Bug #50979 COPYFRAG should retry on temporary errors
Submitted: 8 Feb 2010 7:42
Reporter: Andrew Hutchings Email Updates:
Status: Verified Impact on me:
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S4 (Feature request)
Version:mysql-5.1-telco-6.3 OS:Any
Assigned to: CPU Architecture:Any

[8 Feb 2010 7:42] Andrew Hutchings
When a node is starting it is possible for a node to fail due to temporary errors during COPYFRAG.  For example:

Error data: Killed by node 5 as copyfrag failed, error: 1501
Error object: NDBCNTR (Line: 270) 0x0000000a

1501 - Out of UNDO log

For these kind of errors the copyfrag should probably be retried a certain number of times (maybe with a short pause between) for temporary errors and print a warning in the node's log, this may reduce the node failures for them.

How to repeat:
[25 May 2012 10:24] Gustaf Thorslund
Bug 60654 marked as duplicate