This article addresses scenarios where VMware vSAN reports a Permanent Device Loss (PDL) or "Unhealthy" disk alert, but physical hardware management consoles (such as iDRAC, xClarity, or iLO) report the drive as healthy. This discrepancy typically indicates a controller-level I/O failure on the NVMe device rather than a physical removal or connectivity issue.
Symptoms:
esxcli vsan storage list shows the device as missing or "in CMMDS: false."esx.problem.vob.storage.connectivity.lost or NVMEDEV: Controller failed to recover.The hardware console performs out-of-band electrical presence checks, which only verify if the drive is physically seated in the PCIe slot. ESXi checks for active storage path responsiveness under I/O load. If the NVMe controller on the drive fails to process I/O requests, ESXi will mark the device as PDL even if the drive is technically "powered on" and visible to the server BIOS/iDRAC.
# esxcli vsan storage list.vmkwarning logs for controller recovery failures (e.g., "Controller failed to recover after X attempts").