Newly recreated Supervisor VM is not able to join cluster even after it powered on successfully.
Issue observed after upgrade of vCenter to 7U3c from 7U2c in which supervisor k8s patch version v1.19.1 is not available on upgraded vCenter.
Running Supervisor Version -- 1.19.1
Available version in vCenter 7U3c -- 1.19.12, 1.20.8 and .1.21.0
We can see there is mismatch between supervisor vm ovf and available ovf rpm package version on vCenter terminal. In this scenario, if we recreate any of the VM then Supervisor cluster would go in degraded state.
Refer below command to check the same.
# /usr/lib/vmware-wcp/wcp-db-dump.py | grep master_vm_information -A 4 "master_vm_information": { "vm-2009": "v1.19.1+vmware.2-vsc0.0.9-17882987", "vm-2010": "v1.19.1+vmware.2-vsc0.0.9-17882987", "vm-2011": "v1.19.1+vmware.2-vsc0.0.9-17882987" }, # rpm -qa VMware-wcpovf* vmware-wcpovf-3.0.281-18735554.x86_64
Refer the last digits 17882987 and 18735554. Ideally, it should be same.
This was occurred due to mismatch in Supervisor k8s patch version running on Supervisor VM and new wcp-ovf available on vCenter after vCenter upgrade.
Fix is already been shipped in vCenter 7U3f to handle this patch version mismatch.
New recreated supervisor VM won't able to join existing cluster.