After following standard procedures to unmount and detach a LUN from ESXi hosts, administrators may observe "All Paths Down" (APD) alerts in vCenter or entries in the /var/log/vmkernel.log.
2026-06-25T00:28:31.049Z In(182) vmkernel: cpu44:2098030)StorageApdHandlerEv: 106: Device or filesystem with identifier [naa.600a09803830477174244e6c35725570] has entered the All Paths Down state.
ESXi 8.x
vCenter 8.x
This behavior is by design within the ESXi Pluggable Storage Architecture (PSA). Even after a device is detached, ESXi continues to periodically probe the paths to monitor for state changes. This ensures the system can report whether a device is available for re-attachment. If the LUN is unmapped at the storage array or a link error occurs during these probes, the PSA triggers an APD state to alert the administrator of the device's status.
If standard detach procedures lead to persistent APD alerts, follow the steps below to prevent the host from reporting the device state.
Warning: Disabling autoclaim is a system-wide setting. Ensure it is re-enabled immediately after the procedure (Step 5) to prevent health issues for other LUNs.
Turn the device off: esxcli storage core device set --state=off -d <device id>
Disable autoclaim on the ESX host: esxcli storage core claiming autoclaim --enabled=false
Unclaim the device: esxcli storage core claiming unclaim -t device -d <device id>
Unmap the LUN/device from the storage array: Complete steps 1 through 3 on all affected ESXi hosts before unmapping the device at the storage layer.
Enable autoclaim on ESX hosts: esxcli storage core claiming autoclaim --enabled=true
Detach a LUN device from ESXi hosts
https://knowledge.broadcom.com/external/article/323128/how-to-detach-a-lun-device-from-esxi-hos.html
Removing a LUN containing a datastore from VMware ESXi/ESX
https://knowledge.broadcom.com/external/article/315416/removing-a-lun-containing-a-datastore-fr.html