Details
-
New Feature
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
2.5.14
-
None
-
MXS-SPRINT-184
Description
If master becomes unresponsive (eg: InnoDB clogged by contention issues) we want to be able to promote a replica and then kill the unrecoverable master.
This is currently not possible:
2022-02-03 13:44:35 error : [mariadbmon] Failed to enable read_only on 'brokercloudprod-db-mdb-ms-0': Query 'SET STATEMENT max_statement_time=3 FOR SET GLOBAL read_only=1;' failed on 'brokercloudprod-db-mdb-ms-0': 'Query execution was interrupted (max_statement_time exceeded)' (1969).
|
Proposal: either move forward even if read only cannot be set after some time, or provide an option to switchover to force this when need be. Thank you.
Attachments
Issue Links
- relates to
-
MXS-4759 Force failover flag
- Closed