MySQL Cluster NDB 7.1.22 is a new release of MySQL Cluster,
incorporating new features in the
NDBCLUSTER storage engine and
fixing recently discovered bugs in previous MySQL Cluster NDB
Obtaining MySQL Cluster NDB 7.1. The latest MySQL Cluster NDB 7.1 binaries for supported platforms can be obtained from http://dev.mysql.com/downloads/cluster/. Source code for the latest MySQL Cluster NDB 7.1 release can be obtained from the same location. You can also access the MySQL Cluster NDB 7.1 development source tree at https://code.launchpad.net/~mysql/mysql-server/mysql-cluster-7.1.
This release also incorporates all bugfixes and changes made in previous MySQL Cluster releases, as well as all bugfixes and feature changes which were added in mainline MySQL 5.1 through MySQL 5.1.61 (see Changes in MySQL 5.1.61 (2012-01-10)).
A shortage of scan fragment records in
resulted in a leak of concurrent scan table records and key
DUMP 2303 in the ndb_mgm
client now includes the status of the single fragment scan
record reserved for a local checkpoint.
Cluster Replication: DDL statements could sometimes be missed during replication channel cutover, due to the fact that there may not be any epochs following the last applied epoch when the slave is up to date and no new epoch has been finalized on the master. Because epochs are not consecutively numbered, there may be a gap between the last applied epoch and the next epoch; thus it is not possible to determine the number assigned to the next epoch. This meant that, if the new master did not have all epochs, it was possible for those epochs containing only DDL statements to be skipped over.
The fix for this problem includes modifications to mysqld binary
logging code so that the next position in the binary log
COMMIT event at the end of an
epoch transaction, as well as the addition of two new columns
next_position to the
mysql.ndb_binlog_index table. In addition, a
new replication channel cutover mechanism is defined that
employs these new columns. To make use of the new cutover
mechanism, it is necessary to modify the query used to obtain
the start point; in addition, to simplify prevention of possible
errors caused by duplication of DDL statements, a new shorthand
ddl_exist_errors is implemented for use
with the mysqld option
--slave-skip-errors. It is highly
recommended that you use this option and value on the new
replication slave when using the modified query.
For more information, see Implementing Failover with MySQL Cluster Replication.
Note that the existing replication channel cutover mechanism continues to function as before, including the same limitations described previously. (Bug #11762277, Bug #54854)