Documentation Home
MySQL 8.0 Release Notes
Related Documentation Download these Release Notes
PDF (US Ltr) - 4.2Mb
PDF (A4) - 4.3Mb


MySQL 8.0 Release Notes  /  Changes in MySQL 8.0.13 (2018-10-22, General Availability)

Changes in MySQL 8.0.13 (2018-10-22, General Availability)

For general information about upgrades, downgrades, platform support, etc., please visit https://dev.mysql.com/doc/relnotes/mysql/8.0/en/.

Account Management Notes

  • It is now possible to require that attempts to change an account password be verified by specifying the current password to be replaced. This enables DBAs to prevent users from changing a password without proving that they know the current password. It is possible to establish password-verification policy globally using the password_require_current system variable, as well as on a per-account basis using the PASSWORD REQUIRE option of the CREATE USER and ALTER USER statements. Together with existing password-management capabilities, the new capability of requiring verification provides DBAs more complete control over password management. For more information, see Password Management.

    Important

    The implementation of password-verification capability involves a change to the structure of the mysql.user system table. If you upgrade to this MySQL release from an earlier version, you must run mysql_upgrade (and restart the server) to incorporate this system database change. Until this is done, password changes are not possible.

    (WL #11544)

Compilation Notes

  • Solaris: MySQL now can be compiled on Solaris using gcc. (Bug #27802681)

Configuration Notes

  • The new WITH_LTO CMake option controls whether to enable link-time optimization. Currently, this is supported only by GCC 7 and 8. (Bug #28184537, Bug #28211382)

  • The new WITH_RAPIDJSON CMake option controls whether to compile with the bundled or system RapidJSON library. (Bug #28024992, Bug #90867)

  • The CMAKE_BUILD_TYPE CMake option now supports a Release build type, which is like the RelWithDebInfo build type but omits debugging information to reduce the build size. (Bug #27874068, Bug #32287863)

  • The new sql_require_primary_key system variable makes it possible to have statements that create new tables or alter the structure of existing tables enforce the requirement that tables have a primary key. Enabling this variable helps avoid performance problems in row-based replication that can occur when tables have no primary key. Suppose that a table has no primary key and an update or delete modifies multiple rows. On the master server, this operation can be performed using a single table scan but, when replicated using row-based replication, results in a table scan for each row to be modified on the slave. With a primary key, these table scans do not occur. (Bug #17468242, Bug #69845, Bug #17005592, Bug #69223, WL #11807)

  • To enable the server to listen on a set of addresses, the bind_address system variable now permits a list of comma-separated IP addresses or host names, not just a single address or name. For details, see Server System Variables. (WL #11652)

Data Type Notes

  • MySQL now supports use of expressions as default values in data type specifications. This includes the use of expressions as default values for the BLOB, TEXT, GEOMETRY, and JSON data types, which previously could not be assigned default values at all. For details, see Data Type Default Values. (WL #9418)

Deprecation and Removal Notes

  • Important Change: Setting user variables in statements other than SET is now deprecated due to issues that included those listed here:

    • The order of evaluation for expressions involving user variables was undefined.

    • The default result type of a variable is based on its type at the beginning of the statement, which could have unintended effects when a variable holding a value of one type at the beginning of a statement was assigned a new value of a different type in the same statement.

    • HAVING, GROUP BY, and ORDER BY clauses, when referring to a variable that was assigned a value in the select expression list, did not work as expected because the expression was evaluated on the client and so it was possible for stale column values from a previous row to be used.

    Syntax such as SELECT @var, @var:=@var+1 is still accepted in MySQL 8.0 for backward compatibility, but is subject to removal in a future MySQL version. (WL #12131)

  • InnoDB; Partitioning: Support for placing table partitions in shared tablespaces was removed. Shared tablespaces include the system tablespace and general tablespaces. For information about identifying partitions in shared tablespaces and moving them to file-per-table tablespaces, see Preparing Your Installation for Upgrade. (WL #11571, WL #9286)

  • InnoDB: Support for TABLESPACE = innodb_file_per_table and TABLESPACE = innodb_temporary clauses with CREATE TEMPORARY TABLE is deprecated and will be removed in a future MySQL version. (WL #12179)

  • The utf8mb3 character set is deprecated and will be removed in a future MySQL version. Please use utf8mb4 instead. (WL #11109)

  • Nested comments have never been supported (although they might be permitted under some conditions), but now are considered deprecated and will be removed in a future MySQL version. (WL #12099)

  • The deprecated metadata_locks_cache_size and metadata_locks_hash_instances system variables were removed. (WL #12212)

  • The PAD_CHAR_TO_FULL_LENGTH SQL mode has been deprecated and will be removed in a future MySQL version. (WL #12129)

Error Handling

  • The MySQL client library now returns better error messages for OpenSSL errors. Thanks to Facebook for the patch. (Bug #27855668, Bug #90418)

  • Previously, the ER_NO_REFERENCED_ROW_2 and ER_ROW_IS_REFERENCED_2 error messages for foreign key operations were displayed and revealed information about parent tables, even when the user had no parent table access privileges. Error handling for this situation has been revised:

    An exception is that, for stored programs defined to execute with DEFINER privileges, the user against which privileges are assessed is the user in the program DEFINER clause, not the invoking user. If that user has table-level parent table privileges, parent table information is still displayed. In this case, it is the responsibility of the stored program creator to hide the information by including appropriate condition handlers. (Bug #19477611, WL #8910)

INFORMATION_SCHEMA Notes

Logging Notes

  • Incompatible Change: The system variables have been removed that previously configured error logging to the system log (the Event Log on Windows, and syslog on Unix and Unix-like systems). Where appropriate, the removed system variables were replaced with new system variables managed by the log_sink_syseventlog error log component. The following table shows the old and new variable names.

    Old System Variable New System Variable
    log_syslog_facility syseventlog.facility
    log_syslog_include_pid syseventlog.include_pid
    log_syslog_tag syseventlog.tag
    log_syslog None
    Important

    Installations that used the old system variable names must update their configuration to use the new variable names. For more information, see Error Logging to the System Log.

    (WL #11828)

    References: See also: Bug #27534089.

  • A new system variable, log_error_suppression_list, specifies which diagnostics should not be written to the error log when they occur with a severity of WARNING or INFORMATION. For example, if a particular type of warning occurs frequently but is not of interest (and thus may be considered undesirable noise in the error log), it now can be suppressed. (WL #11393)

  • The code to handle rewriting of account-management statements was refactored to make it easier to maintain and extend. This work resulted in a few minor user-visible effects in the audit, general, and slow query logs:

    • Plaintext passwords are replaced by <string> rather than '<string>'.

    • Default clauses are not written unless specified by the user.

    (WL #12007)

  • Previously, messages written to the error log by several internal server methods were logged using the ER_LOG_PRINTF_MSG error code. Each of these messages now is logged using a unique error code.

Optimizer Notes

  • The optimizer now supports a Skip Scan access method that enables range access to be used in previously inapplicable situations to improve query performance. For more information, see Skip Scan Range Access Method. Thanks to Facebook for the patch on which this access method is based. (Bug #26976512, Bug #88103, WL #11322)

  • MySQL now supports creation of functional index key parts that index expression values rather than column values. Functional key parts enable indexing of values that cannot be indexed otherwise, such as JSON values. For details, see CREATE INDEX Statement. (WL #1075)

  • Performance of SELECT COUNT(*) FROM tbl_name queries for InnoDB tables was improved for single-threaded workloads and when no extra clauses such as WHERE or GROUP BY are used. (WL #10398)

Packaging Notes

  • The libevent library bundled with MySQL was upgraded to version 2.1.8. (Bug #28207237, Bug #29041505, Bug #29055011)

  • Binary packages that include curl rather than linking to the system curl library now use curl 7.60.0 rather than 7.45.0. (Bug #28043702)

  • Debian packaging was updated to reflect the removal of yaSSL and that OpenSSL is the default SSL library for all builds. (Bug #28025599)

  • Test plugins have been moved from server packages to test packages. (Bug #27860172)

  • MySQL Router is now included in the MySQL server source and monolithic binary packages.

Performance Schema Notes

  • A new Performance Schema stage, waiting for handler commit, is available to detect threads going through transaction commit. Thanks to Facebook for the patch. (Bug #27855592, Bug #90417)

    References: See also: Bug #28950397.

Pluggable Authentication

  • Microsoft Windows: On Windows, MySQL Enterprise Edition distributions now bundle the Cyrus SASL library files libsasl.dll and saslSCRAM.dll so that the LDAP authentication plugins can use the SCRAM-SHA-1 authentication method. (WL #11927)

Plugin Notes

  • Previously, MySQL plugins could be written in C or C++. MySQL header files used by plugins now contain C++ code, which means that plugins must be written in C++, not C. (Bug #87725, Bug #26781567)

Security Notes

  • MySQL Enterprise Edition now provides data masking and de-identification capabilities, implemented as a plugin library containing a plugin and a set of loadable functions. Data masking hides sensitive information by replacing real values with substitutes. MySQL Enterprise Data Masking and De-Identification functions enable masking existing data using several methods such as obfuscation (removing identifying characteristics), generation of formatted random data, and data replacement or substitution. For example:

    mysql> SET @ssn = gen_rnd_ssn();
    mysql> SET @masked_ssn1 = mask_ssn(@ssn);
    mysql> SET @masked_ssn2 = mask_outer(mask_inner (@ssn,4,5,'A'), 3,0,'B');
    mysql> SELECT @ssn, @masked_ssn1, @masked_ssn2;
    +-------------+--------------+--------------+
    | @ssn        | @masked_ssn1 | @masked_ssn2 |
    +-------------+--------------+--------------+
    | 980-31-2838 | XXX-XX-2838  | BBB-AA-2838  |
    +-------------+--------------+--------------+

    For more information, see MySQL Enterprise Data Masking and De-Identification. (WL #7722)

Spatial Data Support

  • Incompatible Change: Previously, ST_Area() supported only geometry arguments having a Cartesian spatial reference system (SRS) and produced an error when invoked with geometry arguments that specified a geographic SRS. ST_Area() now supports geometry arguments having a geographic SRS and returns the geodetic area in square meters.

    Note

    If spatial data contains geometry values that now are interpreted differently by ST_Area(), existing queries using this function will return different results, compared to previous MySQL versions.

    (WL #10726)

  • The parser for spatial reference system (SRS) definitions in CREATE SPATIAL REFERENCE SYSTEM statements now is stricter about rejecting invalid constructs.

    The OGC 01-009 WKT grammar does not make AXIS clauses mandatory in GEOGCS spatial reference system definitions. However, if there are no AXIS clauses, MySQL cannot determine whether a definition has axes in latitude-longitude order or longitude-latitude order. MySQL now enforces the nonstandard requirement that each GEOGCS definition must include two AXIS clauses. One must be NORTH or SOUTH, and the other EAST or WEST. The AXIS clause order determines whether the definition has axes in latitude-longitude order or longitude-latitude order. (Bug #28186073, Bug #28147723)

  • Previously, ST_Validate() supported only geometry arguments having a Cartesian spatial reference system (SRS) and produced an error when invoked with geometry arguments that specified a geographic SRS. ST_Validate() now supports geometry arguments having a geographic SRS.

    Note

    If spatial data contains geometry values that now are interpreted differently by ST_Validate(), existing queries using this function will return different results, compared to previous MySQL versions.

    (WL #11087)

  • MySQL now implements the ST_Transform() spatial function for use in converting geometry values from one spatial reference system (SRS) to another. Currently, it supports conversion between geographic SRSs. For details, see Spatial Operator Functions. (WL #8684)

SQL Syntax Notes

  • Incompatible Change: The deprecated ASC or DESC qualifiers for GROUP BY clauses have been removed. Queries that previously relied on GROUP BY sorting may produce results that differ from previous MySQL versions. To produce a given sort order, provide an ORDER BY clause.

    Queries and stored program definitions from MySQL 8.0.12 or lower that use ASC or DESC qualifiers for GROUP BY clauses should be amended. Otherwise, upgrading to MySQL 8.0.13 or higher may fail, as may replicating to MySQL 8.0.13 or higher slave servers. (WL #8693)

XA Transaction Notes

  • Previously, metadata locks for XA transactions in PREPARED state could be dropped when the client performing the transaction disconnected or the server restarted. This could lead to behavior such as one session being able to drop tables used by an XA transaction in progress for another session. Metadata locks for XA transactions in PREPARED state now are maintained across client disconnects and server restarts, until an XA COMMIT or XA ROLLBACK is executed. (Bug #22710164, Bug #79940, WL #9335)

X Plugin Notes

  • Important Change: X Protocol now provides a connection pooling option, which can reduce overhead for applications that open many connections to MySQL server such as small web pages or a REST API endpoint. Use the new mysqlx.getClient(connection, options) operation, which returns a Client object. By using a Client object, executing an open session operation retrieves an existing and currently unused network connection from the pool, resets it, and uses it. Closing a session marks the underlying connection as unused and returns it to the Client object's connection pool. The connection pool is configured using a options data dictionary, which means that a deployed application can switch from using connection pooling or not by simply changing the connection string. (WL #11824)

  • A mutex lock has been added to prevent conflict between multiple X Plugin client sessions when a large number of sessions are being opened and closed. (Bug #28637947)

  • When an X Plugin client attempted to connect to a MySQL server specifying a database that did not exist on the server, the error message indicated that access was denied rather than that the database was not known. The correct error message is now returned. (Bug #28110957)

  • Incorrect copying of an integer value by X Plugin caused an error relating to misaligned memory access. The issue is now fixed. (Bug #28070946, Bug #90983)

  • Now that X Plugin is loaded and enabled by default, the default verbosity setting for the error log meant that no message was visible to indicate that X Plugin was available on the MySQL server. A message is now issued during system startup to confirm that X Plugin has been loaded. (Bug #27287340)

  • X Protocol now provides the ability to broadcast information to interested clients without the requirement to request the information. In addition, changes related to Group Replication are also broadcast. This work also adds the Mysqlx_notified_by_group_replication and Mysqlx_notice_global_sent status variables. (WL #10655)

Functionality Added or Changed

  • Important Change; NDB Cluster; NDB Client Programs: Removed the deprecated --ndb option for perror. Use ndb_perror to obtain error message information from NDB error codes instead. (Bug #81705, Bug #23523957)

    References: See also: Bug #81704, Bug #23523926.

  • InnoDB: The innodb_fsync_threshold system variable permits defining a threshold, in bytes, for flushing data to disk from the operating system cache. By default, when InnoDB creates a new data file, such as a new log file or tablespace file, the file is fully written to the operating system cache before it is flushed to disk, which can cause a large amount of disk write activity to occur at once.

    Specifying a threshold to force smaller, periodic flushes may be beneficial in cases where multiple MySQL instances use the same storage devices. For example, creating a new MySQL instance and its associated data files could cause large surges of disk write activity, impeding the performance of other MySQL instances that use the same storage devices. Configuring a threshold helps avoid such surges in write activity. (Bug #27724600)

  • InnoDB: User-created temporary tables and internal temporary tables created by the optimizer are now stored in session temporary tablespaces that are allocated to a session from a pool of temporary tablespaces. When a session disconnects, its temporary tablespaces are truncated and released back to the pool. In previous releases, temporary tables were created in the global temporary tablespace (ibtmp1), which did not return disk space to the operating system after temporary tables were dropped.

    The innodb_temp_tablespaces_dir variable defines the location where session temporary tablespaces are created. The default location is the #innodb_temp directory in the data directory.

    The INNODB_SESSION_TEMP_TABLESPACES table provides metadata about session temporary tablespaces.

    The global temporary tablespace (ibtmp1) now stores rollback segments for changes made to user-created temporary tables. (WL #11613)

  • InnoDB: The InnoDB data-at-rest encryption feature now supports general tablespaces. Previously, only file-per-table tablespaces could be encrypted. To support encryption of general tablespaces, CREATE TABLESPACE and ALTER TABLESPACE syntax was extended to include an ENCRYPTION clause.

    The INFORMATION_SCHEMA.INNODB_TABLESPACES table now includes an ENCRYPTION column that indicates whether or not a tablespace is encrypted.

    The stage/innodb/alter tablespace (encryption) Performance Schema stage instrument was added to permit monitoring of general tablespace encryption operations. (WL #9286)

  • Replication: The group_replication_member_expel_timeout option has been added to define the time period that the group should wait for a non-responding member before evicting the member from the group. This enables you to configure the eviction process when the connection to members is unreliable. (WL #11570)

  • Group Replication: You can now inspect and configure the maximum number of consensus instances at any time for a group. This maximum is referred to as the event horizon for a group, and is the maximum number of consensus instances that the system can execute in parallel. This enables you to fine tune the performance of your Group Replication deployment.

    To inspect a group's event horizon value at runtime, issue this statement:

    SELECT group_replication_get_write_concurrency()

    To set the maximum number of write consensus instances, issue the statement shown here:

    SELECT group_replication_set_write_concurrency(instances);

    instances is the maximum number of instances used for consensus.

    See Using Group Replication Group Write Consensus for more information. (Bug #27260096, WL #11512)

  • Group Replication: You can now make changes to the configuration of a group while it is running, without having to stop all members to make changes. This functionality relies on loadable functions which are installed with this version of the plugin, and all members of a group must have these functions installed. To use the functions, connect to an online member and issue SELECT function_name;

    Use the group_replication_set_as_primary() function to trigger the election of a specific member as the new primary in a single-primary group, overriding the usual election process. See Changing the Primary for more information.

    In addition, you can configure the mode which a group is using while it is online, changing between single-primary mode and multi-primary mode. To change the mode of an online group, choose one of these options:

    See Changing the Group Mode for more information. (WL #10378)

  • Solaris: On Solaris, MySQL can now be built with Developer Studio 12.6. (Bug #27055190, Bug #88316, Bug #28165246, Bug #91214)

  • Data truncation testing was rewritten to avoid undefined behavior. (Bug #28255956, Bug #91445)

  • Out-of-range checking for float values was improved. (Bug #28225635)

  • The upgrade check that the server runs during the startup process now verifies that partitioned InnoDB tables do not use shared tablespaces. (Bug #28204431)

  • Previously, file I/O performed in the I/O cache in the mysys library was not instrumented, affecting in particular file I/O statistics reported by the Performance Schema about the binary log index file. Now, this I/O is instrumented and Performance Schema statistics are accurate. Thanks to Yura Sorokin for the contribution. (Bug #27788907, Bug #90264)

  • Performance for locating user account entries in the in-memory privilege structures has been improved. Thanks to Eric Herman for the contribution. (Bug #27772506, Bug #90244)

  • If mysqld --initialize fails to complete but creates an unusable data directory, it now displays a message that the data directory is unusable and can safely be removed. (Bug #27675647)

  • Instrumentation is now provided in the Performance Schema for transaction retries by individual applier threads on a single-threaded or multithreaded slave. Previously, the Performance Schema table replication_applier_status_by_worker displayed information about errors that stopped the applier thread, but not about transient errors that occurred before a transaction was eventually applied. With this information, you can identify transient errors that are causing replication lag on replication slaves or Group Replication group members.

    Eight new columns have been added to the Performance Schema replication_applier_status_by_worker table:

    • LAST_APPLIED_TRANSACTION_RETRIES_COUNT - The number of times the last applied transaction was retried by the worker after the first attempt. If the transaction was applied at the first attempt, this number is zero.

    • LAST_APPLIED_TRANSACTION_LAST_TRANSIENT_ERROR_NUMBER - The error number of the last transient error that caused the transaction to be retried.

    • LAST_APPLIED_TRANSACTION_LAST_TRANSIENT_ERROR_MESSAGE - The message text for the last transient error that caused the transaction to be retried.

    • LAST_APPLIED_TRANSACTION_LAST_TRANSIENT_ERROR_TIMESTAMP - The timestamp in 'YYYY-MM-DD hh:mm:ss[.fraction]' format for the last transient error that caused the transaction to be retried.

    • APPLYING_TRANSACTION_RETRIES_COUNT - The number of times the transaction that is currently being applied was retried until this moment. If the transaction was applied at the first attempt, this number is zero.

    • APPLYING_TRANSACTION_LAST_TRANSIENT_ERROR_NUMBER - The error number of the last transient error that caused the current transaction to be retried.

    • APPLYING_TRANSACTION_LAST_TRANSIENT_ERROR_MESSAGE - The message text for the last transient error that caused the current transaction to be retried.

    • APPLYING_TRANSACTION_LAST_TRANSIENT_ERROR_TIMESTAMP - The timestamp in 'YYYY-MM-DD hh:mm:ss[.fraction]' format for the last transient error that caused the current transaction to be retried.

    The existing column APPLYING_TRANSACTION_START_APPLY_TIMESTAMP is no longer reset each time a transaction is retried. It now retains the timestamp from the worker's first attempt to apply the transaction. (WL #9131)

  • Previously, CREATE TEMPORARY TABLE and DROP TEMPORARY TABLE statements were not supported inside transactions, procedures, functions, or triggers when using GTIDs (that is, when the enforce_gtid_consistency system variable is set to ON). It was possible to use these statements with GTIDs enabled, but only outside of any transaction, and only with autocommit=1.

    From MySQL 8.0.13, this restriction has been removed when binlog_format is set to ROW or MIXED. With row-based logging in use, CREATE TEMPORARY TABLE and DROP TEMPORARY TABLE statements can now be used inside transactions, procedures, functions, or triggers when GTIDs are enabled. When binlog_format is set to STATEMENT, the restriction remains. Because of this difference in behavior, some additional restrictions now apply to changing the binlog_format setting at runtime:

    • If a session has open temporary tables, the replication format cannot be changed for the session (SET @@SESSION.binlog_format).

    • If any replication channel has open temporary tables, the replication format cannot be changed globally (SET @@GLOBAL.binlog_format or SET @@PERSIST.binlog_format).

    • If any replication channel applier thread is currently running, the replication format cannot be changed globally (SET @@GLOBAL.binlog_format or SET @@PERSIST.binlog_format).

    Trying to switch the replication format in any of these cases (or attempting to set the current replication format) results in an error. You can, however, use PERSIST_ONLY (SET @@PERSIST_ONLY.binlog_format) to change the replication format at any time, because this action does not modify the runtime global system variable value, and takes effect only after a server restart.

    When binlog_format is set to ROW or MIXED, CREATE TEMPORARY TABLE and DROP TEMPORARY TABLE statements are not written to the binary log and are therefore not replicated to slaves. When they are used in transactions, if the removal of these statements from the transaction results in an empty transaction, the transaction is not written to the binary log. If a transaction involving these statements is rolled back, a warning message is issued stating that the creation or dropping of the temporary tables could not be rolled back. (WL #8872)

  • The MySQL Server code for binary logging has been refactored to create new internal interfaces for accessing events in a binary log and relay log. The new interfaces separate the write and read processes for binary logging into input and output streams, and decouple the process of capturing and retrieving binary log events from the process of writing them to files. A logical binary log file is used to wrapper the storage layer operations.

    The new internal interfaces make it possible for MySQL Server to use alternative storage methods for binary log events besides a standard binary log or relay log file, including a binary log cache or memory buffer. For example, Group Replication uses the new interfaces to serialize events directly to memory buffers and transaction messages in order to co-ordinate transactions in the group. mysqlbinlog also now uses them to read binary log events from standard input.

    The following existing error messages are now marked as obsolete due to the new internal interfaces:

    • ER_BINLOG_CANT_OPEN_LOG

    • ER_BINLOG_CANT_CREATE_CACHE_FOR_LOG

    • ER_BINLOG_ERROR_GETTING_NEXT_LOG_FROM_INDEX

    • ER_RPL_RECOVERY_ERROR_FREEING_IO_CACHE

    • ER_GRP_RPL_REINIT_OF_INTERNAL_CACHE_FOR_READ_FAILED

    • ER_GRP_RPL_APPENDING_DATA_TO_INTERNAL_CACHE_FAILED

    • ER_GRP_RPL_REINIT_OF_INTERNAL_CACHE_FOR_WRITE_FAILED

    • ER_GRP_RPL_FAILED_TO_CREATE_COMMIT_CACHE

    • ER_GRP_RPL_REINIT_OF_COMMIT_CACHE_FOR_WRITE_FAILED

    (WL #10956)

  • The TempTable storage engine now supports storage of binary large object (BLOB) type columns. This enhancement improves performance for queries that use temporary tables containing BLOB data. Previously, temporary tables that contained BLOB data were stored in the on-disk storage engine defined by internal_tmp_disk_storage_engine. (WL #11452)

  • The MySQL Server code for deserialization of binary log events has been refactored to improve the identification and resilient handling of corrupted and invalid event data during replication, and also during processing with mysqlbinlog, whether or not binary log checksums are used on the server. (WL #11567)

  • Previously, executing RENAME TABLE required that there be no tables locked with LOCK TABLES. Now it is possible to rename tables that are locked with a WRITE lock or that are the product of renaming WRITE-locked tables from earlier steps in a multiple-table rename operation. (WL #9826)

Bugs Fixed

  • InnoDB; Partitioning: Removed old InnoDB handler and partitioning code that referenced .frm files, and thus no longer had any purpose. (Bug #27995316)

  • InnoDB: An assertion was raised during a DROP TABLE operation. A thread that was accessing the table through the memcached API released metadata locks before releasing the table. (Bug #28531148)

  • InnoDB: The being_modified bit in a LOB reference was set but the bit modification was not logged, causing an assertion failure. (Bug #28443837)

  • InnoDB: Window functions returned incorrect results when the optimizer used the InnoDB storage engine for internal temporary tables. (Bug #28430650)

  • InnoDB: Adjusting the server time to an earlier time caused periodic redo flushes to be missed. (Bug #28430358, Bug #90670)

  • InnoDB: An ALTER TABLE operation that added a primary key produced a segmentation fault. (Bug #28395278)

    References: This issue is a regression of: Bug #27753193.

  • InnoDB: A conditional check was removed by removing the ReadView::complete() function and splitting its work among other functions. This change helps optimize performance on ARM 64-bit. (Bug #28385211, Bug #91759)

  • InnoDB: Leftover thread_mutex code was removed from InnoDB source code files. (Bug #28363673, Bug #91678)

  • InnoDB: Type changes were implemented to eliminate warnings that occurred when compiling InnoDB with Microsoft Visual Studio 2017. (Bug #28338720)

  • InnoDB: An invalid assertion was raised when a B-tree flag used to mark shared index locks was used to mark a shared-exclusive index lock. (Bug #28317172)

  • InnoDB: The sharp checkpoint mechanism no longer forces preflushing of dirty pages when requesting a checkpoint for the currently available LSN.

    The log checkpointer thread now takes the concurrency margin (the per thread margin for free space in the log) into account when determining if the next checkpoint write is required and whether to wake up page cleaners to force a sync-flush of dirty pages. Page cleaner threads take the concurrency margin into account when determining whether to flush dirty pages and how many pages to flush. (Bug #28297462)

  • InnoDB: A misplaced debug crash point caused a transaction timeout resulting in test failures. (Bug #28295814)

  • InnoDB: InnoDB error message format was modified to remove duplicate text. (Bug #28289789)

  • InnoDB: Unnecessary cycles of freeing and allocating memory caused JSON performance degradation on Windows. (Bug #28278737)

  • InnoDB: InnoDB incorrectly permitted dropping an index used in a foreign key constraint using an in-place ALTER TABLE statement. This operation is now blocked at the SQL layer. (Bug #28268875)

  • InnoDB: To avoid checking hardware support each time a hardware-optimized checksum is computed, asserts were converted to debug-only asserts. (Bug #28267334, Bug #91485)

  • InnoDB: A patch that combined Contention-Aware Transaction Scheduling (CATS) with functionality that releases read locks caused gap locks to be removed without granting locks to waiting transactions, resulting in transaction timeouts. (Bug #28261530)

    References: This issue is a regression of: Bug #28261530.

  • InnoDB: The log_checkpointer thread failed to write new checkpoints in a timely manner when the amount of redo was small. (Bug #28220222)

  • InnoDB: The server exited during an in-place upgrade from MySQL 5.7 to MySQL 8.0 due to an attempted eviction of a foreign-key-related table from the cache. At the end of the upgrade process, tables with FULLTEXT indexes were marked as ready for eviction without checking for foreign key relationships. (Bug #28212734, Bug #91325)

  • InnoDB: The format of the following Performance Schema and INFORMATION_SCHEMA table columns was modified:

    The previous format was trx_id:table_id for table locks and trx_id:space_id:page_no:heap_no for record locks. The new format is trx_immutable_id:table_id:lock_immutable_id for table locks and trx_immutable_id:space_id:page_no:heap_no:lock_immutable_id for record locks.

    lock_immutable_id and trx_immutable_id are 64-bit values that do not change during the lifetime of a lock or transaction, respectively, and are unique among other instance object IDs. (Bug #28176910)

  • InnoDB: The list of permitted lock mode descriptors used by the LOCK_MODE column of the Performance Schema data_locks table was expanded to include REC_NOT_GAP, INSERT_INTENTION, PREDICATE, and PRDT_PAGE. REC_NOT_GAP indicates a record-only lock. INSERT_INTENTION indicates an insert intention lock. PREDICATE and PRDT_PAGE descriptors indicate a spatial index lock. (Bug #28176805)

  • InnoDB: Table names were not compared in lowercase on macOS with a setting of lower_case_table_names=2, which caused instability after restarting the server. (Bug #28170699, Bug #91204)

  • InnoDB: Macros used to define constant values in InnoDB source code were changed to constant expressions. (Bug #28152926)

  • InnoDB: A flag that prevents transactions from being rolled back during commit is now set earlier to prevent scenarios in which a high priority transaction could abort a transaction that is in the process of being committed. (Bug #28140462)

  • InnoDB: A query that scanned the primary key of a table did not return the expected result. (Bug #28104394, Bug #91032)

  • InnoDB: Unnecessary header file inclusions were removed from InnoDB source code files. (Bug #28086759)

  • InnoDB: An apparent hang due to the log writer running out of free space in the redo log caused the server to exit. (Bug #28072385, Bug #90993)

  • InnoDB: A query interruption during a lock wait caused an error. (Bug #28068293)

  • InnoDB: After upgrading from MySQL 5.7 to MySQL 8.0, invalid warnings indicated that undo tablespace IDs were not in the undo tablespace range. The warnings occurred if the MySQL 5.7 installation was configured to use separate undo tablespaces. (Bug #28060337)

  • InnoDB: Error messaging was improved for startup failures on an incomplete cloned data directory. (Bug #28032131)

  • InnoDB: A segmentation fault occurred during an XA COMMIT operation. (Bug #27995891)

  • InnoDB: Unused code was removed from the TempTable storage engine source code. (Bug #27978968)

  • InnoDB: An index record was not found when updating a secondary index defined on a generated column. (Bug #27968952)

  • InnoDB: A spurious negation operator in an IF condition caused the Contention-Aware Transaction Scheduling (CATS) algorithm to be used for table locks. (Bug #27944920)

    References: This issue is a regression of: Bug #27572937.

  • InnoDB: The update log applied as part of an online ALTER TABLE operation did not take into account the computed value of the generated column in the old row while updating the secondary index. (Bug #27921932)

  • InnoDB: Buffer pool memory allocation was not fully accounted for in Performance Schema memory/innodb/buf_buf_pool statistics. (Bug #27917595, Bug #90561)

  • InnoDB: An unsupported DDL operation involving a foreign key constraint raised an assertion. (Bug #27912873)

  • InnoDB: A lock-related debug assertion failure was raised when more than one lock matched a search condition, which could lead to releasing the wrong lock. (Bug #27898384)

  • InnoDB: A function that removes aborted indexes during the prepare phase of an online ALTER TABLE operation did not record its changes. (Bug #27879325)

  • InnoDB: A general tablespace created in MySQL 5.7 with no assigned table caused a failure when upgrading to MySQL 8.0. (Bug #27877485)

  • InnoDB: Concurrent undo tablespace truncation and master key rotation operations raised an assertion. (Bug #27872369)

  • InnoDB: A boolean marker identifying whether a transaction holds a mutex was not placed in the correct location. (Bug #27870035)

  • InnoDB: An attempted foreign key check on a discarded table caused a segmentation fault. (Bug #27804668)

  • InnoDB: B-tree bulk load operations could leave a page in a partially initialized state. (Bug #27802098)

  • InnoDB: Starting the server inside a Docker container on a NUMA enabled operating system raised an mbind: Operation not permitted error. (Bug #27792853)

  • InnoDB: A partitioned table TABLE_ID value stored in a storage-engine-private data field in the data dictionary was not adjusted properly after an ALTER TABLE ... PARTITION operation. (Bug #27784462)

  • InnoDB: The server halted with a log writer overwriting data after checkpoint - waited too long error. (Bug #27779266)

  • InnoDB: With innodb_flush_log_at_trx_commit=2, the log_flusher thread could wait for an event with a timeout period equal to the innodb_flush_log_at_timeout setting, causing an initialization delay. (Bug #27762596)

  • InnoDB: An assertion was raised during an OPTIMIZE TABLE operation. (Bug #27753193)

  • InnoDB: Transaction rollback due to a deadlock caused an assertion failure in debug builds. Initiation of an attachable transaction for accessing the data dictionary was not expected during transaction rollback. (Bug #27729974)

  • InnoDB: With innodb_flush_log_at_trx_commit=0 and binary logging enabled, redo logs were not flushed as expected during the commit phase of DDL operations. (Bug #27691035)

  • InnoDB: With REDUNDANT or COMPRESSED row format and READ COMMITTED isolation level, only the LOB value prefix and possibly the external part of an old LOB value were returned, which could cause a JSON document to be viewed as corrupted. If there was no LOB value prefix, an old LOB value with new values for other fields could be returned, causing data inconsistency. (Bug #27624990)

  • InnoDB: A debug option that permits pausing periodic checkpoints became obsolete after periodic checkpoint ownership was moved from the master thread to a log checkpointer thread. The debug option was replaced by another method of pausing periodic checkpoints. (Bug #27588328)

  • InnoDB: A transaction on a table with a spatial index defined on a column with a spatial reference identifier (SRID) was able to insert into an area selected for update by another transaction. (Bug #27577612)

  • InnoDB: A foreign key constraint name was duplicated during a rename table operation, causing a failure during later query execution. (Bug #27545888)

  • InnoDB: A serialized dictionary information (SDI) deletion failure raised an assertion. (Bug #27493634)

  • InnoDB: A server exit occurred after freeing large object (LOB) index entries during a LOB purge or rollback. (Bug #27419474)

  • InnoDB: In a function called before the execution of a statement in a stored procedure, a read and write operation on trx->lock.start_stmt was not protected by a mutex. (Bug #27325898)

  • InnoDB: The INFORMATION_SCHEMA.FILES and INFORMATION_SCHEMA.INNODB_TABLESPACES tables did not show the actual undo tablespaces that were present in the MySQL instance. Only the two default undo tablespaces were shown. (Bug #26820406)

  • InnoDB: An error occurred during a DDL operation due to a mismatch in a REDUNDANT row format calculation that determines the length of the online log. (Bug #26375771)

  • InnoDB: Enabling innodb_undo_log_truncate negatively affected transaction processing performance. Instead of performing two checkpoints during an undo tablespace truncate operation, pages that belong to the tablespace file are now flushed from disk. (Bug #26322656)

  • InnoDB: A helper class was introduced to improve performance associated with reading from secondary keys when there are multiple versions of the same row.

    Thanks to Domas Mituzas for the contribution. (Bug #25540277, Bug #84958)

  • InnoDB: The location of the Innodb Merge Temp File that reported by the wait/io/file/innodb/innodb_temp_file Performance Schema instrument was incorrect. (Bug #21339079, Bug #77519)

  • Partitioning: When a CREATE TABLE ... PARTITION BY ... statement failed due to an invalid partition definition, the server did not remove any partition files which might have been created prior to encountering the invalid PARTITION clause. (Bug #27798708)

    References: See also: Bug #88043, Bug #26945644.

  • Partitioning: It was possible to perform FLUSH TABLES FOR EXPORT on a partitioned table created with innodb_file_per_table=1 after discarding its tablespace. Attempting to do so now raises ER_TABLESPACE_DISCARDED. (Bug #90545, Bug #27903881)

    References: See also: Bug #80669, Bug #22899690.

  • Partitioning: An extraneous row lock was imposed by an update to a partitioned InnoDB table. (Bug #87253, Bug #26553164)

  • Replication: When a replication slave is restarted by a START SLAVE statement, the columns in the Performance Schema table replication_applier_status_by_worker beginning APPLYING_TRANSACTION are now reset on a slave that is operating in single-threaded mode. These columns were always reset on a multi-threaded slave, because the existing worker threads were terminated by the statement and the information could not be retained. The behavior has now been standardized across the slave configurations by resetting the columns for a single-threaded slave as well. (Bug #28248026)

  • Replication: If a multi-threaded replication slave was stopped, changed to a single-threaded slave (by setting slave_parallel_workers > 0), and restarted, the Performance Schema table replication_applier_status_by_worker showed irrelevant timestamps because the old monitoring information had not been cleared. (Bug #28191382)

  • Replication: When the binlog_group_commit_sync_delay system variable is set to a wait time to delay synchronization of transactions to disk, and the binlog_group_commit_sync_no_delay_count system variable is also set to a number of transactions, the MySQL server exits the wait procedure if the specified number of transactions is reached before the specified wait time is reached. The server manages this process by checking on the transaction count after a delta of one tenth of the time specified by binlog_group_commit_sync_delay has elapsed, then subtracting that interval from the remaining wait time.

    If rounding during calculation of the delta meant that the wait time was not a multiple of the delta, the final subtraction of the delta from the remaining wait time would cause the value to be negative, and therefore to wrap to the maximum wait time, making the commit hang. The data type for the remaining wait time has now been changed so that the value does not wrap in this situation, and the commit can proceed when the original wait time has elapsed. Thanks to Yan Huang for the contribution. (Bug #28091735, Bug #91055)

  • Replication: In debug builds, an assertion failed because more than 255 collations are now available in MySQL. (Bug #28015761)

  • Replication: An assertion was raised in debug builds because the MySQL server recorded a GTID consistency violation, but did not remove the record after the relevant statement failed to execute successfully. The handling of this situation has now been improved to ensure that the server checks at the end of a transaction whether a GTID consistency violation was produced by a failed statement, and if this is the case, restores the previous GTID consistency state. (Bug #27903831, Bug #90551)

  • Replication: With GTIDs in use for replication, transactions including statements that caused a parsing error (ER_PARSE_ERROR) could not be skipped manually by the recommended method of injecting an empty or replacement transaction with the same GTID. This action should result in the slave identifying the GTID as already used, and therefore skipping the unwanted transaction that shared its GTID. However, in the case of a parsing error, because the statement was parsed before the GTID was checked to see if it needed to be skipped, the replication applier thread stopped due to the parsing error, even though the intention was for the transaction to be skipped anyway.

    With this fix, the replication applier thread now ignores parsing errors if the transaction concerned needs to be skipped because the GTID was already used. Note that this behavior change does not apply in the case of workloads consisting of binary log output produced by mysqlbinlog. In that situation, there would be a risk that a transaction with a parsing error that immediately follows a skipped transaction would also be silently skipped, when it ought to raise an error. (Bug #27638268)

  • Replication: When a RESET SLAVE statement was issued on a replication slave with GTIDs in use, the existing relay log files were purged, but the replacement new relay log file was generated before the set of received GTIDs for the channel had been cleared. The former GTID set was therefore written to the new relay log file as the PREVIOUS_GTIDS event, causing a fatal error in replication stating that the slave had more GTIDs than the master, even though the gtid_executed set for both servers was empty. Now, when RESET SLAVE is issued, the set of received GTIDs is cleared before the new relay log file is generated, so that this situation does not occur. (Bug #27636289)

  • Replication: The master's receiver thread for semisynchronous replication held a mutex while reading acknowledgements from slaves, but the same mutex was required to add or remove a semisynchronous slave, causing those operations to be delayed by the acknowledgement activity. The issue has now been fixed by not acquiring the mutex to read the acknowledgements from slaves. (Bug #27610678, Bug #89370)

  • Replication: In code for replication slave reporting, a rare error situation raised an assertion in debug builds, but in release builds, returned leaving a mutex locked. The mutex is now unlocked before returning in this situation. Thanks to Zsolt Parragi for the patch. (Bug #27448019, Bug #89421)

  • Replication: Automatic retrying of transactions on a replication slave, as specified by the slave_transaction_retries system variable, was taking place even if the transaction had a non-temporary error that would repeat on retrying or that indicated wider issues. Now, transactions are only automatically retried if there is either no error, or an error that is only temporary. (Bug #27373559, Bug #89143)

  • Replication: When FLUSH statements for specific log types (such as FLUSH SLOW LOGS) resulted in an error, the statements were still written to the binary log. This stopped replication because the error had occurred on the master, but did not occur on the slave. MySQL Server now checks on the outcome of these FLUSH statements, and if an error occurred, the statement is not written to the binary log. (Bug #24786290, Bug #83232)

  • Replication: The PASSWORD() function, which produces a hash of the password, was deprecated in MySQL 5.7 and removed in MySQL 8.0. When a SET PASSWORD statement that used this function was replicated from a MySQL 5.6 master to a MySQL 5.7 slave, or from a MySQL 5.7 master with the log_builtin_as_identified_by_password system variable set to ON to a MySQL 5.7 slave, the password hash was itself also hashed before being stored on the slave. The issue has now been fixed and the replicated password hash is stored as originally passed to the slave. (Bug #24687073)

  • Replication: If an ORDER BY clause was used in retrieving records from certain Performance Schema tables relating to replication, an empty set was returned. The issue has now been fixed. (Bug #22958077, Bug #80777)

  • Replication: When replication channels are used on a slave for multi-source replication, a START SLAVE statement that does not specify an individual channel (so without the FOR CHANNEL clause) should start the I/O thread and the SQL thread for all of the channels on the replication slave. However, if a RESET SLAVE statement was used on such a slave, a subsequent START SLAVE statement did not start the non-default channels. Now, replication channels that are deinitialized as a result of a RESET SLAVE statement, rather than as a result of an error in the initialization process, are identified and are restarted by a START SLAVE statement that applies to all channels. (Bug #22809607)

  • Replication: Issuing RESET SLAVE on a replication slave does not change any replication connection parameters such as master host, master port, master user, or master password, which are retained in memory. However, these connection parameters are reset if you issue RESET SLAVE ALL. Previously, if the slave mysqld was restarted immediately after issuing RESET SLAVE (including a server crash as well as a deliberate restart), the connection parameters were reset as if RESET SLAVE ALL had been used.

    Now, when master_info_repository=TABLE is set on the server (which is the default from MySQL 8.0), replication connection parameters are preserved in the crash-safe InnoDB table mysql.slave_master_info as part of the RESET SLAVE operation. They are also retained in memory. In the event of a server crash or deliberate restart after issuing RESET SLAVE but before issuing START SLAVE, the replication connection parameters are retrieved from the table and reused for the new connection.

    If master_info_repository=FILE is set on the server (which is the default in MySQL 5.7), replication connection parameters are only retained in memory, so the behavior remains the same as previously. If the slave mysqld is restarted due to a server crash or a deliberate restart immediately after issuing RESET SLAVE, the connection parameters are lost. In that case, you must issue a CHANGE MASTER TO statement after the server start to respecify the connection parameters before issuing START SLAVE.

    If you want to reset the connection parameters intentionally, you need to use RESET SLAVE ALL, which clears the connection parameters. In that case, you must issue a CHANGE MASTER TO statement after the server start to specify the new connection parameters. (Bug #20280946)

  • Replication: When a group member resumes after being suspended for some time and is not able to process all pending messages, it enters the ERROR state. However, the remaining members see it as UNREACHABLE, and wait until the member's suspicion expires to evict it from the group. The behavior has now been modified and a member stopping due to some error tries to connect to a known peer to request its removal from the group, before installing the leave view. (Bug #91433, Bug #28252687)

  • Replication: Compilation warnings related to unused functions in xdr_utils have been reduced. Thanks to Zsolt Parragi for the patch. (Bug #91071, Bug #28099963)

  • Replication: If the group_replication_recovery_retry_count variable was modified while the member was already making a reconnection attempt, the connection attempt could enter an infinite loop. (Bug #91057, Bug #28092714)

  • Replication: The group_replication_exit_state_action variable enables you to specify what action is taken if a member involuntarily leaves the group, but when starting a server with group_replication_start_on_boot enabled the group_replication_exit_state_action variable was being ignored during the following scenarios:

    • valid number of group members was exceeded

    • incompatible configuration of the member system variables (various)

    • the joining member had more transactions than the group

    • the joining member's version was not compatible with the group

    (Bug #90494, Bug #27881311)

  • Replication: Issuing START GROUP_REPLICATION and then forcibly stopping the mysqld process, for example using control-C, could result in an unexpected halt of the server. (Bug #90457, Bug #27873419)

  • Group Replication: Entries in the relay log info log for the group_replication_applier and group_replication_recovery channels were not cleared by RESET SLAVE or RESET SLAVE ALL. (Bug #27411175)

  • Group Replication: When Group Replication was started on server with an invalid group_replication_group_name, the server stopped unexpectedly. (Bug #91347, Bug #28219136)

  • Group Replication: When group_replication_group_seeds contained a DNS based entry which resolved to its own local address, Group Replication was unable to start. (Bug #90483, Bug #27882096, Bug #28074929)

  • Microsoft Windows: An error now is written to the server log when the presence of the NO_AUTO_CREATE_USER value for the sql_mode option in the options file prevents a MySQL 8.0 server from starting. (Bug #28061945, Bug #90967)

  • Microsoft Windows: On Windows, uninstallation of the MySQL Server MSI package through MySQL Installer produced a spurious popup window. (Bug #27463864)

  • Microsoft Windows: On Windows, DBUG_ABORT did not print the custom stack trace and other information. (Bug #21383530)

  • Microsoft Windows: Starting MySQL as a Windows service with the service-installation command that specified a service name following the --install option ignored the directives in the named service group of the my.ini or my.cnf options files and used default options instead. Only the default service names (mysqld, mysql_cluster, server, mysqld-8.0) could load different parameters from an options file. (Bug #90383, Bug #27852209)

  • JSON: The server did not reject creation of a table with a generated column in which the generated column used JSON_TABLE(), even though subqueries, parameters, variables, stored functions, and loadable functions are not permitted in expressions for generated columns. The server now checks more aggressively to make sure any of the disallowed constructs (including JSON_TABLE()) are rejected for use in such expressions. (Bug #28518485)

  • JSON: SELECT ... FROM JSON_TABLE() sometimes failed with a permissions error for a user other than MySQL root. This issue could also occur when such a query was used as the basis for a view, and a SELECT from the view failed. (Bug #28255453, Bug #27923406)

    References: See also: Bug #27189940.

  • JSON: The JSON_TABLE() function subjected integer values greater than or equal to 231 to wraparound. For example, the query SELECT id FROM JSON_TABLE('[{"id":"2147483648"}]', '$[*]' COLUMNS (id BIGINT UNSIGNED PATH '$.id')) AS json returned -2147483648. (Bug #27856835)

  • JSON: In some contexts, the NULLIF() function returned its first argument as a boolean value rather than its actual type. This was noticed when the result of this function was used as an argument to JSON_ARRAYAGG() or JSON_OBJECTAGG(), but could have occurred in other cases in which NULLIF() was used in a similar fashion. (Bug #90833, Bug #28007237)

  • JSON: When a JSON document which contained binary data was converted to base-64 encoded text for display, newline characters in the encoded string were not properly escaped, so that the text representation could not be parsed as JSON, and was thus truncated, corrupted, or both. Now MySQL makes sure that any newline characters in the encoded string are escaped. (Bug #90503, Bug #27891359)

  • filesort operations could cause a server exit. (Bug #28791531, Bug #92777)

  • Debian packaging now supports Ubuntu 18.10 (Cosmic Cuttlefish). (Bug #28765706)

  • On the Fedora 29 platform, OpenSSL 1.0.x is used to build packages because OpenSSL 1.1.1 support is not ready. If you build MySQL from source, it is recommended that you build using the compat-openssl10-devel package. (Bug #28737143)

  • On the Fedora 29 platform, upgrading from MariaDB to MySQL 8.0.13 failed due to missing obsoletes. (Bug #28727698)

  • MySQL binary distributions for SLES 12 now are built using GCC 7. The lowest supported GCC version on this platform is now 5.3 (previously 4.8.5). (Bug #28542723)

  • Trying to add a functional index on a subquery should not be possible, and caused the server to hit an assertion when trying to resolve the indexed expression. Now in such cases, the expression is disallowed, and the servers returns an appropriate error message. (Bug #28526493)

  • Added a range check when performing calculations with exponents. (Bug #28505423)

  • CMAKE -DWITHOUT_SERVER=1 resulted in build errors. (Bug #28501563, Bug #92011)

  • When a prefix index was specified with a length of 8 bytes, the Sub_part column in the output of SHOW INDEXES was NULL. (Bug #28499603)

  • Compilation failed for GCC 8 with MySQL configured to use some system libraries. (Bug #28471072, Bug #91914)

  • For debug builds, if the server bootstrapping phase failed, missing cleanup code caused an assertion to be raised. (Bug #28435378, Bug #91847)

  • Numeric ranges in MySQL builds could differ between the ARM and Intel x64 platforms. (Bug #28401869)

  • For mysqldump --tables output, file names now always include a .txt or .sql suffix, even for file names that already contain a dot. Thanks to Facebook for the contribution. (Bug #28380961, Bug #91745)

  • Concurrent INSERT and SELECT statements on a MERGE table could result in a server exit. (Bug #28379285)

  • On SLES 15, upgrading from MariaDB packages to MySQL packages failed due to incorrect obsoletes information. (Bug #28292138)

  • The data dictionary auto releaser now allocates maps for data dictionary object types dynamically. Also, maps were added for column statistics and resource groups used in size calculations and object removal, and an auto releaser was added for ANALYZE TABLE operations. (Bug #28245522, Bug #91420)

  • It was possible for a UNION query that mixed different character sets to fail with a spurious error. (Bug #28237675)

    References: This issue is a regression of: Bug #83895, Bug #25123839.

  • CHECK TABLE ... FOR UPGRADE on temporary tables could raise an assertion. (Bug #28220374)

    References: This issue is a regression of: Bug #24741307.

  • A divide-by-zero error could occur in the range optimizer. (Bug #28214186)

  • Due to a GCC 8 bug, recursion in foreign key checks could exhaust stack space. (Bug #28200774, Bug #28421040, Bug #91823)

  • When sql_mode was set to the empty string, queries of the form SELECT CONCAT( FORMAT(LPAD(char,2,''), 1) ) were not handled properly; the same was also true for similarly-formed queries using RPAD() in place of LPAD(). (Bug #28197977)

  • Non-privileged users could change their own account password history and reuse properties. (Bug #28191838)

  • The OpenSSL 1.1 OPENSSL_malloc_init() call did not work well on Windows. To address this, alternative allocation wrappers are used instead of the defaults from OpenSSL. Consequently, OpenSSL memory instrumentation is turned off. (Bug #28179051)

  • Allocation of certain object types during data dictionary lookups resulted in a stack buffer overflow. (Bug #28176453)

  • An internal server operation that attempted to perform a commit while fetching table statistics from the INFORMATION_SCHEMA could raise an assertion. (Bug #28165060)

  • Certain stored procedures could cause a server exit. (Bug #28156802)

  • Output for CREATE USER statements could differ in the general query log and audit log. (Bug #28147710)

  • For UPDATE and DELETE statements that produce an error due to sql_safe_updates being enabled, the error message was insufficiently informative. The message now includes the first diagnostic that was produced, to provide information about the reason for failure. For example, the message may indicate that the range_optimizer_max_mem_size value was exceeded or type conversion occurred, either of which can preclude use of an index.

    Additionally: (1) Using EXPLAIN for such statements does not produce an error, enabling users to see from EXPLAIN plus SHOW WARNINGS output why an index is not used. (2) For multiple-table deletes and updates, an error is produced with safe updates enabled only if any target table uses a table scan. (Bug #28145710, Bug #91080)

  • MySQL Server and test RPM packages were missing perl-Data-Dumper as a dependency. (Bug #28144933, Bug #72926)

  • Server startup could fail if a service needed by a component was not yet initialized. (Bug #28142250)

  • SHOW CREATE TABLE could omit foreign key RESTRICT options. This in turn could cause foreign key RESTRICT options to be lost from tables dumped with mysqldump and restored from the dump file. (Bug #28122781, Bug #91110)

  • The mysql client was slow to import large multiple-line statements in batch mode. Memory allocation is now more efficient for this situation. Thanks to Sinisa Milivojevic for the patch. (Bug #28116512, Bug #85155)

  • The SUM() and AVG() functions did not handle correctly a string argument used with a window function. (Bug #28105241)

  • Excessive nesting of geometry collections caused the server to exhaust stack space. The server now raises an error if there is a danger of stack overrun. (Bug #28100563)

  • For the mysql client, the -b short option was associated with two long options, --no-beep and --binary-as-hex. The -b option now is associated only with --no-beep. (Bug #28093271)

  • For a table with an auto-increment primary key, concurrent ALTER TABLE ... ADD ... VIRTUAL and INSERT statements could lead to duplicate-key errors. (Bug #28089240)

  • Handling of floating-point values by SUM() was improved. (Bug #28080199)

  • Some expressions using ST_Simplify() could cause a server exit. (Bug #28079969)

  • In builds with Undefined Behavior Sanitizer enabled, testing returned a member call on null pointer of type 'struct Event_db_repository' error. (Bug #28066155)

  • The WITH_GMOCK CMake option did not handle Windows path names properly. (Bug #28061409, Bug #90964)

  • INFORMATION_SCHEMA queries that attempted to cache table statistics could raise a debugging assertion. (Bug #28035207)

  • If roles were assigned to accounts, certain sql_mode settings could cause the server to be unable to start. (Bug #28030423)

  • Group lookups for LDAP authentication plugins could fail if the user had insufficient privileges. Now, group search operations bind again using root credentials if those are available. (Bug #28016008)

  • ANALYZE TABLE ... UPDATE HISTOGRAM statements produced by mysqldump contained a syntax error. (Bug #28014376, Bug #90846)

  • Generated columns having indexes and that used a string function were not always populated correctly. (Bug #27973409)

  • Fixed-length sort keys, such as those used by priority queues, that fit exactly were assumed to have failed. (Bug #27970481, Bug #92448, Bug #28654343)

  • For an empty result, REPLACE() sometimes returned a null string rather than an empty string. (Bug #27960921)

  • Dropping a table that was created with a user-defined collation no longer available could cause a server exit. (Bug #27952999)

  • Joining the INFORMATION_SCHEMA REFERENTIAL_CONSTRAINTS and TABLE_CONSTRAINTS tables on the CONSTRAINT_NAME failed because the column collations differed. (Bug #27945704, Bug #90690)

  • Some operations on DECIMAL values could cause a server exit. (Bug #27942277)

  • Very long table keys were handled incorrectly on replication slaves. (Bug #27930505)

  • mysqld did not determine its installation directory correctly if that directory was the last one listed in $PATH. (Bug #27922896)

  • During server startup/shutdown, PID files could be mishandled. (Bug #27919254)

  • When the server was started with --skip-grant-tables, some account-management statements were not disabled. (Bug #27906226)

  • On rare occasions, setting a savepoint could raise an assertion. (Bug #27898591)

  • Runtime errors could occur for calls to copy_integer(). myfunc_int(), mysql_sys_var_int(), or thd_killed(); for calls to MyISAM sorting functions; or for values outside the range of representable values of type unsigned long long. (Bug #27894901, Bug #90515, Bug #27871951, Bug #27918095, Bug #90609, Bug #27937522, Bug #90661, Bug #27978325, Bug #27962900)

  • mysql_install_plugin failed to report plugin-specific errors if the server was unable to store plugin metadata in the data dictionary. (Bug #27893406)

  • ssl_fips_mode cannot be set to ON unless the OpenSSL requirements are satisfied, but a failed attempt still resulted in the value being displayed as ON. (Bug #27891890)

  • The symbol for the mysql_result_metadata() C API function was not exported by the client library on Windows. (Bug #27868095)

  • For InnoDB tables, self-referencing foreign keys could be created, causing server misbehavior. (Bug #27864515)

  • A UNION query mixing different character sets produced invalid output when performed in a prepared statement or a stored procedure. (Bug #27849293)

  • Debian packages were missing a dependency for libcurl-dev. (Bug #27844465)

  • --help output for client programs did not include the current --ssl-fips-mode value. (Bug #27838966)

  • A runtime error could occur for calls to Derived_key_comp(). (Bug #27830679)

  • For MEMORY tables, memory overflow errors could occur. (Bug #27799513)

  • When converting from a BLOB (or TEXT) type to a smaller BLOB (or TEXT) type, no warning or error was reported informing about the truncation or data loss. Now an appropriate error is issued in strict SQL mode and a warning in nonstrict SQL mode. (Bug #27788685, Bug #90266)

  • For debug builds, my_strnxfrm_unicode_full_bin() could raise an assertion. (Bug #27752619)

  • In builds with Undefined Behavior Sanitizer enabled, signed integer overflow could occur in GIS calculations. (Bug #27751479, Bug #27744399, Bug #27811282)

  • It was possible in some cases for the internal function regexp::EvalExprToCharset() to bind a reference to a misaligned address. (Bug #27743722)

  • The severity of messages produced by the server about being unable to read key files has been escalated from INFORMATION to WARNING. (Bug #27737195)

  • No temporary table field was created for an expression using window functions combined with a rollup, causing queries containing these to fail. (Bug #27735167)

  • The expression FIND_IN_SET( JSON_UNQUOTE( JSON_SET('{}','$','') ), 1) was not evaluated properly. (Bug #27731699)

  • Failure to create a temporary table during a MyISAM query could cause a server exit. Thanks to Facebook for the patch. (Bug #27724519, Bug #90145)

  • parser_max_mem_size was ineffective when parsing stored program definitions. (Bug #27714748)

  • A BETWEEN clause comparing negative values could lead to erroneous results. (Bug #27691347)

  • Some typos in server error messages were fixed. Thanks to Thomas Tsiakalakis for the contribution. (Bug #27688294, Bug #90048)

  • These scripts are no longer included in RPM packages (they are unnecessary because they are compiled into the mysqld binary): fill_help_tables.sql, mysql_sys_schema.sql, mysql_system_tables.sql, mysql_system_tables_data.sql, mysql_system_users.sql. (Bug #27672991)

  • Unique::io_cache was not cleaned up properly when restarting an index merge operation, leading to an assertion in debug builds. (Bug #27599292)

  • Using a binary or hexadecimal literal value in a ENUM column resulted in a string type with bytes not permitted by the data dictionary character set. (Bug #27592803)

  • It was possible for CONVERT() to fail under the following conditions:

    The size of the result of the conversion was greater than max_allowed_packet

    sql_mode was '' (that is, empty).

    The character set of the result differed from character_set_connection (Bug #27592714)

  • Use of ENUM or SET fields with JSON_TABLE() could result in unexpected behavior. (Bug #27571251)

  • Host name resolution errors could cause the audit_log plugin to fail. (Bug #27567003)

  • Unsuccessful connection attempts were not being written to the error log when log_error_verbosity=3. (Bug #27539838)

  • Persisted system variables with a value of NULL were written to mysqld-auto.cnf incorrectly. (Bug #27512616)

  • An earlier code cleanup caused FEDERATED storage engine failures. (Bug #27493633, Bug #89537)

    References: This issue is a regression of: Bug #25943754.

  • Setting max_execution_time sometimes had no effect when used with full-text search. (Bug #27155294)

  • An attempted read of an uncommitted transaction raised an assertion. (Bug #26876608)

  • ALTER TABLE ... REORGANIZE PARTITION ... could result in incorrect behavior if any partition other than the last was missing the VALUES LESS THAN part of the syntax. (Bug #26791931)

  • In some cases when ER_DATA_TOO_LONG was the appropriate error, ER_TRUNCATED_WRONG_VALUE_FOR_FIELD was reported instead. (Bug #26475237, Bug #87100)

  • A last_checked_for_upgrade column that stores the current server version number was added to the mysql.tables data dictionary table to track successful CHECK TABLE ... FOR UPGRADE operations. This data was previously stored in .frm files to avoid rechecking tables unnecessarily. (Bug #24741307, Bug #83169)

  • The MYSQLparse() function generated from the sql_yacc.yy file had exceeded a size limit that caused the Microsoft Visual C++ compiler to abandon optimization. The /d2OptimizeHugeFunctions compiler switch was enabled to restore optimization (available in Visual Studio 2017 and higher). (Bug #24600301)

  • A range check for the product of a signed and unsigned integer could be performed incorrectly. (Bug #17081376, Bug #69715)

  • Specifying the maximum possible value for a YEAR column failed when expressed as a real constant such as 2155.0E00 or 2.15E3. (Bug #91226, Bug #28172538)

  • Prevented potential member access within a null pointer of type SELECT_LEX. (Bug #90689, Bug #27945614)

  • The base used for calculation of line numbers reported in warnings generated by SELECT statements has been changed from the number of input rows read to the actual row numbers seen by the user. For statements using GROUP BY, this number is now based on the number of aggregate groupings. (Bug #90667, Bug #27938507)

  • Estimates contained by histograms for the number of distinct values in each bucket, which can be employed to estimate the number of distinct values in a table, can now be used by the Optimizer's histograms::Histogram interface. (Bug #90465, Bug #27872281)

  • Selecting from the INFORMATION_SCHEMA.KEYWORDS table failed when no default database had been selected. (Bug #90160, Bug #27729859)

  • ANALYZE TABLE UPDATE HISTOGRAM failed to return an error condition to the client when lock_wait_timeout was exceeded during statement processing. (Bug #89738, Bug #27574375)

  • It was possible to create a table t having a generated column g and then to execute ALTER TABLE t ALTER COLUMN g SET DEFAULT ...; attempting to set the default for a generated column in this fashion now fails with Error 1221 Incorrect usage of DEFAULT and generated column. This is the same behavior as for a similar CREATE TABLE statement. (Bug #88619, Bug #27164393)