Bug #58698 Epoch not in ndb_apply_status when Ignoring binary logging mysql db on master
Submitted: 3 Dec 2010 14:23 Modified: 1 Apr 2011 10:14
Reporter: Geert Vanderkelen Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Cluster: Replication Severity:S3 (Non-critical)
Version:mysql-5.1-telco-7.1 OS:Any
Assigned to: Frazer Clement CPU Architecture:Any

[3 Dec 2010 14:23] Geert Vanderkelen
Description:
When we ignore the binary logging for the mysql database on the master, the epoch is not saved in the mysql.ndb_apply_status on the Slave.

How to repeat:
Running master with --binlog-ignore-db=mysql, we update some data, then check the slave:

slave> select * from mysql.ndb_apply_status;
+-----------+-------+------------------+-----------+---------+
| server_id | epoch | log_name         | start_pos | end_pos |
+-----------+-------+------------------+-----------+---------+
|       100 |     0 | machA-bin.000003 |       107 |     377 |
+-----------+-------+------------------+-----------+---------+

Suggested fix:
We should make sure updates on mysql.ndb_apply_status are going to the binary log anyway, even when --binary-ignore-db=mysql is used.
[1 Apr 2011 10:14] Jon Stephens
Documented bugfix in the NDB-6.3.42, 7.0.23, and 7.1.12 changelogs, as
follows:

      Filtering the binary log using --binlog-ignore-db=mysql caused
      epochs not to be logged in the ndb_apply_status table.

Closed.