1f8c02e The test case for this issue is kept around.
PS-9293, PS-9328: Revert of fix for PS-7221: Bug #100402 (crash on se… (8.0 version).
…il.cc:13097:fil_system->shard_by_id(id)->mutex_owned(). Reverted Percona Server's fix for PS-7666 as it became redundant after Upstream's has fixed the problem in a different way. See the Upstream's fix for bug#32819101 "ASSERTION FAILURE: FIL0FIL.CC: 12089:FIL_SYSTEM->SHARD_BY_ID(ID)->MUTEX_OWNED()" for details: 4a2cf86 Removed Percona's test case as well, as Upstream fix comes with one.
PS-9293, PS-9328: Revert of fix for PS-7666: Assertion failure: fil0f… (8.0 version)
https://perconadev.atlassian.net/browse/PS-9378 This patch was originally created for the 8.4 branch where the only available binlog transaction dependency tracking mode is 'WRITESET'. However, it also makes sense to backport this to 8.0 as the same code paths are used when user explicitly set the 'binlog_transaction_dependency_tracking' system variable to 'WRITESET' (by default in 8.0 series, this variable is set to 'COMMIT_ORDER'). Cherry-picked PS-9302 "Improve performance for binlog_transaction_dependency_tracking=WRITESET" (https://perconadev.atlassian.net/browse/PS-9302) from 8.4. Problem: Comparing to 8.0 where the default value of binlog_transaction_dependency_tracking was COMMIT_ORDER, 8.4 introduces a visible write performance drop. Cause: 8.4 uses WRITESET dependency tracking. For this tracking we maintain a map of binlog_transaction_dependency_history_size for row_id to newest transaction sequence_number which modified a given row. Every new transaction needs to check its dependency by examining the map (find), which is done with logarithmic complexity. Solution: Change std::map to std::unordered_map. This allows us to find a row dependency in constant on average complexity.
…et_perf_backport PS-9378: Backport PS-9302 (WRITESET perf improvements) to 8.0 (8.0)
…n rocksdb_stress.drop_cf_stress tests. Disabled two tests which create too much noise in Jenkins and which we could not/do not plan to fix immediately. binlog.binlog_mysqlbinlog_4g_start_position test requires too much memory/ space so often fails in Jenkins even when run alone. rocksdb_stress.drop_cf_stress test executes concurrently mix of DDL and often fails due to race.
…nent (#5407) https://perconadev.atlassian.net/browse/PS-9379 Fix of problem with -latomic in uuid_vx component of some platofrms. CMakeLists.txtx for the component now check code sample compilation and adds "-latomic" if required. Note! On RHEL platforms the package "gcc-toolset-12-libatomic-devel" must be installed as compile dependency.
PS-9293, PS-9328: Disabled binlog.binlog_mysqlbinlog_4g_start_position (8.0 version)
… has 10K tables or more https://perconadev.atlassian.net/browse/PS-9306 Upstream fixed the problem in 8.0.39 release but no testcase is attached to the bug fix. This commit introduces a mtr testcase. It creates 8192 tables, restarts the server and it should come up
… and leaves orphan truncate log files https://perconadev.atlassian.net/browse/PS-9322 Backported Oracle fix for Bug #35784192 "tablespace is missing for table innodb_undo_002" (commit mysql/mysql-server@bda098a and mysql/mysql-server@97f4285) from 8.4 to 8.0. Bug#35784192 tablespace is missing for table innodb_undo_002 This bug reports following two problems: Querying the information schema reports undo tablespace is missing - while an undo file is present. Purge threads don't remove the undo_space_number_trunc.log file. While the first problem is already fixed through Bug#32104924, Bug#32654667 This bug fix aims to address the second problem. The purge thread attempts to lock dictionary tables, while truncating the undo tablespace. This lock can't be acquired in a read-only MySQL instance, except when the thread has some special flags associated with it. In this case, we use the skip_readonly_check member of the thread descriptor struct, to signal that the readonly check should not be applied to the purge thread. This is done via the set_skip_readonly_check() method. Change-Id: Iea6255ff6654752db8658bdd422cc3e58e3a8a97 Bug#35784192 tablespace is missing for table innodb_undo_002 [postfix] Remove MySQL versions number from MTR test. Change-Id: I3973435edbbc8bf442846c617ade4b5b824bb8c2
…runcation_super_read_only PS-9322 fix: With super_read_only=1, undo truncation cannot update DD and leaves orphan truncate log files (8.0)
PS-9306: MySQL 8.0.38, 8.4.1 and 9.0.0 crashes on restart if database…
RM-1420 PS-8.0.39-30
RM-1420 PS-8.0.39-30
|
@dlenev Could you check the conflict? |
|
@adivinho Resolved. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.