MySQL 5.6 Release Notes  /  Changes in MySQL 5.6.0 (Not released, Milestone 4)

Changes in MySQL 5.6.0 (Not released, Milestone 4)


This is a milestone release, for use at your own risk. Significant development changes take place in milestone releases and you may encounter compatibility issues, such as data format changes that require attention in addition to the usual procedure of running mysql_upgrade. For example, you may find it necessary to dump your data with mysqldump before the upgrade and reload it afterward.

Performance Schema Notes

  • The Performance Schema now includes instrumentation for table input and output. Instrumented operations include row-level accesses to persistent base tables or temporary tables. Operations that affect rows are fetch, insert, update, and delete. For a view, waits are associated with base tables referenced by the view.

Globally Unique Server IDs

  • Replication: Globally unique IDs for MySQL servers were implemented. A UUID is now obtained automatically when the MySQL server starts. The server first checks for a UUID written in the auto.cnf file (in the server's data directory), and uses this UUID if found. Otherwise, the server generates a new UUID and saves it to this file (and creates the file if it does not already exist). This UUID is available as the server_uuid system variable.

    MySQL replication masters and slaves know each other's UUIDs. The value of a slave's UUID can be read in the output of SHOW SLAVE HOSTS. After a slave is started using START SLAVE, the value of the master's UUID is available on the slave in the output of SHOW SLAVE STATUS. (Bug #33815, Bug #11747723)

    References: See also Bug #16927, Bug #11745543.

Functionality Added or Changed

  • Partitioning: It is now possible to exchange a partition of a partitioned table or a subpartition of a subpartitioned table with a nonpartitioned table that otherwise has the same structure using the ALTER TABLE ... EXCHANGE PARTITION statement. This can be used, for example, for importing and exporting partitions.

    For more information and examples, see Exchanging Partitions and Subpartitions with Tables.

  • Replication: These unused and deprecated items have been removed: the --init-rpl-role and --rpl-recovery-rank options, the rpl_recovery_rank system variable, and the Rpl_status status variable. (Bug #54649, Bug #11762095)

    References: See also Bug #34437, Bug #11747900, Bug #34635, Bug #11747961.

  • Replication: The SHOW SLAVE STATUS statement now has a Master_Info_File field indicating the location of the file. (Bug #50316, Bug #11758151)

  • Replication: MySQL now supports delayed replication such that a slave server deliberately lags behind the master by at least a specified amount of time. The default delay is 0 seconds. Use the new MASTER_DELAY option for CHANGE MASTER TO to set the delay to N seconds:


    An event received from the master is not executed until at least N seconds later than its execution on the master.

    START SLAVE and STOP SLAVE take effect immediately and ignore any delay. RESET SLAVE resets the delay to 0.

    SHOW SLAVE STATUS has three new fields that provide information about the delay:

    • SQL_Delay: The number of seconds that the slave must lag the master.

    • SQL_Remaining_Delay: When Slave_SQL_Running_State is Waiting until MASTER_DELAY seconds after master executed event, this field contains the number of seconds left of the delay. At other times, this field is NULL.

    • Slave_SQL_Running_State: The state of the SQL thread (analogous to Slave_IO_State). The value is identical to the State value of the SQL thread as displayed by SHOW PROCESSLIST.

    When the slave SQL thread is waiting for the delay to elapse before executing an event, SHOW PROCESSLIST displays its State value as Waiting until MASTER_DELAY seconds after master executed event.

    The file now contains the delay value, so the file format has changed. See Slave Status Logs. In particular, the first line of the file now indicates how many lines are in the file. If you downgrade a slave server to a version older than MySQL 5.6, the older server will not read the file correctly. To address this, modify the file in a text editor to delete the initial line containing the number of lines.

    The introduction of delayed replication entails these restrictions:

    • Previously the BINLOG statement could execute all types of events. Now it can execute only format description events and row events.

    • The output from mysqlbinlog --base64-output=ALWAYS cannot be parsed. ALWAYS becomes an invalid value for this option in 5.6.1.

    For additional information, see Delayed Replication. (Bug #28760, Bug #11746794)

  • The Romansh locale 'rm_CH' is now a permissible value for the lc_time_names system variable. (Bug #50915, Bug #11758678)

  • mysqlbinlog now has a --binlog-row-event-max-size option to enable large row events to be read from binary log files. (Bug #49932)

  • mysqldump now has an --add-drop-trigger option which adds a DROP TRIGGER IF EXISTS statement before each dumped trigger definition. (Bug #34325, Bug #11747863)

  • mysqlbinlog now has the capability to back up a binary log in its original binary format. When invoked with the --read-from-remote-server and --raw options, mysqlbinlog connects to a server, requests the log files, and writes output files in the same format as the originals. See Using mysqlbinlog to Back Up Binary Log Files.

  • The Unicode character sets now have a xxx_german2_ci collation that provides DIN-2 (phone book) ordering (for example, utf8_german2_ci). See Unicode Character Sets.

  • A new SQL function, WEIGHT_STRING(), returns the weight string for an input string. The weight string represents the sorting and comparison value of the input string. See String Functions.

  • Vietnamese collations were added for the Unicode character sets. Those based on Unicode Collation Algorithm 5.2.0 have names of the form xxx_vietnamese_520_ci (for example, utf8_vietnamese_520_ci). Those based on Unicode Collation Algorithm 4.0.0 have names of the form xxx_vietnamese_ci (for example, utf8_vietnamese_ci). These collations are the same as the corresponding xxx_unicode_520_ci and xxx_unicode_ci collations except for precomposed characters which are accented versions of A, D, E, O, and U. There is no change to ideographic characters derived from Chinese. There are no digraphs.

  • Unicode collation names now may include a version number to indicate the Unicode Collation Algorithm (UCA) version on which the collation is based. Initial collations thus created use version UCA 5.2.0. For example, utf8_unicode_520_ci is based on UCA 5.2.0. UCA-based Unicode collation names that do not include a version number are based on version 4.0.0.

    LOWER() and UPPER() perform case folding according to the collation of their argument. A character that has uppercase and lowercase versions only in a Unicode version more recent than 4.0.0 will be converted by these functions only if the argument has a collation that uses a recent enough UCA version.

    The LDML rules for creating user-defined collations are extended to permit an optional version attribute in <collation> tags to indicate the UCA version on which the collation is based. If the version attribute is omitted, its default value is 4.0.0. See Adding a UCA Collation to a Unicode Character Set.

  • In MySQL 5.5, setting optimizer_search_depth to the deprecated value of 63 switched to the algorithm used in MySQL 5.0.0 (and previous versions) for performing searches. The value of 63 is now treated as invalid.

Bugs Fixed

  • Security Fix: A security bug was fixed. (Bug #49124)

  • InnoDB: The server could crash on shutdown, if started with --innodb-use-system-malloc=0. (Bug #55581, Bug #11762927)

  • Replication: The internal flag indicating whether a user value was signed or unsigned (unsigned_flag) could sometimes change between the time that the user value was recorded for logging purposes and the time that the value was actually written to the binary log, which could lead to inconsistency. Now unsigned_flag is copied when the user variable value is copied, and the copy of unsigned_flag is then used for logging. (Bug #51426, Bug #11759138)

    References: See also Bug #49562, Bug #11757508.

  • The embedded server could crash when determining which directories to search for option files. (Bug #55062, Bug #11762465)

  • Performance Schema code was subject to a buffer overflow. (Bug #53363)

  • On Windows, an IPv6 connection to the server could not be made using an IPv4 address or host name. (Bug #52381, Bug #11760016)

  • Subquery execution for EXPLAIN could be done incorrectly and raise an assertion. (Bug #52317, Bug #11759957)

  • There was a mixup between GROUP BY and ORDER BY concerning which indexes should be considered or permitted during query optimization. (Bug #52081, Bug #11759746)

  • On Windows, the my_rename() function failed to check whether the source file existed. (Bug #51861, Bug #11759540)

  • The ref column of EXPLAIN output for subquery lines could be missing information. (Bug #50257, Bug #11758106)

  • Passwords for CREATE USER statements were written to the binary log in cleartext rather than in ciphertext. (Bug #50172)

  • The BLACKHOLE storage engine failed to load on Solaris and OpenSolaris if DTrace probes had been enabled. (Bug #47748, Bug #11755909)

  • Some error messages included a literal mysql database name rather than a parameter for the database name. (Bug #46792, Bug #11755079)

  • In the ER_TABLEACCESS_DENIED_ERROR error message, the command name parameter could be truncated. (Bug #45355, Bug #11753840)

  • On Windows, mysqlslap crashed for attempts to connect using shared memory. (Bug #31173, Bug #11747181, Bug #59107, Bug #11766072)

  • To forestall the occurrence of possible relocation errors in the future, libmysys, libmystrings, and libdbug have been changed from normal libraries to noinst libtool helper libraries, and are no longer installed as separate libraries. (Bug #29791, Bug #11746931)

  • A suboptimal query execution plan could be chosen when there were several possible range and ref accesses. Now preference is given to the keys that match the most parts and choosing the best one among them. (Bug #26106, Bug #11746406)

  • Searches for data on a partial index for a column using the utf8 character set failed. (Bug #24858)

  • For queries with GROUP BY, FORCE INDEX was not ignored as it should have been when it would result in a more expensive query execution plan. (Bug #18144, Bug #11745649)

Download these Release Notes
PDF (US Ltr) - 1.5Mb
PDF (A4) - 1.5Mb
EPUB - 374.7Kb