When upgrading VMware Identity Broker (vIDB) from version 9.0.2 to 9.1.0, the upgrade workflow halts during the "Perform Component Upgrade" sequence. The task times out after approximately 12 to 15 minutes with the error: "Prepare Postgres Migration. Status: Failed."
VMware Identity Broker 9.1
This issue occurs because the archive_mode parameter on a database replica pod (such as vidb-postgres-instance-1) is stuck in an off state. During the upgrade task, the automation sequence attempts to change the archive mode status to on but fails, resulting in a timeout that fails the entire upgrade process.
This issue is under review with Broadcom Engineering and is slated to be fixed in a future release. Please subscribe to this article to be updated on the fix status refer to Subscribe to a Broadcom knowledge article by article or product
Workaround:
To successfully bypass this issue and complete the upgrade, ensure that the primary vidb-postgres-instance-0 is the elected leader to prevent switchovers during the shutdown process:
kubectl exec #### -n vidb-external -- patronictl listNote: Update vidb-postgres-instance-x with the current leader.kubectl exec -it #### -n vidb-external -- patronictl switchover --leader #### --candidate #### --forcekubectl exec -it #### -n #### -- psql -U postgres -c "SHOW archive_mode; SHOW archive_command; SHOW archive_status;"
kubectl exec -it <postgres-pod> -n <namespace> -- psql -U postgres -c "SHOW archive_mode; SHOW archive_command; SHOW archive_status;"