Aria Operations upgrade stuck at step 8 of 14 due to Witness node in non-CA cluster

Products

• VCF Operations/Automation (formerly VMware Aria Suite)

Issue/Introduction

When upgrading VMware Aria Operations, the process may hang at step 8 of 14. This can occur when the cluster contains a Witness node but does not have Continuous Availability (CA) enabled.

• Aria Operations upgrade from 8.18.6 to 8.18.7 hangs at step 8 of 14 for several hours.
• The Admin UI Software Update page shows the upgrade stalled on the primary node.
• Logs in /var/log/vmware/vcops/vcops_upgrade.log show the process waiting or failing during OS updates.

Environment

VMware Aria Operations 8.18.x

Cause

Witness nodes are only supported and required when Continuous Availability (CA) is enabled. If a Witness node was manually added or left over in a standard cluster, the upgrade scripts may fail to validate the cluster state correctly, causing the hang.

Resolution

1. Take cluster offline snapshots of all nodes in the Aria Operations cluster
2. Log in to the Aria Operations Admin UI (https://<primary-node-FQDN>/admin).
3. If the upgrade is still "Running," attempt to let it time out or Contact Broadcom Support if it remains stuck for >4 hours.
4. Once the cluster is accessible, navigate to the System Status page.
5. Identify the Witness Node and select Remove Node.
6. After the Witness node is removed, verify the cluster status is healthy.
7. Re-apply the Aria Operations upgrade PAK file.
8. Select the checkbox Install the PAK file even if it is already installed and proceed.

Additional Information

For assistance with complex cluster recovery, see Contact Broadcom Support.

To be updated on future fix statuses, subscribe to this article.