We see that that the offline sessions loading is going to hung state in a cluster environment. RH-SSO obtain DB lock twice which are in the following situation

1. Obtained to check if we need a database upgarde.
2. To lock to do the load

In a cluster of 6 members, we found that some instances are locked in the step-1 and others are in step-2 and loading of sessions always failed because the jobs of nodes that are executing step-1 failed to load the segment with error "KeycloakSessionFactory not yet set in cache. Worker skipped" The reason is that the those nodes have not started to do the session loading and they don't have the keycloak provider set in the work cache (done at the init cache). As the segments are loaded in all the nodes, all the nodes in step-1 failed, and the segments are re-tried to infinity.

This issue does not occurs starting cluster member one after another.