This is a Monthly Rapid Update release of the MySQL Enterprise Server 5.0.
This section documents all changes and bug fixes that have been applied since the last MySQL Enterprise Server release (5.0.28).
Functionality Added or Changed
The ndb_config utility now accepts
-c as a short form of the
--bind-address option for
ndbd. This permits a data node process to be
bound to a specific network interface.
NDB storage engine could leak
memory during file operations.
variable was renamed to
HELP command in the Cluster
management client now provides command-specific help. For
HELP RESTART in
ndb_mgm provides detailed information about
If the user specified the server options
--table-cache=, a warning would be given in some cases that some
values were recalculated, with the result that
--table-cache could be assigned
In such cases, both the warning and the increase in the
--table-cache value were
completely harmless. Note also that it is not possible for the
MySQL Server to predict or to control limitations on the maximum
number of open files, since this is determined by the operating
The value of
--table-cache is no
longer increased automatically, and a warning is now given only
if some values had to be decreased due to operating system
CALL statement, stored
procedures that take no arguments now can be invoked without
parentheses. That is,
CALL p() and
CALL p are equivalent.
mysql_upgrade now passes all the parameters
specified on the command line to both
SHOW STATUS is no longer logged
to the slow query log.
mysqldump --single-transaction now uses
START TRANSACTION /*!40100 WITH CONSISTENT SNAPSHOT
*/ rather than
BEGIN to start
a transaction, so that a consistent snapshot will be used on
those servers that support it.
InnoDB exhibited thread thrashing with more
than 50 concurrent connections under an update-intensive
InnoDB showed substandard performance with
multiple queries running concurrently.
MySQL Cluster: Backup of a cluster failed if there were any tables with 128 or more columns. (Bug #23502)
MySQL Cluster: Cluster backups failed when there were more than 2048 schema objects in the cluster. (Bug #23499)
The management client command
ALL DUMP 1000
would cause the cluster to crash if data nodes were connected to
the cluster but not yet fully started.
MySQL Cluster: (NDB API): Inacivity timeouts for scans were not correctly handled. (Bug #23107)
MySQL Cluster: If a node restart could not be performed from the REDO log, no node takeover took place. This could cause partitions to be left empty during a system restart. (Bug #22893)
MySQL Cluster: Multiple node restarts in rapid succession could cause a system restart to fail , or induce a race condition. (Bug #22892, Bug #23210)
(NDB API): Attempting to read a nonexistent tuple using
Commit mode for
caused node failures.
--help output from
NDB binaries did not include
MySQL Cluster: (NDB API): Scans closed before being executed were still placed in the send queue. (Bug #21941)
MySQL Cluster: A scan timeout returned Error 4028 (Node failure caused abort of transaction) instead of Error 4008 (Node failure caused abort of transaction...). (Bug #21799)
The node recovery algorithm was missing a version check for
tables in the
ALTER_TABLE_COMMITTED state (as
opposed to the
which has the version check). This could cause inconsistent
schemas across nodes following node recovery.
MySQL Cluster: Partition distribution keys were updated only for the primary and starting replicas during node recovery. This could lead to node failure recovery for clusters having an odd number of replicas.
For best results, use values for
NumberOfReplicas that are even powers of 2.
MySQL Cluster: The ndb_mgm management client did not set the exit status on errors, always returning 0 instead. (Bug #21530)
Attempting to create an
on a MySQL with an existing non-Cluster table with the same name
in the same database could result in data loss or corruption.
MySQL now issues a warning when a
TABLES or other statement causing table discovery
finds such a table.
MySQL Cluster: Cluster logs were not rotated following the first rotation cycle. (Bug #21345)
When inserting a row into an
table with a duplicate value for a nonprimary unique key, the
error issued would reference the wrong key.
Condition pushdown did not work correctly with
MySQL Cluster: Under some circumstances, local checkpointing would hang, keeping any unstarted nodes from being started. (Bug #20895)
Using an invalid node ID with the management client
STOP command could cause
ndb_mgm to hang.
MySQL Cluster: Data nodes added while the cluster was running in single user mode were all assigned node ID 0, which could later cause multiple node failures. Adding nodes while in single user mode is no longer possible. (Bug #20395)
In some cases where
SELECT COUNT(*) from an
NDB table should have yielded an
MAX_INT was returned instead.
MySQL Cluster: Following the restart of a management node, the Cluster management client did not automatically reconnect. (Bug #19873)
Error messages given when trying to make online changes to
parameters such as
NoOfReplicas that can
only be changed using a complete shutdown and restart of the
cluster did not indicate the true nature of the problem.
MySQL Cluster: ndb_restore did not always make clear that it had recovered successfully from temporary errors while restoring a cluster backup. (Bug #19651)
In rare situations with resource shortages, a crash could result
from an insufficient number of
MySQL Cluster: ndb_mgm -e show | head would hang after displaying the first 10 lines of output. (Bug #19047)
MySQL Cluster: The error returned by the cluster when too many nodes were defined did not make clear the nature of the problem. (Bug #19045)
The ndb_config utility did not perform host
lookups correctly when using the
MySQL Cluster: A problem with takeover during a system restart caused ordered indexes to be rebuilt incorrectly. (Bug #15303)
Replication: Column names were not quoted properly for replicated views. (Bug #19736)
Transient errors in replication from master to slave may trigger
Got fatal error 1236: 'binlog truncated in the
middle of event' errors on the slave.
method, when called with the name of a nonexistent column,
caused a segmentation fault.
Cluster API: When multiple processes or threads in parallel performed the same ordered scan with exclusive lock and updated the retrieved records, the scan could skip some records, which as a result were not updated. (Bug #20446)
There was a race condition in the
References: This bug was introduced by Bug #15653.
yaSSL-related memory leaks were detected by Valgrind. (Bug #23981)
The internal SQL interpreter of
an unnecessary lock on the supremum record with
enabled. This caused an assertion failure when
InnoDB was built with debugging enabled.
M % 0
NULL, but (
M % 0) IS NULL
For not-yet-authenticated connections, the
Time column in
PROCESSLIST was a random value rather than
MySQL failed to build on Linux/Alpha. (Bug #23256)
References: This bug was introduced by Bug #21250.
cause a server crash for several operations on
MyISAM tables: repair table, create index by
sort, repair by sort, parallel repair, bulk insert.
The column default value in the output from
SHOW COLUMNS or
FROM INFORMATION_SCHEMA.COLUMNS was truncated to 64
mysql did not check for errors when fetching data during result set printing. (Bug #22913)
The return value from
my_seek() was ignored.
The optimizer failed to use equality propagation for
predicates with string arguments.
variable sometimes was incremented when no rollback had taken
If a table contains an
inserting into an insertable view on the table that does not
AUTO_INCREMENT column should not
change the value of
LAST_INSERT_ID(), because the
side effects of inserting default values into columns not part
of the view should not be visible. MySQL was incorrectly setting
LAST_INSERT_ID() to zero.
Instance Manager had a race condition involving mysqld PID file removal. (Bug #22379)
Some queries that used
GROUP BY could incorrectly return an empty
statement produced an error, the connection was silently
terminated with no error message. Now the server writes a
warning to the error log.
Use of a DES-encrypted SSL certificate file caused a server crash. (Bug #21868)
Use of a subquery that invoked a function in the column list of the outer query resulted in a memory leak. (Bug #21798)
Selecting from a
MERGE table could result in
a server crash if the underlying tables had fewer indexes than
MERGE table itself.
(Bug #21617, Bug #22937)
FLUSH TABLES WITH
READ LOCK followed by
TABLES, attempts to drop or alter a stored routine
failed with an error that the routine did not exist, and
attempts to execute the routine failed with a lock conflict
statements, storage engines were not notified of duplicate-key
Within a prepared statement,
SELECT (COUNT(*) =
1) (or similar use of other aggregate functions) did
not return the correct result for statement re-execution.
It was possible for a stored routine with a
latin1 name to cause a stack overrun.
TEMPORARY table with the same name
as an existing table that was locked by another client could
result in a lock conflict for
TABLE because the server unnecessarily tried to
acquire a name lock.
Incorrect results could be obtained from re-execution of a
parametrized prepared statement or a stored routine with a
SELECT that uses
JOIN with a second table having only one row.
Within a stored routine, a view definition cannot refer to routine parameters or local variables. However, an error did not occur until the routine was called. Now it occurs during parsing of the routine creation statement.
A side effect of this fix is that if you have already created
such routines, and error will occur if you execute
SHOW CREATE PROCEDURE or
SHOW CREATE FUNCTION. You
should drop these routines because they are erroneous.
In mysql, invoking
\r with very long
host_name parameters caused buffer
SHOW VARIABLES truncated the
Value field to 256 characters.
WITH ROLLUP could group unequal values.
Range searches on columns with an index prefix could miss records. (Bug #20732)
UPDATE that referred to a key
column in the
WHERE clause and activated a
trigger that modified the column resulted in a loop.
LIKE searches failed for indexed
utf8 character columns.
sql_mode = TRADITIONAL, MySQL
incorrectly aborted on warnings within stored routines and
mysqldump --xml produced invalid XML for
FLUSH INSTANCES in Instance Manager triggered
an assertion failure.
For a debug server, a reference to an undefined user variable in
a prepared statement executed with
EXECUTE caused an assertion
Within a trigger for a base table, selecting from a view on that base table failed. (Bug #19111)
DELETE IGNORE could hang for foreign key
InnoDB used table locks (not row locks)
within stored functions.
mysql would lose its connection to the server if its standard output was not writable. (Bug #17583)
mysql-test-run did not work correctly for RPM-based installations. (Bug #17194)
A client library crash was caused by executing a statement such
SELECT * FROM t1 PROCEDURE ANALYSE() using
a server side cursor on a table
t1 that does
not have the same number of columns as the output from
WITH CHECK OPTION for a view failed to
prevent storing invalid column values for
ALTER TABLE was not able to
rename a view.
A literal string in a
GROUP BY clause could
be interpreted as a column name.
Instance Manager didn't close the client socket file when starting a new mysqld instance. mysqld inherited the socket, causing clients connected to Instance Manager to hang. (Bug #12751)
Entries in the slow query log could have an incorrect
Lack of validation for input and output
TIME values resulted in several
SEC_TO_TIME() in some
cases did not clip large values to the
TIME range appropriately;
BIGINT UNSIGNED values as signed; only
truncation warnings were produced when both truncation and
(Bug #11655, Bug #20927)
A locking safety check in
InnoDB reported a
spurious error stored_select_lock_type is 0 inside
SELECT statements in
mode. The safety check was removed.