Incompatible Change: This release introduces the TRIGGER privilege. Previously, the SUPER privilege was needed to create or drop triggers. Now those operations require the TRIGGER privilege. This is a security improvement because you no longer need to grant users the SUPER privilege to enable them to create triggers. However, the requirement that the account named in a trigger's DEFINER clause must have the SUPER privilege has changed to a requirement for the TRIGGER privilege. After upgrading, be sure to update your grant tables by running mysql_upgrade. This will assign the TRIGGER privilege to all accounts that had the SUPER privilege. (After updating, you might also consider whether any of those accounts no longer need the SUPER privilege.) If you fail to update the grant tables, triggers may fail when activated. (Bug #9412)

Incompatible Change: Words with apostrophes are now matched in a FULLTEXT search against nonapostrophe words (for example, a search for Jerry will match against the term Jerry's). Users upgrading to this version must issue REPAIR TABLE ... QUICK statements for tables containing FULLTEXT indexes. (Bug #14194)

Incompatible Change: Before MySQL 5.1.6, the server writes general query log and slow query log entries to log files. As of MySQL 5.1.6, the server's logging capabilities for these logs are more flexible. Log entries can be written to log files (as before) or to the general_log and slow_log tables in the mysql database. If logging is enabled, either or both destinations can be selected. The --log-output option controls the destination or destinations of log output. See Selecting General Query and Slow Query Log Output Destinations.

If logging is enabled, the default destination now is to log to tables, which differs from earlier versions. If you had the server configured for logging to log files formerly, use --log-output=FILE to preserve this behavior after an upgrade to MySQL 5.1.6 or higher.

Important Change; MySQL Cluster; Replication: Replication between MySQL Clusters is now supported. It is now also possible to replicate between a MySQL Cluster and a noncluster database. See MySQL Cluster Replication, for more information.

MySQL Cluster: The NDB storage engine now supports the CREATE INDEX and DROP INDEX statements.

MySQL Cluster: Added the ndb_extra_logging system variable.

Packaging: MySQL 5.1.6 introduces some changes to distribution packaging:

ANALYZE TABLE is now supported for partitioned tables. (Bug #13441)

Added the EVENTS table to INFORMATION_SCHEMA.

Added the PARTITIONS table to INFORMATION_SCHEMA.

Partition support is not an “engine”, but it was included in the output of SHOW ENGINES. Now it is not. The have_partition_engine variable was renamed to have_partitioning. (Bug #14355, Bug #16718)

The mysqldump utility now supports an option for dumping tablespaces. Use -Y or --all-tablespaces to enable this functionality. (Bug #16753)

Server plugins can register their own status variables to be displayed by the SHOW STATUS statement.

The ARCHIVE storage engine now supports the AUTO_INCREMENT column attribute and the AUTO_INCREMENT table option. The ARCHIVE Storage Engine.

Server binaries no longer are stripped, except for RPM distributions.

Distributions include both a mysqld optimized server and mysqld-debug debugging server. There is no separate debug distribution.

MySQL 5.1.6 introduces the Event Scheduler which enables statements to be scheduled for execution at predetermined times. Events can be transient (one-time-only) or recurrent at regular intervals, and may execute queries and statements permitted in stored routines, including compound statements.

Events can be altered after creation, and dropped when no longer needed.

Information about scheduled events can be obtained using the statements SHOW EVENTS and SHOW CREATE EVENT, or by querying the INFORMATION_SCHEMA.EVENTS table. All of these are available beginning in MySQL 5.1.6.

Users must have the EVENT privilege (also added in 5.1.6) to create events.

For more information, see Using the Event Scheduler.

Added the event_scheduler system variable.

Special characters in database and table identifiers now are encoded when creating the corresponding directory names and file names. This relaxes the restrictions on the characters that can appear in identifiers. See Mapping of Identifiers to File Names.

Binary distributions for Unix and Unix-like systems no longer include safe_mysqld as a link to mysqld_safe. safe_mysqld has been deprecated since MySQL 4.0 and now is removed.

Added the FILES table to INFORMATION_SCHEMA.

There is no longer a mysqld-max server. (Note: This changed in MySQL 5.1.9: The mysqld-max server also is included in binary distributions.)

Queries against partitioned tables can now take advantage of partition pruning. In some cases, this can result in query execution that is an order of magnitude faster than the same query against a nonpartitioned version of the same table.

Added the --use-threads option for mysqlslap.

MySQL Cluster: CREATE TABLESPACE statements were incorrectly parsed on 64-bit platforms. (INITIAL SIZE size worked, but INITIAL SIZE = size failed.) (Bug #13556)

MySQL Cluster: A BIT column whose offset and length totaled 32 caused the cluster to crash. (Bug #16125)

MySQL Cluster: Using mysqldump to obtain a dump of a partitioned table employing the NDB storage engine produced a nonfunctional table creation statement. (Bug #13155)

MySQL Cluster: NDB returned an incorrect Can't find file error for OS error 24; this has been changed to Too many open files. (Bug #15020)

MySQL Cluster: NDB leaked disk space when performing repeated INSERT or DELETE statements. (Bug #16771)

MySQL Cluster: Trying to import too many dumped tables requiring resources beyond those allocated in the cluster configuration file caused the server to crash instead of reporting an insufficient resources error. (Bug #16455)

MySQL Cluster: ndb_delete_all ran out of memory when processing tables containing BLOB columns. (Bug #16693)

MySQL Cluster: The ndb_autodiscover test failed sporadically due to a node not being permitted to connect to the cluster. (Bug #15619)

Disk Data: DROP LOGFILE GROUP corrupted the cluster file system and caused ndbd to fail when running more than one node on the same system. (Bug #16193)

Disk Data: NDB returned the wrong error when the tablespace on disk was full. (Bug #16738)

Disk Data: Tablespaces created using parameters with relatively low values (10 MB or less) produced filesizes much smaller than expected. (Bug #16742)

Disk Data: The error message generated by a failed ADD UNDOFILE did not provide any reasons for the failure. (Bug #16267)

Cluster API: Upon the completion of a scan where a key request remained outstanding on the primary replica and a starting node died, the scan did not terminate. This caused incomplete error handling for the failed node. (Bug #15908)

Creating a partitioned table using a storage engine other than the session default storage engine caused the server to crash. (Bug #15966)

The mysql_real_connect() C API function incorrectly reset the MYSQL_OPT_RECONNECT option to its default value. (Bug #15719)

Multi-byte path names for LOAD DATA and SELECT ... INTO OUTFILE caused errors. Added the character_set_filesystem system variable, which controls the interpretation of string literals that refer to file names. (Bug #12448)

ALTER TABLE ... ADD PARTITIONS on a table with one partition crashed the server. (Bug #15820)

Using RANGE partitioning with a CASE expression as the partitioning function would cause records to be placed in the wrong partition. (Bug #15393)

Using ENGINE=... within a PARTITION clause could cause the server to crash. (Bug #13438)

Certain subqueries where the inner query was the result of a aggregate function would return different results with MySQL 5.0 or 5.1 than with MySQL 4.1.

Subselects could also return wrong results when the query cache and grouping were involved. (Bug #15347)

The server would permit foreign keys to be declared in the definition of a partitioned table despite the fact that partitioned tables do not support foreign keys (see Restrictions and Limitations on Partitioning). (Bug #13446)

Using the TRUNCATE() function with a negative number for the second argument on a BIGINT column returned incorrect results. (Bug #8461)

ALTER TABLE ... DROP PARTITION would truncate all DATE column values in the table's remaining partitions to NULL. (Bug #13644)

In some cases the query optimizer did not properly perform multiple joins where inner joins followed left joins, resulting in corrupted result sets. (Bug #15633)

CREATE TABLE ... LIKE did not work if the table whose schema was to be copied was a partitioned table. (Bug #13435)

An indexing error sometimes caused values to be assigned to the wrong RANGE partition. (Bug #16684)

Temporary table aliasing did not work inside stored functions. (Bug #12198)

The error message for specifying values for which no partition exists returned wrong values on certain platforms. (Bug #15910)

When the full-text search parser plugin returned more words than half of the length (in bytes) of the query string, the server would crash. (Bug #16722)

An ALTER TABLE ... PARTITION BY ... statement did not have any effect. (Bug #15523)

Trying to add more than one partition in a single ALTER TABLE ... ADD PARTITION statement caused the server to crash. (Bug #16534)

Using an identifier rather than a literal integer value in the LESS THAN clause of a range-partitioned table could cause the server to crash and corruption of tables. (Bug #13439)

Improper memory handling for stored routine variables could cause memory overruns and binary log corruption. (Bug #15588)

An INSERT statement in a stored procedure corrupted the binary log. (Bug #16621)

Certain permission management statements could create a NULL host name for a user, resulting in a server crash. (Bug #15598)

STR_TO_DATE(1,NULL) caused a server crash. (Bug #15828, CVE-2006-3081)

Inserting a negative value into an integer column used as the partitioning key for a table partitioned by HASH could cause the server to crash. (Bug #15968)

Attempting to insert data into a partitioned table that used the BLACKHOLE storage engine caused mysqld to crash. (Bug #14524)

The absence of a table in the left part of a left or right join was not checked prior to name resolution, which resulted in a server crash. (Bug #15538)

With a table partitioned by LIST, inserting a value which was smaller than any value shown in the partitioning value-lists could cause the server to crash. (Bug #14365)

The DATA DIRECTORY and INDEX DIRECTORY clauses of a CREATE TABLE statement involving partitions did not work. (Bug #14354)

ALTER TABLE ... ADD PARTITION could crash the server or cause an Out of memory error in some circumstances. (Bug #13447)

Parallel builds occasionally failed on Solaris. (Bug #16282)

SHOW CREATE TABLE did not display the PARTITIONS clause for tables partitioned by HASH or KEY. (Bug #14327)

A FULLTEXT query in a prepared statement could result in unexpected behavior. (Bug #14496)

Specifying a value for --tmpdir without a trailing slash had unpredictable results. (Bug #15904)

Issuing a TRUNCATE TABLE statement twice in succession on the same partitioned table would cause the server to crash. (Bug #13442)

A SELECT from a key-partitioned table with a multi-column key could cause the server to crash. (Bug #13445)

Certain Japanese table names were not properly saved during a CREATE TABLE statement. (Bug #3906)

Using a REPLACE statement on a partitioned table caused the server to crash. (Bug #13440)