Details
-
Task
-
Resolution: Fixed
-
Major
-
None
-
None
-
None
-
10 - Path of Vengeance
-
3
Description
Currently we do not check to see if it is safe to proceed with a graceful or hard failover immediately prior to executing any kind of failover.
Our expectations should be:
1) The nodes we know about are part of the cluster and marked as active via /pools/default/
2) There are 0 inactiveFailed nodes
3) There are 0 inactiveAdded nodes
This check should be performed immediately prior to any failover or graceful failover of a node as closely as possible. In the future, we will only do this on versions that do not have MB-61888.
If this checks fail, we should abort any upgrades and allow reconciliation to recover the cluster prior to continuing any upgrades.