Vxrail cluster Decom
Hi All,
i'm in charge of decom a VXrail cluster soon and unfortunetly our Dell support contract is not up anymore, so i'm alone on this point.
I wanted to know if someone can confirm my process or have one to follow ?
Environment:
- VxRail 8.0.370-29422450, standalone (customer-managed vCenter, not VCF/SDDC Manager
- Stretched Cluster (2 fault domains + witness)
- Same vCenter also hosts a second, unrelated VxRail cluster
Goal: fully remove this cluster from vCenter cleanly, without breaking the other cluster or the shared NFS.
Planned approach (unconfirmed):
- Confirm datastore empty (
esxcli vsan debug vmdk list) - Remove hosts from shared VDS uplinks/port groups (not deleting the VDS itself)
- vSAN decommission per Broadcom KB 392097: retreat mode → remove disk groups (Maintenance Mode, "No data migration") → manually remove Fault Domains + Witness (stretched-specific) → disable vSAN VMkernel service → turn off HA → turn off vSAN service cluster-wide
- Put all hosts in Maintenance Mode, then delete the vSphere cluster object directly (not host-by-host removal, since that path is blocked)
- Manually delete the VxRail Manager VM for this cluster afterward, accept potential residual/orphaned state in VxRail Manager's internal DB (hardware is being destroyed anyway, no future redeploy planned)
- Peripheral cleanup: DNS, AD, iDRAC/OMIVV/SCG, firewall rules, Commvault/backup jobs, licensing
My open question: has anyone actually deleted a standalone (non-VCF) VxRail cluster object directly from vCenter after zeroing out vSAN, without going through VxRail Manager's node-removal workflow? Did it leave VxRail Manager in a broken state, and did that actually matter for you (especially if hardware wasn't being reused)?
Any first-hand experience, even partial, appreciated — trying to avoid surprises mid-teardown with no Dell safety net.
Thanks in advance !