Vcenter vSAN corrupter

[jcastillo88]

HI Guys.

This is my setup  1x Vcenter 8.0 and 3x DELL VSAN nodes.

I have a question guys. Our VCSA who managed the vSAN cluster nodes get corrupted  and our backup failed.

If I am going to reinstall the vCenter with the same fqdn name and IP address, what will happen if I add the existing 3 nodes to newly reinstalled VCenter? My VMs and existing vSAN datastore is still accessible right?

Thank you in advance.

Jess Castillo

[drdoc_fred]

When you reinstall the vCenter Server and add the existing 3 nodes to it, vCenter Server will discover all the virtual machines on those managed hosts and add them to the vCenter Server inventory. If a managed host is disconnected and reconnected, any changes to the virtual machines on that managed host are identified, and the vSphere Client updates the list of virtual machines. So, your VMs and existing vSAN datastore should still be accessible.

see https://docs.vmware.com/en/VMware-vSphere/8.0/vsphere-vm-administration/GUID-376174FE-F936-4BE4-B8C2...

[TheBobkin]

Hi Jess,

You first need to prevent this new vCenter from breaking the vSAN cluster while adding it back to inventory - this can be done by temporarily setting the nodes to ignore changes to the unicastagent lists being pushed by vCenter (e.g. in case it pushes a blank or partial list while adding 1st/2nd/3rd node):
Run on all nodes:
# esxcfg-advcfg -s 1 /VSAN/IgnoreClusterMemberListupdates

Next you want to redeploy vCenter and recreate the vSphere cluster in inventory - don't enable vSAN on this yet - add the hosts to vCenter inventory (just at Datacenter level), then reconfigure all the networks to match existing vDS etc. and configure these on the hosts, then move all the nodes into the vSphere cluster, then enable vSAN and ensure to enable the exact features as they were (e.g. if you had deduplication&compression enabled then it should be enabled here etc.). If you don't know which features were enabled then advise checking these ('esxcli vsan storage list' informs of most of them).

Then confirm that cluster still formed and Skyline Health doesn't show any serious warnings relating to cluster or data health - it should at this point still have a red error for 'vCenter is Authoritive', this is because unicastagent is still set to ignore, this can then be reverted ('esxcfg-advcfg -s 0 /VSAN/IgnoreClusterMemberListupdates' on all nodes) and re-push/check of ESXi configuration can be applied from 'vCenter is Authoritive' health check.