PolarDB:PolarDB for MySQL 5.7 release notes

Category

Description

Release date

Bug fixes

• Fixed an issue where updating a NULL value to an empty string in a table that uses the REDUNDANT row format and has instant column addition enabled could cause a cluster crash.

• Fixed an occasional out-of-bounds read/write issue in ROW LOG when performing an online DDL operation on a table with virtual columns.

• Fixed an issue where the memory profile could not be enabled by restarting a 5.7 production cluster.

• Optimized the settings for variables related to global consistency (high-performance mode). The scope is changed from the system level to the connection level, allowing you to configure whether to enable the feature in the database proxy.

• Fixed an issue where a cluster would crash if a read-only node could not find a specified file when opening it during a DDL process.

• Fixed an issue where a "file not found" error occurred when a read-only node executed Show Binlog Events.

• Fixed an issue of potential index corruption for secondary indexes created on virtual columns.

• Fixed a compatibility issue where the database and table extension feature and the file space extension feature were incompatible because the loose_innodb_initialize_in_space_extend parameter could not be set to 0.

• Fixed an issue where the database and table restoration process could be interrupted when using the space extension feature without setting the value to 0.

• Fixed an issue where data could not be recovered using IMPORT TABLESPACE after the optimized BLOB write process was abnormally interrupted.

2025-06-20

Category

Description

Release date

New feature and performance optimization

Added the sm4_encrypt() and sm4_decrypt() functions to implement field-level access encryption using the SM4 algorithm.

2025-01-13

Bug fixes

• Fixed a system variable compatibility issue. No error is reported if duplicate items exist when you set a flagset variable, such as optimizer_switch.

• Fixed an issue of slow redo log synchronization when Innodb_flush_log_at_trx_commit=0, which could cause an SCC consistency timeout.

• Fixed a rare issue where a secondary node would crash during a reverse search on an index structure due to a lost synchronization offset.

• Fixed an issue where a crash during a non-atomic DDL process prevented the cluster from starting.

• Fixed an issue in earlier community versions where the JOIN ORDER in multi-table UPDATE statements was not correctly optimized.

Category

Description

Release date

New features and performance optimizations

• Added support for creating custom temporary tables on read-only nodes. Only the MEMORY and MyISAM engines are supported.

• Optimized the instant column addition feature to support adding columns to partitioned tables instantly.

• Optimized the buffer pool scale-out process and reduced lock contention with user threads to improve the scale-out experience.

2024-11-28

Bug fixes

• Fixed an issue where conflicts between advanced transaction splitting in the proxy and Fast Query Cache affected the result set.

• Fixed an issue in the new database and table restoration flow where the process was interrupted after multiple rounds of restoration on the same table.

• Fixed an issue where executing CHECK TABLE on a read-only (RO) node could cause a cluster crash.

• Fixed a rare issue where modifying the innodb_bulk_load_page_grained_redo_enable parameter within the time window during parallel DDL execution could cause incorrect redo writes for some pages.

• Fixed an issue where the auto-increment value of a table was rolled back when an INSERT ... ON DUPLICATE KEY UPDATE statement was executed after a read-only node was switched to a read/write node.

• Fixed an issue by disabling the index merge intersection access path by default for DML statements to improve performance and prevent deadlocks.

Category

Description

Release date

New features and performance optimizations

• Added support for physical migration from PolarDB 5.6 to PolarDB 5.7.

• Added a feature to apply TDE encryption to new tables after transparent data encryption (TDE) is enabled.

• Added fine-grained data modification tracking to improve the performance of global consistency (high-performance mode).

• Added support for user-defined in-memory temporary tables on read-only nodes. Currently, only the MEMORY and MyISAM engines are supported.

• Optimized the read-only variable max_digest_length to be dynamically configurable.

• Optimized the performance of simulated AIO.

• Optimized the outline library to allow an empty outline library name for each execution.

• Optimized PolarDB IO Threads to be dynamically adjustable without a restart.

2024-08-30

Bug fixes

• Fixed a lock conflict issue during dynamic buffer pool scale-in.

• Fixed issues related to purge pre-reading for the blob data type.

• Fixed an issue where the auto-increment value of a partitioned table was rolled back after an HA failover.

• Fixed a performance issue with large JSON serialization.

• Fixed an issue of insufficient precision for floating-point numbers in JSON.

• Fixed a deadlock issue caused by record locks that resulted from unnecessary unique key (UK) constraints.

• Fixed an issue where user threads on a read-only node waited too long for a free page when the buffer pool was full.

• Fixed an issue where the auto-increment value might be rolled back to 0 after all data in a table was deleted.

Category

Description

Release date

New features and performance optimizations

• Added the preemptive DDL feature to resolve failures in acquiring metadata locks (MDL) during DDL synchronization.

• Optimized thread priorities in PolarFileSystem to ensure timely processing of transaction requests and improve the DDL performance of PolarDB for MySQL.

• Optimized the page read performance on read-only nodes in PolarFileSystem.

2024-07-01

Bug fix

• Fixed an issue where the MDL system could become abnormal if the number of connections exceeded 65,536.

Category

Description

Release date

New features and performance optimizations

• Improved the error message for metadata lock (MDL) synchronization failures on read-only nodes.

• Added the polar_only_allow_discard_with_admin parameter. When enabled, this parameter prohibits users from performing DISCARD operations. The default value is ON. You can configure it as needed.

• Enhanced Fast Query Cache to include information about the number of returned rows. For cache-hit queries, the number of returned rows can be displayed in the audit log.

2024-05-27

Bug fixes

• Fixed an issue where system O&M access affected the slow_queries status metric.

• Fixed an issue where a trailing comma was required when setting loose_optimizer_switch using the SET_VAR hint.

Category

Description

Release date

New features and performance optimizations

• Added support for creating and deleting temporary tables within transactions after the global transaction identifier (GTID) feature is enabled.

• Added support for using the Variable-Setting Hint syntax in global consistency (high-performance mode).

• Optimized the performance of the thread pool when multi-threaded concurrent writing for binary logging is enabled.

2024-03-27

Bug fixes

• Fixed an incompatibility issue between transparent data encryption (TDE) and RDMA memory when a child process is created using fork.

• Fixed an issue where subsequent scale-out requests were blocked because a large user SQL statement occupied many pages for an extended period during a buffer pool scale-in operation.

• Fixed an issue where a node restart was required to recover from an auto-increment value rollback.

• Fixed a deadlock issue with DDL read-ahead in parallel DDL scenarios.

• Fixed an issue where new INSERT requests would encounter an "autoinc duplicate key" error after a DDL statement that modifies partitions was executed on a partitioned table.

Category

Description

Release date

New feature and performance optimization

Improved the speed at which the thread pool processes connection requests in low-concurrency scenarios.

2024-01-31

Bug fixes

• Fixed an issue where using the logical pre-read feature could cause a node to crash.

• Fixed a memory leak issue when using the logical pre-read feature.

• Fixed an issue in the new database and table restoration flow where threads could not exit because the disk was full.

• Fixed an issue in the new database and table restoration flow where adding a new read-only (RO) node failed because the registration offset of the RO node was too low.

Category

Description

Release date

Bug fixes

• Fixed an issue where the auto-increment value was rolled back after a high availability (HA) failover when autoinc_persist was enabled.

• Fixed an issue where the auto-increment value was rolled back after repeated HA failovers between a read/write node and a standby (or read-only) node.

• Fixed an issue where the standby node would lose the physical replication reconnection mechanism after a Degrade operation was performed on the read/write node.

2023-12-11

Category

Description

Release date

New features and performance optimizations

• Optimized the database and table restoration feature. Enhancements include faster restoration speed and support for task failure rollbacks.

• Optimized the log output feature by removing unnecessary error logs from fil_space_extend.

2023-11-23

Bug fixes

• Fixed an issue where dirty page flushing on a read/write (RW) node was blocked because the RW node did not terminate the replication relationship in special cases when the physical server of a read-only (RO) node failed.

• Fixed an issue where the auto-increment value was rolled back due to table cache eviction caused by an excessive number of tables after a high availability (HA) failover.

• Fixed an issue where the migration link was interrupted due to a PolarDB cluster restart during a one-click upgrade from RDS for MySQL to PolarDB for MySQL.

• Fixed an issue of abnormal checkpoints during the undo truncate process.

• Fixed an issue where a read-only node did not use the polarfs dir fast scan interface to connect to the read/write node upon startup.

Category

Description

Release date

Bug fixes

• Fixed a deadlock issue caused by a conflict between a logical pre-read cross-page operation and a Split Merge Operation (SMO) merge operation.

• Fixed an issue where temporary tables in the InnoDB engine could not be flushed.

• Fixed an issue where deleting the last CCL rule of the same SQL statement type that only specified a keyword without a schema and table could cause a node to crash.

2023-09-14

Category

Description

Release date

Bug fix

Fixed an issue where write performance decreased because the page cleaner thread flushed too many dirty pages when the dirty page level was low.

2023-08-28

Category

Description

Release date

Bug fix

Fixed an issue where SQL latency increased due to excessive pressure on the local disk after the audit log was enabled.

2023-08-02

Category

Description

Release date

New features and performance optimizations

• Added the innodb_polar_import_tablespace_strict_mode parameter. The default value is ON, which means an error is reported if the table schema contains a full-text index when you perform a database and table restoration operation.

• Added the session-level innodb_polar_scc parameter to control whether the current connection request requires strong consistency. When global consistency (high-performance mode) is enabled and innodb_polar_scc=true for the current connection, global consistency takes effect.

2023-07-23

Bug fixes

• Fixed an issue where regular users could delete the performance_schema database.

• Fixed an issue where the checkpoint LSN did not advance as expected.

• Fixed an issue where an error might occur when you perform a query on a read-only (RO) node while an ALTER TABLE ... IMPORT TABLESPACE operation is being performed on the read/write (RW) node.

• Fixed an issue where the secondary node did not trigger the reconnection mechanism when the primary node was shut down and restarted.

• Fixed an issue where the GROUP_CONCAT function returned an incorrect result due to an arithmetic overflow when the group_concat_max_len parameter was set to a large value.

• Fixed a mutex lock contention issue in high-concurrency range update scenarios after PolarTrans was dynamically enabled.

• Fixed an issue where a read-only (RO) node would crash if you connected to the database through the RO node to read data after the undo log recycle mechanism was triggered on the read/write (RW) node.

Category

Description

Release date

Bug fix

Fixed an issue where an abnormal internal counter in the thread pool could cause SQL requests to become stacked during concurrent recording of slow query logs.

2023-06-14

Category

Description

Release date

New features and performance optimizations

• Added support for recording DDL and LOCK TABLE statements to system tables.

• Made the thread pool feature compatible with CCL. After the thread pool feature is enabled, if a user connection is blocked by CCL, the connection is marked within the thread pool. The thread pool then creates extra threads to execute SQL statements on other connections.

• When you use Statement Outline, if a specified Index Hint does not exist, only a warning is reported instead of an error.

• Optimized show full tables from db_name like table_name.

• Added the Innodb_scc_timeout_count parameter for global consistency (high-performance mode) to record the number of timeouts during queries.

• Added the Innodb_scc_timeout_degrade_count parameter for global consistency (high-performance mode) to record the number of times a query timed out and was switched to an asynchronous query.

• Optimized the process of using Statement Outline by skipping digest hash calculation for statement execution when no outline is present.

• Improved the efficiency of truncating temporary tables.

• Added support for querying GTID information on read-only nodes.

• Optimized the internal lock granularity of PolarFS to significantly improve performance in some high-load scenarios.

2023-05-08

Bug fixes

• Fixed an issue where an event could not be deleted after it was created.

• Fixed an issue where a database connection could not be established due to a read timeout when physical replication was established on a read-only node.

• Fixed an issue where undo logs could not be truncated.

• Fixed a replication delay issue that occurred when a read-only node applied redo logs of a wide table.

Category

Description

Release date

Bug fixes

• Fixed an issue where enabling global consistency (high-performance mode) could cause a hot replica to crash unexpectedly.

• Fixed an issue where PolarTrans logs could not be properly reclaimed after a read/write node was restarted.

2023-03-25

Category

Description

Release date

Bug fixes

• Fixed an issue where the AUTO_INCREMENT value in a CREATE TABLE statement was lost after a failover between primary and secondary nodes.

• Fixed an issue where the innodb_buffer_pool_in_core_file parameter in the my.cnf file did not take effect at cluster startup.

• Fixed an issue of slow communication for the transparent data encryption (TDE) plugin on instances with large memory specifications.

2023-02-27

Category

Description

Release date

New features and performance optimizations

• Ignored the semicolon at the end of the SQL statement in the query parameter of the dbms_ccl.add_ccl_digest_rule, dbms_outln.add_index_outline, and dbms_outln.add_optimizer_outline built-in stored procedures.

• Statement Outline now supports all UNION clauses.

• Adjusted the timeout parameters related to physical replication connections to prevent connection issues when the physical server of the read/write node fails.

• Thread pool optimizations:

  • You can configure a list of high-priority users using the thread_pool_high_priority_users parameter. When the thread pool allocates threads, it prioritizes the requests of high-priority users.

  • Added a bypass mechanism. You can configure privileged connections using the bypass_thread_pool_ips parameter. When a user connects to the database from an IP address in bypass_thread_pool_ips, the connection can bypass the thread pool.

  • Added a slow query timeout handling mechanism. The thread pool ignores the number of threads in a slow query state when calculating the total number of threads. This ensures that the total number of threads is no longer limited by the number of slow query threads.

2023-01-06

Bug fix

Fixed an issue where database performance could slow down in long-running transaction scenarios.

Category

Description

Release date

New features and performance optimizations

• Added the Innodb_scc_timeout_count parameter for global consistency (high-performance mode) to record the number of timeouts during queries.

• Added the Innodb_scc_timeout_degrade_count parameter for global consistency (high-performance mode) to record the number of times a query timed out and was switched to an asynchronous query.

2022-11-11

Bug fixes

• Fixed an issue where the physical replication module on a read/write node consumed extra memory.

• Fixed an issue where undo logs could not be purged in special scenarios.

• Fixed an issue where control health checks failed in special scenarios to improve system stability in high-concurrency scenarios.

Category

Description

Release date

Bug fix

Fixed a crash issue on read-only nodes caused by a filename hash conflict.

2022-09-17

Category

Description

Release date

New features and performance optimizations

• Added the global consistency (high-performance mode) feature. After this feature is enabled, read-only nodes of PolarDB can provide strong consistency read services. For more information, see Overview.

• Optimized the DDL read-ahead feature to improve the efficiency of index creation.

• Optimized DDL sorting. Two-way merging is replaced with multi-way merging to reduce I/O operations and improve the performance of DDL index addition.

• Improved the performance of the thread pool under high concurrency.

2022-08-30

Bug fix

Fixed an issue where deleting a database was not recorded in the slow log.

Category

Description

Release date

New feature and performance optimization

Added support for the DDL read-ahead feature to improve the efficiency of index creation.

2022-06-17

Bug fixes

• Optimized the selection of the ordering index path, which did not consider the cost of table access by index.

• Fixed an issue where the LRU list cleanup was blocked, preventing pages from being released when the buffer pool had too many dirty pages and flushing was slow.

Category

Description

Release date

New feature and performance optimization

Supported millisecond-level latency for physical replication between primary and secondary nodes.

2022-05-20

Bug fixes

• Fixed an issue of excessive physical replication delay under low cluster loads.

• Optimized the selection of the ordering index to prioritize index paths that can both use index conditions and provide ordering.

• Fixed an issue where the optimizer did not choose a better index range path for some GROUP BY statements.

• Fixed an issue where a checkpoint was not triggered when a read-only node registered with a read/write node.

Category

Description

Release date

New feature and performance optimization

Changed OF, COMMIT_ON_SUCCESS, ROLLBACK_ON_FAIL, and TARGET_AFFECT_ROW from reserved keywords to non-reserved keywords in SQL parsing.

2022-04-07

Category

Description

Release date

New features and performance optimizations

• Added support for the hot row optimization feature. For more information, see Hot row optimization.

• Added support for the hot replica feature to further improve high availability efficiency.

2022-01-14

Category

Description

Release date

Bug fix

Optimized the policy for manually triggering checkpoints.

2021-12-14

Category

Description

Release date

New feature and performance optimization

Added the Fast Query Cache feature. For more information, see Fast Query Cache.

2021-11-17

Category

Description

Release date

Bug fixes

• Provided an optimizer hint that lets users control whether to force a view merge. For more information, see the official MySQL documentation.

• Added support for the SELECT FOR UPDATE WAIT N syntax. N indicates the timeout period in seconds and applies only to a single row lock. If a query needs to lock multiple rows, the system does not accumulate the time for multiple rows but only checks the timeout for a single row. If the wait time exceeds N seconds, the system returns a lock wait timeout error: Lock wait timeout exceeded; try restarting transaction.

2021-10-19

Category

Description

Release date

New feature and performance optimization

A read-only node in the secondary cluster of a GDN can be switched to a primary node using the alter polar to slave command. This implements high availability for the secondary cluster.

2021-09-10

Category

Description

Release date

New features and performance optimizations

• Supported multiple connections between the primary cluster and synchronization nodes in a GDN to improve the throughput of physical replication.

• Added support for obtaining binary logs from read-only nodes. For more information, see Remotely obtain and parse binary logs of PolarDB for MySQL.

2021-08-27

Bug fixes

• Fixed an issue where rules for specified databases and tables in Statement Concurrency Control could not be matched.

• Accelerated the application of redo logs on read-only nodes and secondary clusters to improve the synchronization efficiency of the primary node.

Category

Description

Release date

New feature and performance optimization

Added the polar_replica_work_on_nonblock_mdl_mode parameter. When this parameter is enabled, uncommitted transactions at the RU/RC isolation level on a read-only node no longer block DDL operations on the primary node. In this case, the transactional properties of table definitions are no longer guaranteed on the read-only node.

2021-07-08

Category

Description

Release date

New feature and performance optimization

Added support for the SELECT FOR UPDATE/SHARE SKIP LOCKED/NOWAIT syntax for compatibility with MySQL 8.0. For more information, see MySQL 8.0.

2021-06-15

Bug fix

Added support for flushing InnoDB temporary table data to disk on read-only nodes.

Category

Description

Release date

Bug fixes

• Added support for Quick Launch of the storage engine in multi-table scenarios.

• Fixed a database crash issue caused by rolling back an UPDATE operation on a virtual column.

2021-05-13

Category

Description

Release date

New feature and performance optimization

Specifies whether to immediately execute the checkpoint policy on the primary node when a read-only node establishes a replication relationship.

2021-04-19

Bug fixes

• Changed KICKOUT to a non-reserved keyword.

• Fixed an issue where encryption information in a TDE-encrypted table might be lost after a primary/secondary failover, leading to decryption failure.

Category

Description

Release date

New features and performance optimizations

• Added the parallel DDL feature to improve DDL performance. For more information, see Parallel DDL.

• Added the innodb_buffer_pool_in_core_file parameter to remove the buffer pool from the core file.

• Added support for the following keywords to Index Hints, which allow the optimizer to use or ignore specified indexes when processing queries:

  • GROUP_INDEX or NO_GROUP_INDEX: Use or ignore the specified index for an index scan with a GROUP BY operation.

  • INDEX or NO_INDEX: Force the server to use or ignore the specified index.

  • JOIN_INDEX or NO_JOIN_INDEX: Force MySQL to use or ignore the specified index for any access method.

• Added support for the following keywords to Join Order Hint, which allow the optimizer to select an appropriate table join order:

  • JOIN_FIXED_ORDER: Force the optimizer to join tables in the order they appear in the FROM clause.

  • JOIN_ORDER: Instruct the optimizer to join tables in the specified order. This hint applies to named tables. The optimizer can place unnamed tables anywhere in the join order, including between specified tables.

  • JOIN_PREFIX: Instruct the optimizer to join tables in the specified order for the first few tables of the join execution plan. This hint applies to named tables. The optimizer places all other tables after the named tables.

  • JOIN_SUFFIX: Instruct the optimizer to join tables in the specified order for the last few tables of the join execution plan. This hint applies to named tables. The optimizer places all other tables before the named tables.

2021-03-10

Bug fixes

• Partitioned the token number comparison table and reserved some tokens to fix an issue where the digest hash value of the same statement was inconsistent after a new token was inserted.

• Fixed an upgrade/downgrade failure caused by a read-only node failing to register with the primary node during the upgrade/downgrade process.

• Fixed an issue where creating or deleting a temporary table in a stored procedure caused the cluster to exit abnormally when the session_track_temporary_tables system variable was enabled.

Category

Description

Release date

Bug fix

Fixed an issue where the result of SHOW PROCESSLIST was incorrect in some scenarios.

2021-02-22

Category

Description

Release date

Bug fix

Fixed an occasional issue where error code 1032 (Can't find record in TABLE) occurred when executing a query that used Index Merge optimization on a read-only node.

2021-02-10

Category

Description

Release date

Bug fixes

• Fixed an issue of abnormal dirty page flushing on the read/write node.

• Fixed an issue where the primary zone could not be changed on a cluster that had undergone a primary/secondary failover.

• Fixed a database crash issue caused by an incorrect call to the Thread Descriptor (THD) when executing SHOW PROCESSLIST.

2021-02-02

Category

Description

Release date

New features and performance optimizations

• Added support for the Statement Queue feature. For more information, see Statement Queue.

• Added support for the Instant add column feature. For more information, see Instantly add a column at the end of a table.

• Added support for the Returning feature. For more information, see Returning.

• The default value of the innodb_adaptive_hash_index parameter is changed from ON to OFF.

2021-01-29

Bug fix

Fixed an issue where system tables were lost when upgrading the PolarDB kernel version.

Category

Description

Release date

Bug fixes

• Fixed an issue where a read-only node would crash when a query was performed on it after the key was modified on the primary node.

• Fixed an issue where a read-only node would crash when a partitioned table was analyzed on it.

• Fixed an issue where the new primary node would crash due to inconsistent tablespaces after a primary/secondary failover.

2020-12-31

Category

Description

Release date

Bug fixes

• Fixed an issue where data could not be inserted into a temporary table after a primary/secondary failover.

• Fixed an issue where a read-only node became unavailable when the temporary tablespace was extended on it.

• Fixed an issue where Simulated AIO did not work as expected.

2020-11-17

Category

Description

Release date

New feature and performance optimization

Added support for the transparent data encryption (TDE) feature.

2020-10-21

Bug fixes

• Fixed an issue of inconsistent performance statistics between a read-only node and the primary node.

• Fixed an issue where an error message was displayed in the result of SHOW PROCESSLIST.

Category

Description

Release date

Bug fixes

• Fixed an issue where .IDB files could not be extended after migration from RDS for MySQL 5.7.

• Prohibited the execution of CF_STATUS_COMMAND-related commands during a primary/secondary failover to ensure normal operation.

• Fixed an issue where the primary node became unavailable due to page contention between the background thread for updating statistics and the Truncate logic.

2020-08-28

Category

Description

Release date

Bug fixes

• Fixed an issue where MaxScale could not be accessed because some PolarDB-specific commands were missing their corresponding command numbers.

• Fixed an issue where a cluster created by restoring from the recycle bin became unavailable because the tablespace could not be found.

• Fixed an issue where a file was not closed when the promote replica command was executed, causing a crash when unmounting the PFS file.

2020-07-30