חדשות, עדכונים, מדריכים ועזרים | עדכוני תוכנות ואפליקציות - (09.08.12) - גרסה חדשה: MySQL 5.1.65

(09.08.12) - גרסה חדשה: MySQL 5.1.65

עדכוני תוכנות ואפליקציות

חדשות, עדכונים, מדריכים ועזרים


פרטים נוספים:
http://dev.mysql.com/downloads/

מה חדש:

Changes in MySQL 5.1.65 (August 9, 2012)
 
Functionality Added or Changed
 
* Important Change: The YEAR(2) data type is now deprecated
 because it is problematic. Support for YEAR(2) will be removed
 in a future release of MySQL. For more information, see
 Section 11.3.4, "YEAR(2) Limitations and Migrating to
 YEAR(4)."
 
* Important Change: Replication: The SHOW BINARY LOGS statement
 (and its equivalent SHOW MASTER LOGS) may now be executed by a
 user with the REPLICATION CLIENT privilege. (Formerly, the
 SUPER privilege was necessary to use either form of this
 statement.)
 
Bugs Fixed
 
* The server did not build with gcc 4.7. (Bug #14238406)
 
* InnoDB: If a row was deleted from an InnoDB table, then
 another row was re-inserted with the same primary key value,
 an attempt by a concurrent transaction to lock the row could
 succeed when it should have waited. This issue occurred if the
 locking select used a WHERE clause that performed an index
 scan using a secondary index. (Bug #14100254, Bug #65389)
 
* InnoDB: In a transaction using the REPEATABLE READ isolation
 level, an UPDATE or DELETE statement for an InnoDB table could
 sometimes overlook rows recently committed by other
 transactions. As explained in Consistent Nonlocking Reads
 (http://dev.mysql.com/doc/refman/5.1/...onsistent-read
 .html), DML statements within a REPEATABLE READ transaction
 apply to rows committed by other transactions, even if a query
 could not see those rows. (Bug #14007649, Bug #65111)
 
* InnoDB: Using the KILL statement to terminate a query could
 cause an unnecessary message in the error log:
 [ERROR] Got error -1 when reading table table_name
 (Bug #13933132)
 
* InnoDB: For an InnoDB table with a trigger, under the setting
 innodb_autoinc_lock_mode=1, sometimes auto-increment values
 could be interleaved when inserting into the table from two
 sessions concurrently. The sequence of auto-increment values
 could vary depending on timing, leading to data inconsistency
 in systems using replication. (Bug #12752572, Bug #61579)
 
* InnoDB: The CHECK TABLE statement could fail for a large
 InnoDB table due to a timeout value of 2 hours. For typical
 storage devices, the issue could occur for tables that
 exceeded approximately 200 or 350 GB, depending on I/O speed.
 The fix relaxes the locking performed on the table being
 checked, which makes the timeout less likely. It also makes
 InnoDB recognize the syntax CHECK TABLE QUICK, which avoids
 the possibility of the timeout entirely. (Bug #11758510, Bug
 #50723)
 
* Replication: It was theoretically possible for concurrent
 execution of more than one instance of SHOW BINLOG EVENTS to
 crash the MySQL Server. (Bug #13979418)
 
* Replication: An event whose length exceeded the size of the
 master dump thread's max_allowed_packet caused replication to
 fail. This could occur when updating many large rows and using
 row-based replication.
 As part of this fix, a new server option
 --slave-max-allowed-packet is added, which permits
 max_allowed_packet to be exceeded by the slave SQL and I/O
 threads. Now the size of a packet transmitted from the master
 to the slave is checked only against this value (available as
 the value of the slave_max_allowed_packet server system
 variable), and not against the value of max_allowed_packet.
 (Bug #12400221, Bug #60926)
 
* Replication: Statements using AUTO_INCREMENT,
 LAST_INSERT_ID(), RAND(), or user variables could be applied
 in the wrong context on the slave when using statement-based
 replication and replication filtering server options (see How
 Servers Evaluate Replication Filtering Rules
 (http://dev.mysql.com/doc/refman/5.1/...ion-rules.html
 )). (Bug #11761686, Bug #54201)
 References: See also Bug #11754117, Bug #45670, Bug #11746146,
 Bug #23894.
 
* Replication: An INSERT into a table that has a composite
 primary key that includes an AUTO_INCREMENT column that is not
 the first column of this composite key is not safe for
 statement-based binary logging or replication. Such statements
 are now marked as unsafe and fail with an error when using the
 STATEMENT binary logging format. For more information, see
 Determination of Safe and Unsafe Statements in Binary Logging
 (http://dev.mysql.com/doc/refman/5.1/...ion-rbr-safe-u
 nsafe.html), as well as Replication and AUTO_INCREMENT
 (http://dev.mysql.com/doc/refman/5.1/...ion-features-a
 uto-increment.html).
 Note
 Tables using the InnoDB storage engine are not affected by
 this issue, since InnoDB does not allow the creation of a
 composite key that includes an AUTO_INCREMENT column, where
 this column is not the first column in the key.
 (Bug #11754117, Bug #45670)
 References: See also Bug #11761686, Bug #54201, Bug #11746146,
 Bug #23894.
 
* Replication: After upgrading a replication slave to MySQL
 5.5.60 or later, enabling the query cache eventually caused
 the slave to fail. (Bug #64624, Bug #14005409)
 
* Incorrect stored program caching could cause statements within
 a stored program that included a GROUP BY clause to return
 different results across multiple program invocations. (Bug
 #13805127)
 
* For queries with ORDER BY COUNT(*) and LIMIT, the optimizer
 could choose an execution plan that produced incorrect
 results. (Bug #12713907)
 
* SHOW TABLES was very slow unless the required information was
 already in the disk cache. (Bug #60961, Bug #12427262)
 
* mysqlbinlog exited with no error code if file write errors
 occurred. (Bug #55289, Bug #11762667)
 
* yaSSL rejected valid SSL certificates that OpenSSL accepts.
 (Bug #54348, Bug #11761822)
 
* Sessions could end up deadlocked when executing a combination
 of SELECT, DROP TABLE, KILL, and SHOW ENGINE INNODB STATUS.
 (Bug #60682, Bug #12636001)
 
Thanks,
 On Behalf of Oracle MySQL RE Team