When decommissioning a vSAN cluster, the disk groups should be deleted as per the following KB article:
Steps to decommission the vSAN datastore.
Failure to perform this step may result in being unable to delete them when the hosts have been reinstalled using an earlier ESXi release.
vSAN OSA 7.x
vSAN OSA 8.x
Failure to delete vSAN disk groups when decommissioning a vSAN cluster and subsequently re-installing a lower version of ESXi that is incompatible with the higher vSAN on-disk format version causes the issue.
Refer to the following KB:
Understanding vSAN on-disk format versions and compatibility
Attempts to delete an Unhealthy disk group through the GUI or command line fail and the following type of entries will be observed in the vmkernel.log on the host:
WARNING: PLOG: PLOGProbeDevice:6857: Wrong metadata configuration read from disk naa.################################:1: Number of mapped disks: 1, Metadata version: 4, Disk format version: 20,WARNING: PLOG: PLOGFindDisks:1708: Invalid disk args suppliedPLOG: PLOG_ExecVSIOp:2233: PLOG Op DISKGROUP DESTROY for SSD:1 completed with status Failure
Need to select Erase Partitions for all vSAN disks as per the instructions in the following KB article:
vSAN disk shows as Ineligible and cannot be claimed for vSAN disk group
Once the partitions have been erased, the disks will be available to create vSAN disk groups.