AliSQL is a MySQL branch that is developed by Alibaba Cloud. AliSQL provides all the features of the MySQL Community Edition. AliSQL also provides some similar features that you can find in the MySQL Enterprise Edition. These similar features include enterprise-grade backup and restoration, thread pool, and parallel query. ApsaraDB RDS for MySQL with AliSQL provides a wide range of features, such as security, backup and restoration, monitoring, performance optimization, and read-only instances. This topic describes the release notes of minor AliSQL versions.

Note
  • This topic lists all minor AliSQL versions. When a minor AliSQL version is being maintained, it may be unavailable in the ApsaraDB RDS console. You can select only the minor AliSQL versions that are displayed in the ApsaraDB RDS console.
  • For more information about the dedicated proxy versions of ApsaraDB RDS for MySQL, see Release notes for dedicated proxy versions.

MySQL 8.0 on RDS Basic Edition or High-availability Edition

Minor AliSQL version Description
20201231
  • New features:
    • The feature changes in open source MySQL 8.0.19, MySQL 8.0.20, MySQL 8.0.21, and MySQL 8.0.22 are incorporated. For more information, see Changes in MySQL 8.0.19, Changes in MySQL 8.0.20, Changes in MySQL 8.0.21, and Changes in MySQL 8.0.22.
    • The feature that is used to adjust the size of the buffer pool is supported.
    • The feature that is used to asynchronously read multiple blocks at a time is supported.
    • The feature that is used to place the UPDATE statement and the DELETE statement into a concurrency control (CCL) queue is supported. These statements must contain a primary key or a unique key.
    • The logic that is used to scan the linked list of the buffer pool and obtain idle pages is optimized. The linked list is managed based on the least recently used (LRU) algorithm.
    • Three session trackers are added. These trackers are session_track_prepare_statements, session_track_temporary_tables, and session_track_user_variables.
    • The feature that is used to record the time that ApsaraDB RDS needs to wait before it can process a TCP-based write request is supported. The time information is recorded in the P_S.events_statements_summary_by_digest_supplement table.
    • The feature that is used to record DDL operations in the redo log is supported. This feature is supported for DDL operations that are performed by using the in-place method.
  • Bugs fixed:
    • The bug that causes a stack overflow is fixed by introducing a protection mechanism.
    • The bug that prevents the mysqld program from starting is fixed. This bug is triggered if the Binlog in Redo feature and X-Engine are both enabled. The Binlog in Redo feature is controlled by a system variable named persist_binlog_to_redo.
    • The bug that causes memory leaks is fixed. This bug is triggered if the pthread_getattr_np function is invoked.
    • The bug that causes ApsaraDB RDS to delete cached pages is fixed. The cached pages are generated by the performance agent feature.
    • The bug in database outlines is fixed.
    • The bug that causes ApsaraDB RDS to reset the session tracker in use is fixed. This bug is triggered if ApsaraDB RDS encounters errors when it executes statements.
    • The ccl_wait_timeout variable is renamed.
    • The bug that prevents ApsaraDB RDS from deleting socket files when the database process stops is fixed. The bug prevents the database process from starting.
    • The priorities of two user-defined functions, nextval and currval functions, are adjusted.
    • The bug that causes the net_length_size function to return an invalid value, 251, is fixed.
20201031
  • New features:
    • The restoration of tables from the recycle bin is supported. For more information, see Recycle bin.
    • The automatic restoration of slow query log files at instance initialization is supported.
  • Performance optimization:

    If X-Engine is used, the Binlog in Redo feature cannot be enabled. For more information, see Binlog in Redo.

  • Bugs fixed:
    • The bug that causes errors in executing the ASSERT statement is fixed. This bug is triggered if the value of the key for a unique index is abnormally large.
    • The bug that prevents the COM_DAEMON process from stopping is fixed.
    • The bug that causes a cache overflow is fixed. This bug is triggered if you run Full-Text Search (FTS)-based queries.
    • The bug that prevents a proper rollback is fixed. This bug is triggered if instant DDL operations become abnormal.
20200831
  • New features:
    • An option is added to disable parallel scans for the count(*) function.
    • Start global transaction identifiers (GTIDs) and end GTIDs are introduced to the mysqlbinlog plug-in.
    • Log sequence numbers (LSNs) are supported in the redo log.
      • innodb_lsn: the LSN of each record in the redo log.
      • innodb_log_checkpoint_lsn: the LSN of the last checkpoint.
      • innodb_log_write_lsn: the LSN of each record that is written to the redo log.
      • innodb_log_ready_for_write_lsn: the LSN of the last record that is written to the log buffer.
      • innodb_log_flush_lsn: the LSN of each record that is flushed from the redo log to the disk.
      • innodb_log_dirty_pages_added_up_to_lsn: the LSN of each record that logs a page as dirty.
      • innodb_log_oldest_lsn: the LSN of each record that logs an update to a page.
  • Performance optimization:
    • The CCL mechanism is optimized to better schedule and concurrently run transactions.
    • The CCL mechanism is optimized to better prioritize the stored procedures that are to run.
  • Bugs fixed:
    • The bug that prevents the recursively called interpreter from checking the memory size is fixed.
    • The bug that prevents you from modifying table definitions is fixed. This bug is triggered if transparent data encryption (TDE) is enabled.
    • The bug that causes the event scheduler to leak memory is fixed.
20200630
  • New features:
    • The faster DDL feature is introduced. It provides an optimized buffer pool management mechanism. This mechanism reduces the impact of DDL operations on database performance and increases the number of online DDL operations that can be concurrently run. For more information, see Faster DDL.
    • The maximum number of connections that are allowed is increased to 500,000.
  • Performance optimization:
    • The thread pool feature is optimized.
    • The memory allocation mechanism is optimized. You can specify the maximum memory consumption allowed for PERFORMANCE SCHEMA based on the instance type.
    • SQL log files are no longer analyzed.
    • TDE is optimized to cache the keys that are provided by Key Management Service (KMS).
    • The statuses of threads that are managed by the CCL mechanism are modified. For more information, see Statement concurrency control.
  • Bugs fixed:
    • The bug that causes ApsaraDB RDS to consider a semicolon (;) to be a part of a statement is fixed. This bug is triggered if the statement is used to create a database outline.
    • The bug that causes an RDS instance to unexpectedly exit in the event of table modifications is fixed.
    • The bug that causes earlier versions to disallow two keywords that are supported in later versions is fixed. The keywords are memory and array.
    • The bug that causes ApsaraDB RDS to incorrectly count the number of waits is fixed. This bug is triggered if commands are read from a database client.
    • The bug that causes failures in minor engine version updates is fixed.
20200430
  • New features:
    • The Binlog in Redo feature is introduced. It allows ApsaraDB RDS to write binary log records to the redo log before the binary log records are written to the disk. This reduces I/O consumption and improves database performance. For more information, see Binlog in Redo.
    • A restructured row caching mechanism is introduced to X-Engine.
    • The XA_RECOVER_ADMIN permission is provided.
  • Performance optimization:
    • The code that is used to scan data when operations are performed on a temporary InnoDB table is optimized. This allows ApsaraDB RDS to scan only dirty pages instead of the entire buffer pool.
    • The global parameter opt_readonly_trans_implicit_commit is renamed as rds_disable_explicit_trans. This ensures compatibility with MySQL 5.6.
    • The SQL Explorer (SQL audit) feature is optimized. This feature no longer logs the upgrades to an RDS instance.
    • Memory resources that are consumed by the DDL operations on X-Engine tables are reduced.
  • Bugs fixed:
    • The bug that causes the sizes of X-Engine tables on the disk to be inconsistent with the statistical information in the INFORMATION_SCHEMA database is fixed.
    • The bug that causes ApsaraDB RDS to initialize X-Engine log records is fixed. This bug is triggered if the error log file is re-opened.
20200331
  • New feature:

    The TRUNCATE TABLE statement is supported. After you execute this statement on a table, this statement moves the table to a dedicated directory that is used for the recycle bin. Then, this statement creates a table by using the schema of the table that you have truncated. For more information, see Recycle bin.

  • Performance optimization:
    • The output of TCP errors is disabled by default.
    • The performance of the thread pool feature with the default configuration is improved.
  • Bugs fixed:
    • The bug that causes invalid databases and tables is fixed. This bug is triggered if the names of partitioned tables are separated with the combination of a pound key and a letter p (#p).
    • The bug that causes the CCL-managed statements to be case-sensitive is fixed.
  • Changes incorporated: The feature changes in open source MySQL 8.0.17 and MySQL 8.0.18 are incorporated. For more information, see Changes in MySQL 8.0.17 and Changes in MySQL 8.0.18.
20200229
  • New features:
    • The performance agent feature is introduced. For more information, see Performance Agent. This feature is provided as a MySQL plug-in. This feature allows ApsaraDB RDS to compute and analyze the metrics that are used to measure the performance of an RDS instance.
    • Network round-trip time is introduced as a metric to measure the performance of an RDS instance that runs in semi-synchronous mode.
    • Online DDL operations are supported for X-Engine.
  • Performance optimization:
    • The CCL mechanism is optimized to control the concurrency of statements on read-only RDS instances.
    • The database outline mechanism is optimized to support database outlines on secondary RDS instances.
    • The database proxy feature is enhanced to optimize short-lived connections.
    • The time that is required to execute the PAUSE statement is reduced in various CPU architectures.
    • A memory summary table is introduced to provide the statuses of thread pools.
  • Bugs fixed:
    • The bug that causes ApsaraDB RDS to forbid the ppoll function and replace the ppoll function with the poll function is fixed. This bug is found in Linux kernels earlier than version 4.9.
    • The bug that causes errors in invoking the wrap_sm4_encrypt function is fixed.
    • The bug that causes ApsaraDB RDS to lock global variables at the rotation of SQL logs is fixed.
    • The bug that causes errors in restoration inconsistency checks is fixed.
    • The bug that causes inaccurate time values in the io_statistics table is fixed.
    • The bug that causes an RDS instance to unexpectedly exit is fixed. This bug is triggered if invalid compression algorithms are called.
    • The bug that causes user columns in MySQL 8.0 and MySQL 5.6 to be incompatible is fixed.
  • Programs optimized:
    • The faster DDL feature is introduced. It provides an optimized buffer pool management mechanism. This mechanism reduces the impact of DDL operations on database performance and increases the number of online DDL operations that can be concurrently run. For more information, see Faster DDL.
    • The thread pool feature is optimized to improve database performance.
    • The mechanism that is used for reference counting is optimized to prevent memory leaks in the buffer.
20200110
  • New feature:

    Three hints are introduced. These hints can be used in the following statements to commit and roll transactions back at high speeds: SELECT, UPDATE, INSERT, and DELETE. This increases the throughput of your application. For more information, see Inventory Hint.

  • Performance optimization:
    • The CCL mechanism is optimized. When an RDS instance starts, CCL queue structures are initialized before CCL rules are initialized.
    • The file deletion mechanism is optimized. When ApsaraDB RDS asynchronously deletes files, the links to small files are canceled.
    • The thread pool feature is optimized. For more information, see Thread Pool.
    • The restoration inconsistency check feature is disabled by default.
    • The permissions that are required to configure variables are changed.
      • The user role that is authorized to configure the following variables is changed to standard account:
        • auto_increment_increment
        • auto_increment_offset
        • bulk_insert_buffer_size
        • binlog_rows_query_log_events
      • The user role that is authorized to configure the following variables is changed to superuser or system variable administrator:
        • binlog_format
        • binlog_row_image
        • binlog_direct
        • sql_log_off
        • sql_log_bin
20191225
  • New feature:

    The recycle bin feature is introduced. All the tables that you have deleted are moved to the recycle bin. You can specify a retention period within which you can retrieve tables from the recycle bin. For more information, see Recycle bin.

  • Performance optimization:
    • The mechanism that is used to process short-lived connections is optimized.
    • A dedicated thread is used to serve the maintain user. This prevents high availability (HA) failures.
    • The locking mechanism is optimized. If an error occurs when binary log records are flushed by using the redo log, ApsaraDB RDS can explicitly release the locks that are triggered by file synchronization.
    • The deletion of unnecessary TCP error logs is supported.
    • The thread pool feature is enabled by default.
  • Bugs fixed:
    • The bug that causes errors in updating slow query logs is fixed.
    • The bug that causes an inappropriate lock scope is fixed.
    • The bug that causes errors in core dumps is fixed. This bug is triggered if the select() function is invoked for TDE.
20191115

New feature:

The statement queue feature is introduced. It allows statements to queue in the same bucket. These statements may be executed on the same resources. For example, these statements are executed on the same row of a table. This reduces overheads that are caused by possible conflicts. For more information, see Statement Queue.

20191101
  • New features:
    • The SM4 encryption algorithm is supported for TDE. For more information, see Configure TDE for an ApsaraDB RDS for MySQL instance.
    • Data protection is supported for secondary RDS instances. Only the accounts with the SUPER or REPLICATION_SLAVE_ADMIN role have the permissions to insert, delete, and modify data in the slave_master_info, slave_relay_log_info, and slave_worker_info tables.
    • A mechanism is introduced to increase the priorities of auto-increment keys. If a table does not have a primary key or it does not have a unique key without a null value, the auto-increment key without a null value is assigned the highest priority.
    • A mechanism is introduced to prevent the automatic conversion of tables from the MEMORY storage engine to the MyISAM storage engine. These tables include system tables. These tables also include the tables that are called by the threads in the initializing state.
    • A mechanism is introduced to flush binary log records to the disk before redo log records.
    • A mechanism is introduced to stop the creation of temporary tables on an RDS instance when the RDS instance is locked.
    • X-Engine is provided to store transactions based on a log-structured merge (LSM) tree.
  • Performance optimization:
    • The thread pool feature is optimized to reduce mutexes. For more information, see Thread Pool.
    • The performance insight feature is optimized to monitor thread pools. For more information, see Performance Insight.
    • Parameter adjustment:
      • primary_fast_lookup: a session parameter. Default value: true.
      • thread_pool_enabled: a global parameter. Default value: true.
20191015
  • New features:
    • The TDE feature is introduced to support real-time I/O encryption and decryption on data files. This feature encrypts data before the data is written to the disk and decrypts data before the data is read from the disk to the memory. For more information, see Configure TDE for an ApsaraDB RDS for MySQL instance.
    • The returning feature is introduced. It allows DML statements to return result sets. In addition, the DBMS_TRANS package is provided for you to use this feature. For more information, see Returning.
    • The forced conversion from the MyISAM or MEMORY storage engine to the InnoDB storage engine is supported. If the global variable force_mysiam_to_innodb or force_memory_to_innodb is set to ON, a table is converted at table creation or modification from the MyISAM or MEMORY storage engine to the InnoDB storage engine.
    • A mechanism is introduced to forbid standard accounts from performing primary/secondary switchovers. Only privileged accounts have the permissions to perform primary/secondary switchovers.
    • A performance proxy plug-in is provided. It obtains performance data and saves the data as TXT files to your computer. These files are deleted by using a round-robin algorithm. Only the latest files at the single-digit second level are retained.
    • A configurable timeout period is introduced for mutexes in InnoDB. This timeout period can be changed by setting the global variable innodb_fatal_semaphore_wait_threshold. The default value of the global variable is 600.
    • Index hint errors can be ignored by setting the global variable ignore_index_hint_error. The default value of the global variable is false.
    • The SSL encryption feature can be disabled. For more information, see Configure SSL encryption on an ApsaraDB RDS for MySQL instance.
    • The output of TCP errors is supported. TCP errors in read, read-wait, and write-wait events along with their error codes are returned by using end_connection events. In addition, log records with information about the errors are generated.
  • Bugs fixed:
    • The bug that prevents a Linux operating system from merging local asynchronous I/O (AIO) requests before linear Read Ahead is triggered is fixed.
    • The bug that prevents the proper collection of table and index statistics is fixed.
    • The bug that prevents ApsaraDB RDS direct access to the primary key index of a table is fixed. This bug is triggered if a primary key is specified for the table.
20190915

Bug fixed:

The bug that causes memory leaks is fixed. This bug is triggered if you run the Cmd_set_current_connection process.

20190816
  • New features:
    • The thread pool feature is introduced to separate threads from sessions. If a large number of sessions are created, ApsaraDB RDS can run a small number of threads to complete the tasks in active sessions. For more information, see Thread Pool.
    • The CCL mechanism is introduced. It allows you to specify the maximum number of concurrent requests that are allowed. This enables ApsaraDB RDS to handle traffic bursts, process statements that consume excessive resources, and adapt to changes of SQL models. This also ensures the continuity and stability of your database service. For more information, see Statement concurrency control.
    • The statement outline feature is introduced to support optimizer hints and index hints. These hints are used to ensure that ApsaraDB RDS stably runs query plans. For more information, see Statement outline.
    • The Sequence engine is introduced to simplify the acquisition of sequence values. For more information, see Sequence Engine.
    • The Purge Large File Asynchronously feature is introduced to asynchronously delete files. Before you delete a tablespace, ApsaraDB RDS renames the files in the tablespace to mark them as temporary files. Then, ApsaraDB RDS starts background threads to asynchronously delete the temporary files. For more information, see Purge Large File Asynchronously.
    • The performance insight feature is introduced to support load monitoring, association analysis, and performance optimization at the instance level. This feature allows you to evaluate the loads on an RDS instance. This feature also allows you to locate performance issues and ensure the stability of your database service. For more information, see Performance Insight.
    • An optimized instance locking mechanism is introduced. It allows you to drop and truncate tables on an RDS instance even if the RDS instance is locked.
  • Bugs fixed:
    • The bug that causes ApsaraDB RDS to incorrectly calculate file sizes is fixed.
    • The bug that allows irrelevant processes to reuse idle memory resources is fixed.
    • The bug that causes a host to unexpectedly exit is fixed. This bug is triggered if the available cache size on the host is 0.
    • The bug that causes conflicts between implicit primary keys and CTS statements is fixed.
    • The bug that causes ApsaraDB RDS to incorrectly log slow queries is fixed.
20190601
  • Performance optimization:
    • Metadata locking on logging tables is reduced.
    • The code for termination options is restructured.
  • Bugs fixed:
    • The bug that prevents the SQL Explorer (SQL audit) feature from logging precompiled statements is fixed.
    • The bug that prevents ApsaraDB RDS from filtering out error log records from logging tables is fixed. This bug is triggered if the names of the logging tables are invalid.

MySQL 8.0 on RDS Enterprise Edition

Minor AliSQL version Description
20210305 Bug fixed:

The bug that causes PERFORMANCE SCHEMA to leak memory is fixed.

20200918 New feature:

SSL encryption can be enabled for the dedicated proxy of an RDS instance.

20200805 Bug fixed:

A metadata-related bug is fixed.

20200608
  • New feature:

    The TRUNCATE TABLE statement is supported. After you execute this statement on a table, this statement moves the table to a dedicated directory that is used for the recycle bin. Then, this statement creates a table by using the schema of the table that you have truncated. For more information, see Recycle bin.

  • Performance optimization:
    • The output of TCP errors is disabled by default.
    • The performance of the thread pool feature with the default configuration is improved.
  • Bugs fixed:
    • The bug that causes invalid databases and tables is fixed. This bug is triggered if the names of partitioned tables are separated with the combination of a pound key and a letter p (#p).
    • The bug that causes the CCL-managed statements to be case-sensitive is fixed.
  • Changes incorporated: The feature changes in open source MySQL 8.0.17 and MySQL 8.0.18 are incorporated. For more information, see Changes in MySQL 8.0.17 and Changes in MySQL 8.0.18.
20200317
  • New features:
    • The performance agent feature is introduced. For more information, see Performance Agent. This feature is provided as a MySQL plug-in. This feature allows ApsaraDB RDS to compute and analyze the metrics that are used to measure the performance of an RDS instance.
    • Network round-trip time is introduced as a metric to measure the performance of an RDS instance that runs in semi-synchronous mode.
  • Performance optimization:
    • The CCL mechanism is optimized to control the concurrency of statements on read-only RDS instances.
    • The database outline mechanism is optimized to support database outlines on secondary RDS instances.
    • The database proxy feature is enhanced to optimize short-lived connections.
    • The time that is required to execute the PAUSE statement is reduced in various CPU architectures.
    • A memory summary table is introduced to provide the statuses of thread pools.
  • Bugs fixed:
    • The bug that causes ApsaraDB RDS to forbid the ppoll function and replace the ppoll function with the poll function is fixed. This bug is found in Linux kernels earlier than version 4.9.
    • The bug that causes errors in invoking the wrap_sm4_encrypt function is fixed.
    • The bug that causes ApsaraDB RDS to lock global variables at the rotation of SQL logs is fixed.
    • The bug that causes errors in restoration inconsistency checks is fixed.
    • The bug that causes inaccurate time values in the io_statistics table is fixed.
    • The bug that causes an RDS instance to unexpectedly exit is fixed. This bug is triggered if invalid compression algorithms are called.
    • The bug that causes user columns in MySQL 8.0 and MySQL 5.6 to be incompatible is fixed.
  • Programs optimized:
    • The faster DDL feature is introduced. It provides an optimized buffer pool management mechanism. This mechanism reduces the impact of DDL operations on database performance and increases the number of online DDL operations that can be concurrently run. For more information, see Faster DDL.
    • The thread pool feature is optimized to improve database performance.
    • The mechanism that is used for reference counting is optimized to prevent memory leaks in the buffer.

MySQL 5.7 on RDS Basic Edition or High-availability Edition

Minor AliSQL version Description
20210630
  • New features:
    • The audit log format MySQL_V3 is supported.
    • Slow query logs and general logs can be truncated.
    • The check of out of memory (OOM) exceptions in the thread stack is supported.
    • The thread_pool_strict_mode parameter is supported. This parameter is used to control the maximum number of worker threads.
  • Bugs fixed:
    • The bug that causes the deletion of outdated socket lock files is fixed. This bug is triggered if you run the mysqld_safe script to start the mysqld process.
    • The bug that causes memory leaks in the recycle bin is fixed.
20210430
  • New features:
    • The recycle bin feature is introduced. For more information, see Recycle bin.
    • The Sequence engine feature is introduced. For more information, see Sequence Engine.
  • Performance optimization:
    • More metrics are added to the I_S.PERF_STATISTICS table.
    • The number of index scans is added to the I_S.INDEX_STATISTICS table.
    • The TDE feature is optimized to improve database performance.
  • Bug fixed:

    The bug that causes column generation exceptions during a rollback is fixed. This bug is found in the MySQL Community Edition.

20201231
  • New features:
    • Changes incorporated: The feature changes in open source MySQL 5.7.32 are incorporated. For more information, visit GitHub.
    • The session tracker feature is supported by the fast query cache. For more information, see Fast query cache.
  • Performance optimization:
    • More help information is added to the file_help_tables.sql file.
    • The timeout variable ccl_wait_timeout is added to the statement CCL feature. For more information, see Statement concurrency control.
  • Bugs fixed:
    • The bug that causes an RDS instance to unexpectedly exit is fixed. This bug is triggered if you run the rds_change_user command on the RDS instance.
    • The bug that causes an RDS instance to unexpectedly exit due to an excessively small amount of remaining space in the table cache is fixed. This bug is triggered if you run a large number of concurrent queries.
    • The bug that causes memory leaks is fixed. This bug is triggered if an incorrect reference count occurs in the buffer.
20201031

Bugs fixed:

  • The bug that causes the ROW_SEARCH_MVCC function to unexpectedly exit in the event of concurrent updates is fixed.
  • The bug that prevents an RDS instance from starting is fixed. This bug is triggered if you reconfigure the innodb_undo_tablespaces parameter for the RDS instance.
  • The bug that causes a cache overflow is fixed. This bug is triggered if you run FTS-based queries.
20200831
  • New features:
    • Changes incorporated: The feature changes in open source MySQL 5.7.30 are incorporated. For more information, visit GitHub.
    • The CCL mechanism is optimized to better schedule and concurrently run transactions.
    • Start GTIDs and end GTIDs are introduced to the mysqlbinlog plug-in.
    • LSNs are supported in the redo log.
      • innodb_lsn: the LSN of each record in the redo log.
      • innodb_log_write_lsn: the LSN of each record that is written to the redo log.
      • innodb_log_checkpoint_lsn: the LSN of the last checkpoint.
      • innodb_log_flushed_lsn: the LSN of each record that is flushed from the redo log to the disk.
      • innodb_log_Pages_flushed: the LSN of each record that logs an update to a page.
  • Performance optimization:

    The CCL mechanism is optimized to better prioritize the stored procedures that are to run.

  • Bug fixed:

    The bug that an incorrect reference count causes memory leaks is fixed. The bug is triggered if an incorrect reference count occurs in the temporary tables. This bug may cause inefficient flushing of dirty blocks from the buffer pool to the disk. As a result, no idle pages are available in the buffer pool, and the database performance deteriorates. For more information, see the official MySQL documentation.

20200630
  • New features:
    • Three hints are introduced. These hints can be used in the following statements to commit and roll transactions back at high speeds: SELECT, UPDATE, INSERT, and DELETE. This increases the throughput of your application. For more information, see Inventory Hint.
    • The CCL mechanism is introduced. It allows you to specify the maximum number of concurrent requests that are allowed. This enables ApsaraDB RDS to handle traffic bursts, process statements that consume excessive resources, and adapt to changes of SQL models. This also ensures the continuity and stability of your database service. For more information, see Statement concurrency control.
    • The statement queue feature is introduced. It allows statements to queue in the same bucket. These statements may be executed on the same resources. For example, these statements are executed on the same row of a table. This reduces overheads that are caused by possible conflicts. For more information, see Statement Queue.
    • The statement outline feature is introduced to support optimizer hints and index hints. These hints are used to ensure that ApsaraDB RDS stably runs query plans. For more information, see Statement outline.
    • The faster DDL feature is introduced. It provides an optimized buffer pool management mechanism. This mechanism reduces the impact of DDL operations on database performance and increases the number of online DDL operations that can be concurrently run. For more information, see Faster DDL.
    • The maximum number of connections that are allowed is increased to 500,000.
  • Performance optimization:
    • The call dbms_admin.show_native_procedure(); command is provided to display all the procedures of an RDS instance.
    • A new function is provided to delete orphan tables.
    • The thread pool feature is optimized.
    • Query caching is optimized.
    • The memory allocation mechanism is optimized. You can specify the maximum memory consumption allowed for PERFORMANCE SCHEMA based on the instance type.
  • Bug fixed:

    The bug that causes an audit update thread to enter an infinite loop is fixed.

20200430
  • Performance optimization:

    Read-write locks are no longer supported in the query cache. The default hash function is changed from LF_hash to murmur3 hash.

  • Bugs fixed:

    Two bugs that occur when ApsaraDB RDS hits the requested data in the query cache are fixed. These bugs are triggered if you run transactions at the REPEATABLE_READ isolation level.

20200331
  • New features:
    • The fast query cache is introduced. It is developed by Alibaba Cloud based on the native MySQL query cache. It uses a new design and a new implementation mechanism to improve query performance. For more information, see Fast query cache.
    • Two metadata locks are introduced from Percona Server 5.7: LOCK TABLES FOR BACKUP (LTFB) and LOCK BINLOG FOR BACKUP (LBFB).
  • Performance optimization:
    • The thread pool feature is optimized to ensure compatibility with earlier MySQL versions.
    • The output of TCP errors is disabled by default.
    • The performance of the thread pool feature with the default configuration is improved.
  • Bugs fixed:
    • The bug that causes ApsaraDB RDS to delete temporary files is fixed. This bug is triggered if you delete large files.
    • The bug that causes dump threads in thread pools to time out is fixed.
    • The bug that causes ApsaraDB RDS to incorrectly count the value of the IPK field in the procedure context is fixed.
    • The bug that causes the leak and release of PFS threads is fixed. This bug is triggered if you run the rds_change_user command.
  • Changes incorporated: The feature changes in open source MySQL 5.7.28 are incorporated. For more information, visit GitHub.
20200229
  • New features:
    • The performance agent feature is introduced. For more information, see Performance Agent. This feature is provided as a MySQL plug-in. This feature allows ApsaraDB RDS to compute and analyze the metrics that are used to measure the performance of an RDS instance.
    • Network round-trip time is introduced as a metric to measure the performance of an RDS instance that runs in semi-synchronous mode.
  • Performance optimization:
    • The time that is required to execute the PAUSE statement is reduced in various CPU architectures.
    • The database proxy feature is enhanced to optimize short-lived connections.
    • A memory summary table is introduced to provide the statuses of thread pools.
  • Bugs fixed:
    • The bug that impairs the security of DDL-related redo log records is fixed.
    • The bug that causes inaccurate time values in the io_statistics table is fixed.
    • The bug that causes an RDS instance to unexpectedly exit in the event of table modifications is fixed.
    • The bugs in MySQL test cases are fixed.
20200110

Performance optimization:

  • The file deletion mechanism is optimized. When ApsaraDB RDS asynchronously deletes files, the links to small files are canceled.
  • The thread pool feature is optimized. For more information, see Thread Pool.
  • The default value of the thread_pool_enabled parameter is changed to OFF.
20191225
  • New feature:

    The management of internal accounts is supported. This allows you to manage user permissions and protect your data.

  • Performance optimization:
    • The mechanism that is used to process short-lived connections is optimized.
    • A dedicated thread is used to serve the maintain user. This prevents HA failures.
    • The deletion of unnecessary TCP error logs is supported.
    • The thread pool feature is optimized.
  • Bugs fixed:
    • The bug that causes the mysqld process to unexpectedly exit is fixed. This bug is triggered if the read/write splitting feature is enabled.
    • The bug that causes errors in core dumps is fixed. This bug is triggered if ApsaraDB RDS uses a keyring.
20191115

Bug fixed:

The bug that causes ApsaraDB RDS to display variables in SQL logs is fixed. This bug is triggered if a primary/secondary switchover is performed.

20191101
  • New features:
    • The SM4 encryption algorithm is supported for TDE. For more information, see Configure TDE for an ApsaraDB RDS for MySQL instance.
    • The bug that prevents ApsaraDB RDS direct access to the primary key index of a table is fixed. This bug is triggered if a primary key is specified for the table.
    • A mechanism is introduced to prevent the automatic conversion of tables from the MEMORY storage engine to the MyISAM storage engine. These tables include system tables. These tables also include the tables that are called by the threads in the initializing state.
  • Performance optimization:
    • The thread pool feature is optimized to reduce mutexes. For more information, see Thread Pool.
    • An SQL log caching mechanism is introduced to increase SQL logging performance.
    • The performance insight feature is optimized to monitor thread pools. For more information, see Performance Insight.
    • The thread pool feature is enabled by default. For more information, see Thread Pool.
  • Bugs fixed:
    • The bug that causes the release of locks on user tables is fixed. This bug is triggered if these tables are being managed or maintained.
    • More TCP errors are added.
20191015
  • New features:
    • The rotation of slow query logs is supported. Every CSV slow query log file is assigned a unique name and saved as a different new file. This prevents data loss at the collection of slow query logs. You can run the show variables like '%rotate_log_table%'; command to check whether the rotation of slow query logs is enabled.
    • A performance proxy plug-in is provided. It obtains performance data and saves the data as TXT files to your computer. These files are deleted by using a round-robin algorithm. Only the latest files at the single-digit second level are retained.
    • The forced conversion from the MEMORY storage engine to the InnoDB storage engine is supported. If the global variable rds_force_memory_to_innodb is set to ON, a table is converted at table creation or modification from the MEMORY storage engine to the InnoDB storage engine.
    • The keyring-rds plug-in is introduced to TDE. This plug-in allows ApsaraDB RDS to communicate with the administration system or KMS.
    • The output of TCP errors is supported. TCP errors in read, read-wait, and write-wait events along with their error codes are returned by using end_connection events. In addition, log records with information about the errors are generated.
  • Bug fixed:

    The bug that causes Error 1290 in DDL operations is fixed.

20190925

Parameter adjustment:

  • The default value of the system variable auto_generate_certs is changed from true to false.
  • The global read-only variable auto_detact_certs is introduced. Valid values: true and false. Default value: false. This variable is supported for an RDS instance only when code is compiled by using OpenSSL on the RDS instance. This variable specifies whether the RDS instance automatically searches for SSL certificate files and key files in the data directory.
20190915

New feature:

The thread pool feature is introduced to separate threads from sessions. If a large number of sessions are created, ApsaraDB RDS can run a small number of threads to complete the tasks in active sessions. For more information, see Thread Pool.

20190815
  • New features:
    • The Purge Large File Asynchronously feature is introduced to asynchronously delete files. Before you delete a tablespace, ApsaraDB RDS renames the files in the tablespace to mark them as temporary files. Then, ApsaraDB RDS starts background threads to asynchronously delete the temporary files. For more information, see Purge Large File Asynchronously.
    • The performance insight feature is introduced to support load monitoring, association analysis, and performance optimization at the instance level. This feature allows you to evaluate the loads on an RDS instance. This feature also allows you to locate performance issues and ensure the stability of your database service. For more information, see Performance Insight.
    • An optimized instance locking mechanism is introduced. It allows you to drop and truncate tables on an RDS instance even if the RDS instance is locked.
  • Bugs fixed:
    • The bug that allows you to set the rds_prepare_begin_id parameter in the set rds_current_connection command is fixed.
    • The bug that prevents ApsaraDB RDS from updating information about locked accounts is fixed.
    • The bug that allows you to use actual as a keyword in table names is fixed.
    • The bug that causes an overflow of timestamps in slow query logs is fixed.
20190510

New feature:

The creation of temporary tables in transactions is supported.

20190319

New feature:

The configuration of thread IDs for proxies in handshake packets is supported.

20190131
  • Performance optimization:
    • The upgrade to MySQL 5.7.25 is supported.
    • JeMalloc that is used for memory management is disabled.
  • Bug fixed:

    The bug that causes ApsaraDB RDS to incorrectly calculate the value of the internal variable net_lenth_size is fixed.

20181226
  • New feature:

    Dynamic modifications to the system variable binlog-row-event-max-size are supported. This expedites the replication of tables that do not have a primary key.

  • Bug fixed:

    The bug that prevents the proxy instances of an RDS instance from applying for memory resources is fixed.

20181010

Performance optimization:

  • Implicit primary keys are supported.
  • The replication of tables between primary and secondary RDS instances is accelerated. This is supported if these tables do not have a primary key.
  • Native AIO is provided to improve I/O performance.
20180431

New features:

  • The RDS High-availability Edition is supported.
  • The SQL audit feature is supported. For more information, see SQL audit.
  • The protection for RDS instances on which snapshot backups are being created is enhanced.

ApsaraDB RDS for MySQL 5.7 on RDS Enterprise Edition

Minor AliSQL version Description
20191128
  • New feature:

    The read/write splitting feature is supported.

  • Bugs fixed:
    • The bug that causes ApsaraDB RDS to incorrectly calculate the value of the Second_Behind_Master metric for a follower is fixed.
    • The bug that causes deadlocks is fixed. This bug is triggered if you re-execute table-level parallel replication transactions.
    • XA-related bugs are fixed.
20191016
  • New features:
    • The upgrade from the RDS High-availability Edition to the RDS Enterprise Edition is supported for RDS instances that are equipped with local SSDs.
    • GTIDs that are provided by open source MySQL are supported. GTIDs are disabled by default.
    • All the proprietary AliSQL features that are released by Alibaba Cloud in the RDS Basic Edition and the RDS High-availability Edition in 20190915 and earlier minor engine versions are incorporated.
  • Bug fixed:

    The bug that causes ApsaraDB RDS to disable the binary log for an RDS instance is fixed. This bug is triggered if you reset the secondary RDS instance of the RDS instance.

20190909
  • New features:
    • A mechanism is introduced to ensure that large transactions are executed at fast speeds. This mechanism is supported when the synchronous mode is used to replicate data between primary and secondary RDS instances that run the RDS Enterprise Edition.
    • The dumping of binary log records from a leader or a follower is supported.
    • The creation of read-only RDS instances is supported.
    • The InnoDB storage engine is used to store system tables by default.
  • Bugs fixed:
    • The bug that invalidates commands is fixed. These commands are run by a follower to delete logs.
    • The bug that causes slave threads to unexpectedly exit is fixed. This bug is triggered if the slave_sql_verify_checksum parameter is set to OFF and the binlog_checksum parameter is set to crc32.
20190709

New features:

  • The RDS Enterprise Edition is supported.
  • The disabling of the semi-sync plug-in is supported.
  • Table-level parallel replication and write set-level parallel replication are supported.
  • The pk_access module is introduced to expedite queries that are run based on primary keys.
  • The thread pool feature is supported.
  • All the proprietary AliSQL features that are released by Alibaba Cloud in the RDS Basic Edition and the RDS High-availability Edition in 20190510 and earlier minor engine versions are incorporated.

MySQL 5.6

Minor AliSQL version Description
20210630
  • New feature:

    The OOM exceptions in the thread stack is supported.

  • Bugs fixed:
    • The bugs that are found in the MySQL 5.6 Community Edition are fixed.
    • The operation limits on the SHOW CREATE statement in the mysql system database are removed.
20210430
  • Performance optimization:

    More metrics are added to the I_S.PERF_STATISTICS table.

  • Bugs fixed:

    The bugs that cause memory leaks in the MySQL Community Edition are fixed.

20201231
  • New features:
    • Three session trackers are introduced to track user variables.
    • A mechanism is introduced to lock only the first sub-partitioned table. This reduces CPU utilization.
  • Performance optimization:

    More help information is added to the file_help_tables.sql file.

  • Bugs fixed:
    • The bug that causes lock conflicts is fixed. This bug is triggered if you execute the SHOW BINLOG EVENTS statement.
    • The bug that causes ApsaraDB RDS to return an inaccurate value for the net_length_size parameter is fixed.
20201031 Bugs fixed:
  • The bug that invalidates subqueries is fixed. This bug is triggered if these subqueries nested in IN clauses.
  • The bug that causes errors in the permissions on processes is fixed.
  • The bug that causes errors in the permissions on users in the kill_user_list table is fixed.
  • The bug that causes errors in executing the DROP DATABASE statement is fixed.
  • The bug that causes ApsaraDB RDS to incorrectly calculate the value of the SECONDS_BEHIND_MASTER parameter due to PREVIOUS_GTID events is fixed.
20200831
  • New feature:

    LSNs are supported in the redo log.

    • innodb_lsn: the LSN of each record in the redo log.
    • innodb_log_write_lsn: the LSN of each record that is written to the redo log.
    • innodb_log_checkpoint_lsn: the LSN of the last checkpoint.
    • innodb_log_flushed_lsn: the LSN of each record that is flushed from the redo log to the disk.
    • innodb_log_Pages_flushed: the LSN of each record that logs an update to a page.
  • Bugs fixed:
    • The bug that causes the SHOW_HA_ROWS statement to return invalid enumeration types is fixed.
    • The bug that causes ApsaraDB RDS to incorrectly count the value of the IPK field in the procedure context is fixed.
    • The bug that causes an RDS instance to unexpectedly exit is fixed. This bug is triggered if you query data from the INFORMATION_SCHEMA database.
    • The bug that causes an audit update thread to enter an infinite loop is fixed.
    • The bug that prevents secondary RDS instances from reporting data replication latencies is fixed.
20200630
  • New features:
    • The performance agent feature is introduced. For more information, see Performance Agent. This feature is provided as a MySQL plug-in. This feature allows ApsaraDB RDS to compute and analyze the metrics that are used to measure the performance of an RDS instance.
    • The maximum number of connections that are allowed is increased to 500,000.
    • The faster DDL feature is introduced. It provides an optimized buffer pool management mechanism. This mechanism reduces the impact of DDL operations on database performance and increases the number of online DDL operations that can be concurrently run. For more information, see Faster DDL.
  • Performance optimization:
    • The global parameter max_execution_time is introduced. If the time that is required to execute an SQL statement exceeds the value of this parameter, the SQL statement is paused.
    • The thread pool feature is optimized.
  • Bug fixed:

    The bug that causes ApsaraDB RDS to incorrectly count the number of waits is fixed. This bug is triggered if commands are read from a database client.

20200430
  • New features:

    The mdl_info table is provided to store the information about metadata locks.

  • Bug fixed:

    The bug that causes conflicts between the thread pool feature and the ic_reduce feature is fixed.

20200331

Performance optimization:

  • The performance of the thread pool feature with the default configuration is improved.
  • The output of TCP errors is disabled by default.
20200229
  • New feature:

    The read/write splitting feature is supported for database proxies.

  • Performance optimization:
    • The thread pool feature is optimized.
    • The time that is required to execute the PAUSE statement is reduced in various CPU architectures.
  • Bug fixed:

    The bug that causes ApsaraDB RDS to partially commit XA transactions is fixed.

20200110
  • New feature:

    The thread pool feature is introduced to separate threads from sessions. If a large number of sessions are created, ApsaraDB RDS can run a small number of threads to complete the tasks in active sessions. For more information, see Thread Pool.

  • Performance optimization:

    The file deletion mechanism is optimized. When ApsaraDB RDS asynchronously deletes files, the links to small files are canceled.

  • Bugs fixed:
    • The bug that causes ApsaraDB RDS to incorrectly calculate the sleep time of the page cleaner is fixed.
    • The bug that causes the SELECT @@global.gtid_executed statement to trigger failover failures is fixed.
    • The bug that causes the "IF CLIENT KILLED AFTER ROLLBACK TO SAVEPOINT PREVIOUS STMTS COMMITTED" error is fixed. For more information, see IF CLIENT KILLED AFTER ROLLBACK TO SAVEPOINT PREVIOUS STMTS COMMITTED.
20191212 Performance optimization:

The deletion of unnecessary TCP error logs is supported.

20191115 Bug fixed:

The bug that causes an overflow of timestamps in slow query logs is fixed.

20191101 Bugs fixed:
  • The bug that causes ApsaraDB RDS to rotate slow query logs is fixed. This bug is triggered if common logs are updated.
  • Some display-related bugs are fixed.
20191015
  • New features:
    • The rotation of slow query logs is supported. Every CSV slow query log file is assigned a unique name and saved as a different new file. This prevents data loss at the collection of slow query logs. You can run the show variables like '%rotate_log_table%'; command to check whether the rotation of slow query logs is enabled.
    • A new SM4 encryption algorithm is introduced to replace the previously used SM4 encryption algorithm.
    • The Purge Large File Asynchronously feature is introduced to asynchronously delete files. Before you delete a tablespace, ApsaraDB RDS renames the files in the tablespace to mark them as temporary files. Then, ApsaraDB RDS starts background threads to asynchronously delete the temporary files. For more information, see Purge Large File Asynchronously.
    • The output of TCP errors is supported. TCP errors in read, read-wait, and write-wait events along with their error codes are returned by using end_connection events. In addition, log records with information about the errors are generated.
    • An SQL log caching mechanism is introduced to increase SQL logging performance.
  • Bugs fixed:
    • The bug that prevents responses to the pstack command is fixed by disabling the pstack command. This bug is triggered if a large number of connections are established.
    • The bug that causes conflicts between implicit primary keys and the CREATE TABLE AS SELECT statement is fixed.
    • The bug that prevents ApsaraDB RDS from deleting temporary files is fixed. These temporary files are created from binary log files.
20190815 Performance optimization:

An optimized instance locking mechanism is introduced. It allows you to drop and truncate tables on an RDS instance even if the RDS instance is locked.

20190130 Bugs fixed:

The bugs that compromise database stability are fixed.

20181010 Performance optimization:

The rocksdb_ddl_commit_in_the_middle parameter is introduced to MyRocks. If this parameter is set to on, some DDL statements call the COMMIT operation when they are executed.

201806** New feature:

Microsecond-level time precision is supported for slow query logs.

20180426
  • New feature:

    Invisible indexes are supported. For more information, see AliSQL 5.6.32 Release Notes (2017-07-16).

  • Bugs fixed:
    • The bug that causes ApsaraDB RDS to apply threads on secondary RDS instances is fixed.
    • The bug that compromises database performance is fixed. This bug is triggered if the updates to partitioned tables are applied on secondary RDS instances.
    • The bug that causes TokuDB to rebuild tables is fixed. This bug is triggered if you execute the ALTER TABLE COMMENT statement on these tables. For more information, see AliSQL 5.6.32 Release Note (2018-05-01).
    • The bug that causes deadlocks is fixed. This bug is triggered if you execute the SHOW SLAVE STATUS statement or the SHOW STATUS statement.
20171205 Bugs fixed:
  • The bug that causes deadlocks is fixed. This bug is triggered if you execute the OPTIMIZE TABLE statement and the ONLINE ALTER TABLE statement at the same time.
  • The bug that causes conflicts between sequences and implicit primary keys is fixed.
  • The bug that prevents ApsaraDB RDS from properly executing the SHOW CREATE SEQUENCE statement is fixed.
  • The bug that prevents ApsaraDB RDS from accurately collecting statistics about TokuDB tables is fixed.
  • The bug that causes deadlocks is fixed. This bug is triggered if you execute the OPTIMIZE statement on multiple tables in parallel.
  • The bug that prevents ApsaraDB RDS from accurately recording character sets in QUERY_LOG_EVENT is fixed.
  • The bug that prevents an RDS instance from stopping due to signal processing issues is fixed. For more information, see AliSQL 5.6.32 Release Notes (2017-10-10).
  • The bug that is caused by the RESET MASTER statement is fixed.
  • The bug that causes secondary RDS instances to be in a constant waiting state is fixed.
  • The bug that causes status errors in primary and secondary RDS instances is fixed. This bug is triggered if you switch your database service over between the primary and secondary RDS instances and these instances run the RDS Enterprise Edition.
  • The bug that may cause the database process to unexpectedly exit is fixed. This bug is triggered if you execute the SHOW CREATE TABLE statement.
20170927 Bug fixed:

The bug that causes ApsaraDB RDS to query tables from TokuDB based on inappropriate indexes is fixed.

20170901
  • New features:
  • Bug fixed:

    The bug that causes ApsaraDB RDS to return an inaccurate result set for a NOT IN query is fixed.

20170530 New feature:

The privileged account of an RDS instance is granted the permissions to close the connections that are established by all the standard accounts of the RDS instance.

20170221 New feature:

The read/write splitting feature is supported. For more information, see Enable the proxy terminal feature for an ApsaraDB RDS for MySQL instance.

MySQL 5.5

Minor AliSQL version Description
20181212 Bug fixed:

The bug that causes the gettimeofday(2) function to return an inaccurate time value is fixed. The returned time value is used to calculate the timeout period. If the time value that is returned is inaccurate, some operations never time out.