One or more objects are inaccessible after improper replacement of witness appliance in a stretch cluster
When running:
esxcli vsan debug object list --health=inaccessible
Object layout looks similar to below:
VMware vSAN (All versions)
Quorum lost due to object not having enough votes to meet quorum, resulting in object going inaccessible
Please see the following for more info on inaccessible objects in vSAN: vSAN Health Service - Data Health - vSAN Object Health - inaccessible objects
Open up a support SR with Broadcom for assistance with recovery analysis and options.