Decommission an offline VCSA

Share on:

Not long ago I decided to retire a vCenter from my lab - but I didn't retire the VCSA properly, leaving its linked mode partners with an error. Offline VCSA Error

While there is a KB for this, I wanted to share some notes on my experience. The key piece of information that you need is the Primary Network Identifier of the "lost" node. Typically this is the FQDN of the appliance or in rarer instances the IP address. If you aren't sure what the pnid of the node you wish to remove is, and it's truly not available, there are some commands like vdcrepadmin that you can use to poke around and find that information.

Once you've found the commands you are looking for it's a pretty straight forward process to clean things up, although I should probably remind you about the importance of making sure the node you are removing in this fashion never ever comes back online again. That said, the cleanup should be nothing more complicated than running a single cmsso-util command:

 1root@lab-comp02-vcenter [ ~ ]# cmsso-util unregister --node-pnid lab-mgmt01-vcenter.lab.sentania.net --username administrator@vsphere.local
 2    Password:
 3    Solution users, computer account and service endpoints will be unregistered
 4    2021-01-02T22:23:12.063Z  Running command: ['/usr/lib/vmware-vmafd/bin/dir-cli', 'service', 'list', '--login', 'administrator@vsphere.local']
 5    2021-01-02T22:23:12.111Z  Done running command
 6    Stopping all the services ...
 7    All services stopped.
 8    Starting all the services ...
 9    Started all the services.
10    Success

If all goes well, you shouldn't see any more error messages related to the missing member.