Hi
I’m looking for guidance on the best practice to upgrade a Metro Availability setup with a Witness VM and would like to confirm the recommended and supported approach.
Scenario:
-
Two AHV clusters configured with Metro Availability
-
External Witness VM providing quorum on a third site
-
Prism Central deployed (one per cluster)
Upgrade-related questions:
-
During maintenance, is it recommended or even possible to:
-
Temporarily pause or disable the Witness service
-
Upgrade Prism Central
-
Re-enable the Witness afterward
Or is this not required / not supported?
-
-
Can Prism Central be upgraded without impacting Metro Availability, assuming both clusters remain healthy and communication between them is intact?
-
My concern is the following scenario:
-
During a Prism Central upgrade, communication with the Witness is temporarily lost due to the reboot of the VM or the services
-
Both clusters are still communicating correctly with each other all time
In that case, should the environment continue operating normally, or could this trigger any automatic failover / stonith / VM stun behavior?
-
-
Regarding the Witness VM lifecycle:
-
How often should the Witness VM be updated?
-
Should it be upgraded every time Prism Central or AOS is upgraded?
-
Or is the Witness typically a “deploy once, rarely touch” component, updated only for major compatibility or security reasons?
-
Thanks in advance!
