Dear MySQL users, MySQL Server 5.7.9 (GA) is a new version of the world's most popular open source database. This is the first official release of MySQL 5.7.
[Due to size limitations on forum/email, this announcement had to be split into two parts. This is part 1.] The new features in this release are now deemed to be of Release quality. Note that 5.7.9 includes all features in MySQL 5.6 and previous 5.7 Development Milestone Releases. An overview of what's new in MySQL 5.7 is available online at http://dev.mysql.com/doc/refman/5.7/en/mysql-nutshell.html For information on installing MySQL 5.7.9 on new servers, please see the MySQL installation documentation at http://dev.mysql.com/doc/refman/5.7/en/installing.html MySQL Server 5.7.9 is available in source and binary form for a number of platforms from the "Development Releases" selection of our download pages at http://dev.mysql.com/downloads/mysql/ MySQL Server 5.7.9 is also available from our repository for Linux platforms, go here for details: http://dev.mysql.com/downloads/repo/ Windows packages are available via the Installer for Windows or .ZIP (no-install) packages for more advanced needs. The point and click configuration wizards and all MySQL products are available in the unified Installer for Windows: http://dev.mysql.com/downloads/installer/ 5.7.9 also comes with a web installer as an alternative to the full installer. The web installer doesn't come bundled with any actual products and instead relies on download-on-demand to fetch only the products you choose to install. This makes the initial download much smaller but increases install time as the individual products will need to be downloaded. We welcome and appreciate your feedback, bug reports, bug fixes, patches, etc.: http://bugs.mysql.com/report.php The following section lists the changes in MySQL 5.7.9 since the 5.7.8 Release Candidate. http://dev.mysql.com/doc/relnotes/mysql/5.7/en/news-5-7-9.html Enjoy! ============================================================================== Changes in MySQL 5.7.9 (2015-10-21) * Audit Log Plugin Notes * C API Notes * Compilation Notes * Packaging Notes * Performance Schema Notes * Plugin Notes * Security Notes * Spatial Data Support * sys Schema Notes * Functionality Added or Changed * Bugs Fixed Audit Log Plugin Notes * Some events cannot be terminated. Previously, if an audit plugin returned nonzero status for a nonterminable event, the server ignored the status and continued processing the event. However, if an audit plugin used the my_message() function to terminate a nonterminable event, a server exit occurred. Now the server correctly handles termination of nonterminable events using my_message(). (Bug #21458066) C API Notes * The mysql_options() C API function has two new options, MYSQL_OPT_MAX_ALLOWED_PACKET and MYSQL_OPT_NET_BUFFER_LENGTH, that set the max_allowed_packet and net_buffer_length system variables, respectively. Each option name also now can be passed to the mysql_get_option() C API function to retrieve its value. For more information, see mysql_options() (http://dev.mysql.com/doc/refman/5.7/en/mysql-options.html), and mysql_get_option() (http://dev.mysql.com/doc/refman/5.7/en/mysql-get-option.html). (Bug #20821550) * Previously, it was necessary to call mysql_thread_end() for each mysql_thread_init() call to avoid a memory leak. C API internals have been reimplemented to reduce the amount of information allocated by mysql_thread_init() that must be freed by mysql_thread_end(): + For release/production builds without debugging support enabled, mysql_thread_end() need not be called. + For debug builds, mysql_thread_init() allocates debugging information for the DBUG package (see The DBUG Package (http://dev.mysql.com/doc/refman/5.7/en/dbug-package.html)). mysql_thread_end() must be called for each mysql_thread_init() call to avoid a memory leak. (Bug #20621281, Bug #21802367) Compilation Notes * MySQL distributions now contain a mysqlclient.pc file that provides information about MySQL configuration for use by the pkg-config command. This enables pkg-config to be used as an alternative to mysql_config for obtaining information such as compiler flags or link libraries required to compile MySQL applications. For more information, see Building C API Client Programs Using pkg-config (http://dev.mysql.com/doc/refman/5.7/en/c-api-building-clients-pkg-config.html). A new INSTALL_PKGCONFIGDIR CMake option is available to specify the directory in which to install the mysqlclient.pc file. The default value is INSTALL_LIBDIR/pkgconfig, unless INSTALL_LIBDIR ends with /mysql, in which case that is removed first. (Bug #20637746) Packaging Notes * The shell and Perl versions of mysql_install_db have been removed from MySQL distributions. The executable C++ version of mysql_install_db implemented in MySQL 5.7.5 is still present, but remains deprecated (use mysqld --initialize instead) and will be removed in a future MySQL release. (Bug #21625471) * The deprecated _r versions of the libmysqlclient libraries are no longer installed. (Bug #21311067) Performance Schema Notes * Previously, the transaction instrument in the setup_instruments table was disabled by default, and the events_transactions_current and events_transactions_history consumers in the setup_consumers table were enabled by default. This setup is inconsistent, and having the consumers enabled could lead to the impression that transactions were instrumented by default. Now, the consumers are also disabled by default. To monitor transactions, enable the instrument and the applicable consumers. (Bug #21780891) * With the show_compatibility_56 system variable disabled, SHOW VARIABLES and SHOW STATUS statements failed if MySQL was compiled without Performance Schema support. Consequently, it is no longer possible to compile without the Performance Schema. If it is desired to compile without particular types of instrumentation, that can be done with the following CMake options: DISABLE_PSI_COND DISABLE_PSI_FILE DISABLE_PSI_IDLE DISABLE_PSI_MEMORY DISABLE_PSI_METADATA DISABLE_PSI_MUTEX DISABLE_PSI_PS DISABLE_PSI_RWLOCK DISABLE_PSI_SOCKET DISABLE_PSI_SP DISABLE_PSI_STAGE DISABLE_PSI_STATEMENT DISABLE_PSI_STATEMENT_DIGEST DISABLE_PSI_TABLE DISABLE_PSI_THREAD DISABLE_PSI_TRANSACTION For example, to compile without mutex instrumentation, configure MySQL using the -DDISABLE_PSI_MUTEX=1 option. (Bug #21669500) * Two changes were made regarding the effect of show_compatibility_56: + Previously, when show_compatibility_56=OFF, selecting from the following INFORMATION_SCHEMA system and status variable tables returned an empty result and a deprecation warning: INFORMATION_SCHEMA.GLOBAL_VARIABLES INFORMATION_SCHEMA.SESSION_VARIABLES INFORMATION_SCHEMA.GLOBAL_STATUS INFORMATION_SCHEMA.SESSION_STATUS This caused confusion for applications that were not aware that such selects could be empty: An empty result and a warning was not sufficient notice to signal the need to migrate to the corresponding Performance Schema system and status variable tables. To address this issue, selecting from the INFORMATION_SCHEMA system and status tables now produces an error, to make it more evident that an application is operating under conditions that require modification, as well as where the problem lies. The error code is ER_FEATURE_DISABLED_SEE_DOC. The error message indicates which table is disabled and that the show_compatibility_56 documentation should be consulted. + Previously, when show_compatibility_56=ON, selecting from the following Performance Schema status variable tables returned an empty result: performance_schema.global_status performance_schema.session_status This made it more difficult to migrate applications from the INFORMATION_SCHEMA status variable tables to the corresponding Performance Schema tables: Successfully selecting from the Performance Schema tables required knowing both that the server is from MySQL 5.7 and that show_compatibility_56=OFF. To address this issue, selecting from the Performance Schema status variable tables now produces the same result regardless of the value of show_compatibility_56. Thus, it is necessary to know only that the server is from MySQL 5.7.9 or higher. (If so, select from the Performance Schema tables. Otherwise, select from the INFORMATION_SCHEMA tables.) For additional information about the effects of show_compatibility_56 and migration issues, see Server System Variables (http://dev.mysql.com/doc/refman/5.7/en/server-system-variables.html), and Migrating to Performance Schema System and Status Variable Tables (http://dev.mysql.com/doc/refman/5.7/en/performance-schema-variable-table-migration.html). (Bug #21606701) * These Performance Schema tables now are world readable and accessible without the SELECT privilege: global_variables, session_variables, global_status, and session_status. An implication of this change is that SHOW VARIABLES and SHOW STATUS no longer require privileges on the underlying Performance Schema tables from which their output is produced when show_compatibility_56=OFF. (Bug #21251297) * With the show_compatibility_56 system variable enabled, the reported values of the Last_query_cost and Last_query_partial_plans status variables were incorrect. With the show_compatibility_56 system variable disabled, the reported values of the Created_tmp_tables, Handler_external_lock, and Table_open_cache_% status variables were incorrect. (Bug #20483278, Bug #21788549, Bug #21788887) * The Performance Schema session_account_connect_attrs had mistakenly been changed to require the SELECT privilege. It requires no special privileges again. (Bug #77702, Bug #21436364) References: This bug was introduced by Bug #14569746. * In the setup_timers table, a CYCLE timer for ARM64 platforms is now available. (Bug #77620, Bug #21374923) * The Performance Schema now includes these instruments for monitoring I/O on binary log and relay log cache files: wait/io/file/sql/binlog_cache wait/io/file/sql/binlog_index_cache wait/io/file/sql/relaylog_cache wait/io/file/sql/relaylog_index_cache In addition, the default value of performance_schema_max_file_classes has been increased from 50 to 80. (Bug #76225, Bug #20675180) * The Performance Schema threads table now contains a THREAD_OS_ID column that indicates the thread or task identifier as defined by the underlying operating system. For example, the column value corresponds to the Process Explorer thread ID on Windows and the gettid() value on Linux. For more information, see The threads Table (http://dev.mysql.com/doc/refman/5.7/en/threads-table.html). If you upgrade to this MySQL release from an earlier version, you must run mysql_upgrade (and restart the server) to incorporate this change into the performance_schema database. Plugin Notes * The audit plugin API has been extensively revised to support a finer breakdown of the general event type (MYSQL_AUDIT_GENERAL_CLASS) into more specific events. This enables audit plugins to more precisely indicate the types of events in which they are interested and reduces overhead for plugins that have use for only a few event types. The API also now permits early termination of event execution. For more information, see Writing Audit Plugins (http://dev.mysql.com/doc/refman/5.7/en/writing-audit-plugins.html). The general event type is still available but is deprecated and will be removed in a future MySQL release. In addition, a security_context plugin service is now available. Audit plugins can use this service to examine or modify the security context of threads associated with audited events. See MySQL Services for Plugins (http://dev.mysql.com/doc/refman/5.7/en/plugin-services.html). Security Notes * For RPM-based packages, the permissions used to create the data directory (/var/lib/mysql) have been changed from 755 to 751. This tightens the data directory permissions while still permitting world access to the mysql.sock file in that directory. (Bug #21066592) Spatial Data Support * The required version of the Boost library for server builds has been raised from 1.58.0 to 1.59.0. (Bug #21567456) * Spatial functions such as ST_MPointFromText() and ST_GeomFromText() that accept WKT-format representations of MultiPoint values now permit individual points within values to be surrounded by parentheses. For example, both of the following function calls are valid, whereas previously the second one produced an error: ST_MPointFromText('MULTIPOINT (1 1, 2 2, 3 3)') ST_MPointFromText('MULTIPOINT ((1 1), (2 2), (3 3))') In addition, functions such as ST_AsText() and ST_AsWKT() that produce WKT-format results now display MultiPoint values with parentheses surrounding each point. (Bug #54065, Bug #11761559) sys Schema Notes * The sys schema included in MySQL distributions was updated to version 1.5.0. This includes new features and fixes for several issues: + sys schema objects now have a DEFINER of 'mysql.sys'@'localhost'. (Previously, the DEFINER was 'root'@'localhost'.) Use of the dedicated mysql.sys account avoids problems that occur if a DBA renames or removes the root account. + The following sys schema session views now provide progress reporting for long-running transactions: processlist session x$processlist x$session The progress column of these views shows the percentage of work completed for stages that support progress reporting. + A new diagnostics() stored procedure enables DBAs and other support personnel to collect diagnostic information for investigating MySQL instances. A new metrics view and statement_performance_analyzer() stored procedure provide supporting infrastructure for the diagnostics() procedure. + The sys schema ps_is_instrument_default_enabled() and ps_is_instrument_default_timed() stored functions returned incorrect results in some cases. + The ENABLED and HISTORY columns that were added to the Performance Schema setup_actors table in earlier MySQL 5.7 releases caused the sys schema ps_setup_reset_to_default() stored procedure not to work. + Handing of event-timing information in the sys schema was updated to handle changes to Performance Schema event-timing columns in MySQL 5.7.8. + mysql_upgrade previously checked for an exact object-count value in the sys schema to determine whether an upgrade was needed. If local objects had been added, the resulting reinstallation removed those objects. Now it checks for at least the expected number of objects. If you upgrade to this MySQL release from an earlier version, you must run mysql_upgrade to incorporate these changes into the sys schema. Thanks to Daniël van Eeden, Jesper Wisborg Krogh, Shlomi Noach, and Morgan Tocker for their contributions to this update. (Bug #78115, Bug #21647101, Bug #77927, Bug #21550271, Bug #78720, Bug #21966366) Functionality Added or Changed * Incompatible Change; InnoDB: To better manage redo log format changes, the redo log header of the first redo log file (ib_logfile0) now includes a format version identifier and a text string that identifies the MySQL version that created the redo log files. A new boolean configuration option, innodb_log_checksums, replaces the innodb_log_checksum_algorithm option. innodb_log_checksums=ON enables a CRC-32C checksum, making it the only supported checksum for redo log pages. This patch also removes unused fields from the redo log header and checkpoint pages. Due to redo log format changes introduced by this patch, upgrading to or downgrading from MySQL 5.7.9 and higher requires a clean shutdown and, in some cases, removal of existing redo log files. For instructions related to this change, see Changes Affecting Upgrades to MySQL 5.7 (http://dev.mysql.com/doc/refman/5.7/en/upgrading-from-previous-series.html), and Changes Affecting Downgrades from MySQL 5.7 (http://dev.mysql.com/doc/refman/5.7/en/downgrading-to-previous-series.html). (Bug #21759424) * Important Change; InnoDB: DYNAMIC replaces COMPACT as the implicit default row format for InnoDB tables. A new configuration option, innodb_default_row_format, specifies the default InnoDB row format. Permitted values include DYNAMIC (the default), COMPACT, and REDUNDANT. The COMPACT row format remained the default row format until this release to ensure compatibility with older versions of InnoDB in MySQL 5.1 and earlier. Now that MySQL 5.1 has reached the end of its product lifecycle, the newer DYNAMIC row format becomes the default. For information about advantages of the DYNAMIC row format, see DYNAMIC and COMPRESSED Row Formats (http://dev.mysql.com/doc/refman/5.7/en/innodb-row-format-dynamic.html). Newly created tables use the row format defined by innodb_default_row_format when a ROW_FORMAT option is not specified explicitly or when ROW_FORMAT=DEFAULT is used. Existing tables retain their current row format if a ROW_FORMAT option was specified explicitly. If a ROW_FORMAT option was not specified explicitly or if ROW_FORMAT=DEFAULT was used, any operation that rebuilds a table also silently changes the row format of the table to the format defined by innodb_default_row_format. For more information, see Specifying the Row Format for a Table (http://dev.mysql.com/doc/refman/5.7/en/innodb-row-format-specification.html). * Important Change: Introduced the -> JSON column-path operator. column->path is now supported as a synonym of JSON_EXTRACT(column, path), where column is a JSON column, and path is a valid JSON path. An expression with ->, like its equivalent that uses JSON_EXTRACT() instead, can be used in place of a column identifier wherever the latter can occur within a valid SQL statement. For example, the following CREATE TABLE and SELECT statements are valid: CREATE TABLE t1 ( a JSON, b INT, g INT GENERATED ALWAYS AS (a->"$.id"), h INT GENERATED ALWAYS AS (a->"$.storeid"), INDEX i (g), INDEX j (h) ); SELECT CONCAT(a->"$.fname", ' ', a->"$.lname") AS name, a->"$.id" AS id, a->"$.storeid" AS store FROM t1 WHERE g > 500 ORDER BY a->"$.storeid", a->"$.lname"; A column-path expression can be used for any column value that is read in a SELECT column list, or in a WHERE, ORDER BY, or GROUP BY clause in any SQL statement; such expressions cannot be used to set values. When an SQL statement contains one or more expressions using -> notation, each of these is translated into an equivalent expression that employs the JSON_EXTRACT() function instead. This can be seen in the output from EXPLAIN when used on such a statement. Like JSON_EXTRACT(), the -> operator returns as NULL if no matching value for an otherwise valid path is found. For more information about -> and JSON_EXTRACT(), see Functions That Search JSON Values (http://dev.mysql.com/doc/refman/5.7/en/json-search-functions.html). See Searching and Modifying JSON Values (http://dev.mysql.com/doc/refman/5.7/en/json.html#json-paths), for information about JSON path support. See also Secondary Indexes and Virtual Generated Columns (http://dev.mysql.com/doc/refman/5.7/en/create-table.html#create-table-secondary-indexes-virtual-columns), for additional information and examples. * InnoDB: A new INNODB_METRICS server operations counter (innodb_dict_lru_count) counts the number of tables evicted from the table cache LRU list. Thanks to Daniël van Eeden for the patch. (Bug #21682332, Bug #78190) * InnoDB: The new innodb_numa_interleave read-only configuration option allows you to enable the NUMA interleave memory policy for allocation of the InnoDB buffer pool. When innodb_numa_interleave is enabled, the NUMA memory policy is set to MPOL_INTERLEAVE for the mysqld process. After the InnoDB buffer pool is allocated, the NUMA memory policy is set back to MPOL_DEFAULT. This option is only available on NUMA-enabled systems. Thanks to Stewart Smith for the patch. (Bug #18871046, Bug #72811) * MySQL distributions now include these header files because my_sys.h depends on them: my_thread_local.h, thr_cond.h, thr_mutex.h, thr_rwlock.h. (Bug #21909332) * MySQL Server RPM packages now obsolete MySQL Connector C. Installing MySQL Server causes older libmysqlclient from any MySQL Connector C packages to be removed and replaces them with the current libmysqlclient. (Bug #21900800) * yaSSL was upgraded to version 2.3.8. (Bug #21888925) * Insert overhead for the MEMORY storage table was reduced by caching computed hash values. (Bug #21866029) * RPM .spec files were updated so that MySQL Server builds from source RPM packages will include the proper files to take advantage of operating system NUMA capabilities. This introduces a runtime dependency on libnuma.so.1. RPM and yum detect this and refuse to install if that library is not installed. (Bug #21775221) * The initial-password message written by mysqld --initialize to the error log has been reduced from a [Warning] to a [Note]. (Bug #21680457) * The JSON_APPEND() function was renamed to JSON_ARRAY_APPEND(). (Bug #21560934) * Unneeded scripts and test files were removed from the tests directory and that directory was renamed to testclients. (Bug #21490075) * The JSON value-updating functions JSON_APPEND(), JSON_SET(), JSON_REPLACE(), and JSON_INSERT() now treat SQL NULL values as JSON null literals, which is consistent with JSON_OBJECT() and JSON_ARRAY(). (Bug #21450922) * The server now generates a warning when the default_storage_engine or default_tmp_storage_engine system variable is set to a disabled storage engine named in the disabled_storage_engines system variable. (Bug #21405865) * Metadata locking for tablespaces has been extended so that, for DDL statements that refer to multiple tablespaces, a metadata lock is acquired on all used tablespaces. (Bug #21376265) * Support for building with Solaris Studio 5.13 was added. (Bug #21185883) * mysql_ssl_rsa_setup now is less noisy by default. Output from openssl commands is displayed only if --verbose is given. (Bug #21024979) * yaSSL was upgraded to version 2.3.7d. This fixes a connection-failure issue when used with the thread pool plugin. (Bug #20774956) * mysqlpump now supports a --version option. (Bug #77894, Bug #21534277) * Performance Schema digests in DIGEST_TEXT columns have ... appended to the end to indicate when statements exceed the maximum statement size and were truncated. This is also now done for statement text values in SQL_TEXT columns. (Bug #75861, Bug #20519832) * Output from mysql_upgrade is now less noisy and more informative. (Bug #59077, Bug #11766046) * A new SHUTDOWN SQL statement is available. This provides a SQL-level interface to the same functionality previously available using the mysqladmin shutdown command or the mysql_shutdown() C API function. See SHUTDOWN Syntax (http://dev.mysql.com/doc/refman/5.7/en/shutdown.html). * Work was done to clean up the source code base, including: Removal of unneeded CMake checks; removing unused macros from source files; reorganizing header files to reduce the number of dependencies and make them more modular, removal of function declarations without definitions. Bugs Fixed * Important Change; InnoDB; Partitioning: There was no way to upgrade existing partitioned tables to use the native partitioning implemented for InnoDB tables in MySQL 5.7.6. This fix adds support to both mysql_upgrade and the mysql client for upgrading partitioned InnoDB tables created in previous releases, which used the ha_partition handler, to use InnoDB native partitioning instead. + mysql_upgrade now checks for all InnoDB tables that were created using the generic ha_partition handler and attempts to upgrade them to InnoDB native partitioning. + In the mysql client, pre-5.7.6 partitioned InnoDB tables can be upgraded one by one to native partitioning using the ALTER TABLE ... UPGRADE PARTITIONING statement that is implemented in this release. (Bug #20727344, Bug #76374) * Important Change; Replication: The START SLAVE and STOP SLAVE statements can no longer be used for the group_replication_recovery channel. See Replication Channels (http://dev.mysql.com/doc/refman/5.7/en/replication-channels.html), for more information. (Bug #21680074) * Important Change; Replication: In MySQL 5.7, the binlog_max_flush_queue_time system variable introduced in MySQL 5.6 no longer has any effect. This variable is now deprecated, and has been marked for eventual removal in a future MySQL release. (Bug #21347087) * Important Change; Replication: When using a single-threaded slave, the status of the applier thread is now reported as part of the replication_applier_status_by_worker table, instead of replication_applier_status_by_coordinator. This means that replication_applier_status_by_coordinator is now empty when using a single-threaded slave; it should be noted that such reporting for multi-threaded slaves has not been changed, and continues to be shown in that table. (Bug #74765, Bug #20001173) * InnoDB: The undo log contained insufficient information about virtual columns and virtual column indexes, which could cause a server exit when adding or dropping virtual columns. As a result, a slow shutdown (using innodb_fast_shutdown=0) is required prior to performing an in-place upgrade or downgrade from MySQL 5.7.8. For more information, refer to Upgrading or Downgrading MySQL (http://dev.mysql.com/doc/refman/5.7/en/upgrading-downgrading.html). (Bug #21869656) * InnoDB: A table-rebuilding ALTER TABLE operation that created an index on a virtual column raised an assertion. (Bug #21847170) * InnoDB: A SELECT ... FOR UPDATE operation on a table with virtual generated columns raised an assertion. InnoDB unnecessarily retrieved a non-indexed virtual column for a covered secondary index scan. (Bug #21827963) * InnoDB: Creating a table with large rows failed when using a ROW_FORMAT=DYNAMIC or ROW_FORMAT=COMPRESSED due to an incorrect undo log record size calculation. The same operation succeeded with ROW_FORMAT=COMPACT or ROW_FORMAT=REDUNDANT. ROW_FORMAT=DYNAMIC and ROW_FORMAT=COMPRESSED now permit a row length violation at DDL time if innodb_strict_mode is disabled. (Bug #21816041) * InnoDB: Adding an index on a generated virtual column with an index prefix length that exceeded the maximum length caused a server exit. (Bug #21812026) * InnoDB: A memory leak occurred after crash recovery. Memory allocated in fil_space_read_name_and_filepath() was not freed. (Bug #21811321) * InnoDB: Altering the data type or computation method of a virtual generated column caused an error. (Bug #21810004) * InnoDB: An assertion was raised when creating a spatial index. InnoDB failed to count virtual columns that preceded the spatial index column. (Bug #21807340) * InnoDB: Tablespace discovery modifications in MySQL 5.7.5 included the removal of code related to MLOG_FILE_CREATE2 redo log records. As a result, the redo log did not contain sufficient information about file creation. (Bug #21801423) * InnoDB: The redo log provided no indication that redo logging is disabled for index page writes during concurrent DDL operations. As a result, external hot backup tools could produce corrupt backups. (Bug #21796691) * InnoDB: In debug builds, recovery asserted during a transparent page compression test. A torn page from the doublewrite buffer caused an LSN debug check failure. (Bug #21796092) * InnoDB: Creating a virtual generated column on a partitioned table caused a server exit. (Bug #21790751, Bug #78326) * InnoDB: An old version of numactl headers on the build host caused a compilation error when building a MySQL version that includes NUMA memory policy support. (Bug #21785074) * InnoDB: A SELECT .. FOR UPDATE operation on an indexed virtual generated column raised an assertion. (Bug #21775459) * InnoDB: The mysql_system_tables_fix_for_downgrade.sql script, provided to facilitate system table alterations when downgrading from MySQL 5.7.6 or higher, was removed from the MySQL installation directory. The script was no longer valid for all downgrade paths. For more information, see Changes Affecting Downgrades from MySQL 5.7 (http://dev.mysql.com/doc/refman/5.7/en/downgrading-to-previous-series.html). (Bug #21753832) * InnoDB: The tablespace identifier (space_id) was logged twice for an MLOG_TRUNCATE redo log record. (Bug #21744589) * InnoDB: Crash recovery issued an invalid error message indicating that an isl file could not be opened or is not correct. (Bug #21691438) * InnoDB: The InnoDB Monitor displayed incorrect mutex creation information for a mutex with a long semaphore wait. (Bug #21682997, Bug #78179) * InnoDB: Error message formatting was corrected in the os_file_write_page() function. (Bug #21681433) * InnoDB: The i_s_dict_fill_sys_tablespaces() function could free memory associated with a file name before printing an error message. (Bug #21680518, Bug #78180) * InnoDB: In debug builds, scanned and applied redo log records are now printed with a string identifier instead of a numerical identifier when --debug=d,ib_log is passed to mysqld. For example, rec MLOG_2BYTES is now printed instead of rec 2. (Bug #21664268) * InnoDB: The number of system calls made by the InnoDB page compression feature was reduced. (Bug #21654695) * InnoDB: The recv_parse_log_rec function returned the length of the redo log record instead of 0 when encountering an incomplete MLOG_CHECKPOINT record. (Bug #21640085, Bug #78058) References: This bug is a regression of Bug #17798076. * InnoDB: In debug builds, a parenthesis mismatch in a MATCH() ... AGAINST clause raised a full-text parser assertion. (Bug #21638907) * InnoDB: Debug code was added to avoid a doublewrite buffer assertion that was raised during Valgrind testing. (Bug #21631197) * InnoDB: Log messages were improved to help identify out-of-space errors that could occur when adding rollback segments. (Bug #21629618) * InnoDB: In debug builds, enabling the innodb_log_checkpoint_now debug option while a table-rebuilding ALTER TABLE operation is running could result in an infinite loop. (Bug #21628087, Bug #78056) * InnoDB: UNIV_INLINE was not defined for the dict_table_has_indexed_v_cols function. (Bug #21628058, Bug #78055) * InnoDB: Altering a virtual column data type is not supported as an in-place operation. (Bug #21617377) * InnoDB: Sorting was skipped by an ALTER TABLE statement that changed the primary key and dropped the last column of the previous primary key. (Bug #21612714, Bug #78020) * InnoDB: During recovery, an invalid isl file was treated as a missing isl file, resulting in the tablespace being opened using the file location defined in an MLOG_FILE_* record in the redo log. Recovery no longer opens the tablespace if the isl file is invalid. (Bug #21577278) * InnoDB: Code that provided a timeout mechanism intended to reduce adaptive hash index search latch (btr_search_latch) contention was removed. The code became obsolete after the introduction of adaptive hash index search system partitioning in MySQL 5.7.8. (Bug #21569876, Bug #77957) * InnoDB: The COMPRESSION option was displayed incorrectly in SHOW CREATE TABLE output. (Bug #21557723, Bug #77940) * InnoDB: An innodb_data_file_path mismatch raised an assertion, as did initializing the database with a data file size that was too small for the specified page size. (Bug #21551464) * InnoDB: A check was added to prevent accessing full-text index tables that are in an inconsistent state. (Bug #21529012) * InnoDB: A schema mismatch error occurred when importing a tablespace that was altered by DROP INDEX operation on the source server. (Bug #21514135, Bug #77659) * InnoDB: A virtual generated column on a table that uses index condition pushdown (ICP) caused an assertion. (Bug #21507796, Bug #21478287) * InnoDB: Creating an index on a virtual generated column after adding a full-text index on a preceding column raised an assertion. (Bug #21478389) * InnoDB: An INSERT operation raised a btr_search_enabled assertion. Assertion code was too restrictive. (Bug #21457373) * InnoDB: Reloading a table that was evicted while empty caused an AUTO_INCREMENT value to be reset. (Bug #21454472, Bug #77743) * InnoDB: A crash during a TRUNCATE TABLE operation caused the server to exit on startup. (Bug #21451922) * InnoDB: The .isl was not removed when dropping a general tablespace that was created outside of the MySQL data directory. (Bug #21446772) * InnoDB: An inconsistent read occurred under the REPEATABLE READ transaction isolation level. Transactions that operated on the same row were removed from transaction ID list in the incorrect order. (Bug #21433768) References: This bug is a regression of Bug #17320977. * InnoDB: Attempting to create a general tablespace data file on a Windows root drive caused an error. (Bug #21419888) * InnoDB: After disabling the adaptive hash index feature, an adaptive hash index latch was unnecessarily obtained and released. (Bug #21407023) * InnoDB: An ALTER TABLE operation on a table with an index defined on a virtual column incorrectly modified the data of the virtual column. (Bug #21376546, Bug #77628) * InnoDB: Resizing the buffer pool online raised an assertion due to a memory full condition. (Bug #21348684, Bug #77564) * InnoDB: An ALTER TABLE ... IMPORT TABLESPACE operation with innodb_page_size=4K and ROW_FORMAT=DYNAMIC raised an assertion. (Bug #21341030) * InnoDB: For spatial indexes, InnoDB unnecessarily stored a 3072-byte prefix in undo log records instead of just the maximum bounding rectangle (MBR). For columns with externally stored data, both the prefix and MBR are logged. (Bug #21340268) * InnoDB: Invalid init_ftfuncs() assertion code was removed. (Bug #21300774) * InnoDB: Memory allocation sanity checks were added to the memcached code. (Bug #21288106) * InnoDB: An incorrect reference count caused a hang in the TrxInInnoDB constructor. innobase_close_connection() released the transaction object before destroying the TrxInInnoDB object where the reference count is adjusted. (Bug #21280816) * InnoDB: A MySQL 5.7.8 patch that reintroduced SHOW ENGINE INNODB MUTEX functionality caused a performance regression. (Bug #21266784) * InnoDB: A memcached flush_all command raised an assertion. A function that starts a transaction was called from within assertion code. (Bug #21239299, Bug #75199) * InnoDB: A shutdown hang occurred when a high priority transaction waited for a victim transaction to exit while the victim transaction waited for an asynchronous rollback to complete. (Bug #21143276) * InnoDB: A data corruption occurred on ARM64. GCC builtins did not issue the correct fences when setting or unsetting the lock word. (Bug #21102971, Bug #76135) * InnoDB: Server shutdown was delayed waiting for the purge thread to exit. To avoid this problem, the number of calls to trx_purge() was reduced, and the trx_purge() batch size was reduced to 20. (Bug #21040050) * InnoDB: In READ COMMITTED mode, a REPLACE operation on a unique secondary index resulted in a constraint violation. Thanks to Alexey Kopytov for the patch. (Bug #21025880, Bug #76927) * InnoDB: The IBUF_BITMAP_FREE bit indicated that there was more free space in the leaf page than was actually available. (Bug #20796566) * InnoDB: Moving the data directory before recovering a crashed database caused tablespace discovery to fail for file-per-table tablespaces created outside of the MySQL data directory. (Bug #20698468) * InnoDB: The innodb_buf_flush_list_now debug setting failed to flush all dirty pages to disk. (Bug #20582189) * InnoDB: An ALTER TABLE ... ADD FULLTEXT INDEX operation raised an assertion. A thread attempted to use a lower priority transaction that was being rolled back before the rollback operation completed. (Bug #20481175) * InnoDB: Running an ALTER TABLE operation on a referencing table with a cascading foreign key constraint during a concurrent DML operation on the referenced table caused a loss of referential integrity. (Bug #20367116) * InnoDB: Setting lower_case_table_names=0 on a case-insensitive file system could result in a hang condition when running an INSERT INTO ... SELECT ... FROM tbl_name operation with the wrong tbl_name letter case. An error message is now printed and the server exits when attempting to start the server with --lower_case_table_names=0 on a case-insensitive file system. (Bug #20198490, Bug #75185) * InnoDB: The server failed to start with an innodb_force_recovery setting greater than 3. InnoDB was set to read-only mode before redo logs were applied. DROP TABLE is now supported with an innodb_force_recovery setting greater than 3. (Bug #19779113) * InnoDB: The trx_sys_read_pertable_file_format_id() function reported the wrong file format. (Bug #19206671) * Packaging; OS X: Using user=mysql during installation on OS X did not allow the mysql database to be installed. To fix this problem, OS X packages now use the --no-defaults option when creating this database. This also means that having a my.cnf file on the system no longer affects the installation. (Bug #21364902) * Partitioning: Error handling for failed partitioning-related ALTER TABLE operations against non-partitioned tables was not performed correctly (Bug #20284744) * Partitioning: ALTER TABLE when executed from a stored procedure did not always work correctly with tables partitioned by RANGE. (Bug #16613004) * To be continued in part 2.... -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/mysql
