MySQL Bugs: #49868: RESET SLAVE should not clear ndb_binlog

Bug #49868	RESET SLAVE should not clear ndb_binlog_index
Submitted:	22 Dec 2009 5:31	Modified:	14 Feb 2010 8:26
Reporter:	Mikiya Okuno	Email Updates:
Status:	No Feedback	Impact on me:	None
Category:	MySQL Cluster: Replication	Severity:	S3 (Non-critical)
Version:	mysql-5.1-telco-6.2	OS:	Any
Assigned to:	Assigned Account	CPU Architecture:	Any
Tags:	6.2.x

Description:
RESET SLAVE clears not only ndb_apply_status but also ndb_binlog_index at this moment. However, it must clear only a slave-related table, ndb_apply_status. ndb_binlog_index is a master table, so RESET SLAVE should not clear ndb_binlog_index. ndb_binlog_index should be cleared by RESET MASTER.

This problem affects multi-master setup.

How to repeat:
1. Configure dual-master cluster replication.
2. Kill one cluster.
3. Restore data from surviving cluster.
4. Re-configure cluster replication.
5. RESET SLAVE will clear on the surviving cluster.

Suggested fix:
Do not let RESET SLAVE to reset ndb_binlog_index.

Cannot reproduce.
The code looks as follows, so a purging (clearing)
of ndb_binlog_index should not be done by the slave.

static int
ndbcluster_binlog_index_purge_file(THD *thd, const char *file)
{
  if (!ndb_binlog_running || thd->slave_thread)
    return 0;
... 

Please attach all logs from the scenario for analysis.

No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".