Details
-
Story
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
Description
Based on an IRC conversation it's not immediately obvious that a scheduler backup restore will need to wait for a reconciliation run to become fully effective. For example, restoring a full cluster given an empty set of slaves will leave a user with a bunch of RUNNING tasks that actually aren't.
My suggestion is to add a flag in the admin commit_recovery command that will force an immediate reconciliation once the new scheduler has come back up. It may also make sense to add a reconcile_now command as well.