Datastore un-mounting | Dead path issue (APD)

Vsphere 5.1 onwards the process of removal datastore has been changed, before removing any datastore from ESXi hosts and cluster Right-click the datastore and unmounting.

It is not only the process to remove LUN from ESXi hosts but there are few additional pre-checks and post tasks like detaching the device from the host is must before we request storage administrator to unpresenting the LUN from the backend storage array.

This process needs to be followed properly otherwise it may cause bad issue like Dead path or APD ( All paths down) condition on the ESXi host.

1.  If the LUN is being used as a VMFS datastore, all objects (such as VM, snapshots ,template and HA configuration stored on the VMFS datastore.
2. Ensue the Datastore is not used for vSphere HA heartbeat.
3. Ensure the datastore is not part of a SDRS.
4. Storage I/O control should be disabled.
5. ISO mount and scripts or utilities are accessing the datastore.

Process to remove datatstore or LUN  from ESXi 5.x hosts

• Select the ESXi host -> Configuration - > Storage - > Datastores

(note down the naa id for the datatore which starts from naa.xxx)

•                 Right-click the datastore, which you want to unmounts and select unmounts.

•    Confirm datastore unmounts pre-check is all marked with Green check mark and click on OK.
•     Select the ESXi host -> Configuration -> storage -> devices Match the devices with the naa.id (naa.xx) Right-click the Device and Detach. Verify all the Green checks and click on OK to detach the LUN.

           Repeat the same step for all ESXi hosts, where you want to unpresent this datastore.

•      Inform your storage team to physically unpresent the LUN from the ESXi host using the appropriate array tools.

Rescan the ESXi host and verify detached LUNs are disappeared from the ESXi host.