Some On Premise customers with clustered environments (two nodes or more) may be unable to log in or may experience significant login delays. This issue is likely caused by IIS recycling the application pools on all servers simultaneously, resulting in orphaned records.
To resolve this issue:
Stop all the nodes in the cluster, then delete all entries in both tables of the clustering database (tblClusterEngineGroup and tblClusterEngineGroupConnection). Once the tables are empty, the servers can be restarted.
To prevent this issue:
Configure a 15 to 30 second delay between IIS restarts to ensure the nodes are not restarted at the same time.