Aurora MySQL database engine updates 2020-06-02 (version 2.08.0) (Deprecated)

Improved binary log (binlog) processing to reduce crash recovery time and commit time latency when very large transactions are involved.

Launching Database Activity Streams (DAS) feature for Aurora MySQL. This feature provides a near-real-time data stream of the database activity in your relational database to help you monitor activity. For more information, see Monitoring Amazon Aurora with Database Activity Streams in the Amazon Aurora User Guide.

Updated timezone files to support the latest Brazil timezone change.

Introduced new keywords in SQL to exercise the hash join functionality for a specific table and/or inner table: HASH_JOIN, HASH_JOIN_PROBING, and HASH_JOIN_BUILDING. For additional details, see Aurora MySQL hints in the Amazon Aurora User Guide.

Introduced join order hint support in Aurora MySQL 5.7 by backporting a MySQL 8.0 feature. The new hints are JOIN_FIXED_ORDER, JOIN_ORDER, JOIN_PREFIX, and JOIN_SUFFIX. For detailed documentation of join order hint support, see WL#9158: Join order hints.

Aurora Machine Learning now supports user-defined functions with MEDIUMINT as the return type.

The lambda_async() stored procedure now supports all MySQL utf8 characters.

Fixed an issue that could cause a reader DB instance to return incomplete results for an FTS query after the INFORMATION_SCHEMA.INNODB_SYS_TABLES table is queried on the writer DB instance.

CVE-2019-5443

CVE-2019-3822

Fixed an issue that resulted in a database restart after a multi-query statement that accesses multiple tables or databases is executed with the query cache enabled.

Fixed a race condition in the lock manager that resulted in a database restart or failover during transaction rollback.

Fixed an issue that triggered database restart or failover when multiple connections are trying to update the same table with a Full-Text Search index.

Fixed an issue that could trigger a database restart or failover during a kill session command. If you encounter this issue, contact Amazon support to enable this fix on your instance.

Fixed an issue that caused reader DB instance to restart during a multi-statement transaction with multiple SELECT statements and a heavy write workload on the writer DB instance with AUTOCOMMIT enabled.

Fixed an issue that caused reader DB instance to restart after executing long-running queries while the writer DB instance is under a heavy OLTP write workload.

Improved database recovery time and commit latency for long running transactions when binlog is enabled.

Improved the algorithm to generate better statistics for estimating distinct value counts on indexed columns, including columns with skewed data distributions.

Reduced the response time and CPU utilization of join queries that access MyISAM temporary tables and the results spill to local storage.

Fixed an issue that prevented Aurora MySQL 5.6 snapshots with database or table names containing spaces from being restored to a new Aurora MySQL 5.7 cluster.

Included victim transaction info when deadlock is resolved in show engine innodb status.

Fixed an issue that caused connections to get stuck when clients of multiple different versions are connected to the same database and are accessing the query cache.

Fixed a memory leak resulting from multiple invocations of the Zero-Downtime Patch (ZDP) or Zero-Downtime Restart (ZDR) workflow throughout the lifetime of a database instance.

Fixed an error message in Zero-Downtime Patch (ZDP) or Zero-Downtime Restart (ZDR) operations wrongly stating that the last transaction was aborted if the auto-commit flag is turned off.

Fixed an issue in Zero-Downtime Patch (ZDP) operations that could lead to a server failure error message when restoring user session variables in the new database process.

Fixed an issue in Zero Downtime Patch (ZDP) operations that might cause intermittent database failures when there are long running queries during patching.

Fixed an issue where queries including an Aurora Machine Learning function returned empty error messages due to an incorrectly handled error response from Machine Learning services such as Amazon SageMaker and Amazon Comprehend.

Fixed an issue in the out-of-memory monitoring functionality that did not honor a custom value of the table_definition_cache parameter.

The error message "Query execution was interrupted" is returned if an Aurora Machine Learning query is interrupted. Previously, the generic message "Internal error in processing ML request" was returned instead.

Fixed an issue that could cause a binlog worker to experience a connection timeout when the slave_net_timeout parameter is less than the aurora_binlog_replication_max_yield_seconds parameter and there is low workload on the binlog master cluster.

Improved monitoring of the binlog recovery progress by outputting informational messages in the error log at a frequency of one message per minute.

Fixed an issue that could cause active transactions not to be reported by the SHOW ENGINE INNODB STATUS query.

Bug #25289359: A full-text cache lock taken when data is synchronized was not released if the full-text cache size exceeded the full-text cache size limit.

Bug #29138644: Manually changing the system time while the MySQL server was running caused page cleaner thread delays.

Bug #25222337: A NULL virtual column field name in a virtual index caused a server exit during a field name comparison that occurs while populating virtual columns affected by a foreign key constraint.

Bug #25053286: Executing a stored procedure containing a query that accessed a view could allocate memory that was not freed until the session ended.

Bug #25586773: Executing a stored procedure containing a statement that created a table from the contents of certain SELECT statements could result in a memory leak.

Bug #28834208: During log application, after an OPTIMIZE TABLE operation, InnoDB did not populate virtual columns before checking for virtual column index updates.

Bug #26666274: Infinite loop in performance schema buffer container due to 32-bit unsigned integer overflow.

Asynchronous key prefetch (AKP). For more information, see Optimizing Aurora indexed join queries with asynchronous key prefetch in the Amazon Aurora User Guide.

Hash joins. For more information, see Optimizing large Aurora MySQL join queries with hash joins in the Amazon Aurora User Guide.

Native functions for synchronously invoking Amazon Lambda functions. For more information, see Invoking a Lambda function with an Aurora MySQL native function in the Amazon Aurora User Guide.

Scan batching. For more information, see Aurora MySQL database engine updates 2017-12-11 (version 1.16) (Deprecated).

Migrating data from MySQL using an Amazon S3 bucket. For more information, see Migrating data from MySQL by using an Amazon S3 bucket in the Amazon Aurora User Guide.

Group replication plugin

Increased page size

InnoDB buffer pool loading at startup

InnoDB full-text parser plugin

Multisource replication

Online buffer pool resizing

Password validation plugin

Query rewrite plugins

Replication filtering

The CREATE TABLESPACE SQL statement