VCS cluster node lose a fencing race using a Coordination Point as the cpserver database is locked
Article ID: 100046679
Updated On:
Description
Error Message
The cpserver log will have entries similar to the following:
2019-08-29 16:05:30 CPS CRITICAL V-97-1400-501 Error writing to database! :database is locked
2019-08-29 16:05:30 CPS ERROR V-97-1400-143 Error unregistering node <{12345678-abcd-1234-abcd-1234567890abc}, 1>
2019-08-29 16:05:30 CPS INFO V-97-1400-146 CPS: unreg_node_in_config: END
2019-08-29 16:05:30 CPS ERROR V-97-1400-62 Error in unregistering victim node 1 (server02)
2019-08-29 16:05:30 CPS INFO V-97-1400-405 One or more victim nodes could not be preemptedCause
In the case that there are multiple database write requests at the same time, the cpserver is unable to cope with the many simultaneous requests, causing the un-register request to fail.
Resolution
The issue effects 6.2 version of VCS and is fixed in HotFix 6.2.1.301 and above. Please request the hotfix from support.
Issue/Introduction
VCS cluster node lose a fencing race using a Coordination Point as the cpserver database is locked
Additional Information
ETrack: 3939537
JIRA: STESC-3622
Was this article helpful?
Yes
No
Powered by
