Bug #21816 Slave Cluster restart produces epoch issues for slave MySQLD
Submitted: 24 Aug 2006 18:13 Modified: 24 Aug 2006 18:36
Reporter: Jonathan Miller Email Updates:
Status: Duplicate Impact on me:
None 
Category:MySQL Cluster: Replication Severity:S3 (Non-critical)
Version:5.1.12 OS:Linux (Linux)
Assigned to: CPU Architecture:Any

[24 Aug 2006 18:13] Jonathan Miller
Description:
I thought that this might be the same as the one Jonas reported in #21806, but in talking with Martin, he stated it was not.

Shuting down and restarting the slave cluster produces:
060824 16:23:09 [ERROR] NDB Binlog: cluster has been restarted --initial or with older filesystem. ndb_latest_handled_binlog_epoch: 92726, while current epoch: 0. RESET MASTER should be issued. Resetting ndb_latest_handled_binlog_epoch.

The slave process continues most of the time.

How to repeat:
Start slave replication, shutdown cluster, restart cluster.
[24 Aug 2006 18:36] Jonas Oreland
it the same (case 1)