Details
-
Task
-
Status: Open (View Workflow)
-
Major
-
Resolution: Unresolved
-
25.01, 25.10, 26.10
-
None
-
MXS-SPRINT-283
Description
During switchover, the monitor only relaxes Con/C timeouts to the demoting server, as that can get stuck on commands like set global read_only=1 and stop slave.
However, with semisync replication, a replica can also get stuck running stop slave. Limit this issue by also relaxing timeouts to the promotion target, or perhaps all replicas. This is mainly an issue with failover, due to MDEV-42004.