Impact of vVol Datastore recovery out of an Unplanned PDL Event.

Products

• VMware vSphere ESXi

Issue/Introduction

Symptoms:

• Once the vVol datastore is recovered out of PDL the datastore fails to connect back to the host.
• vVols datastores do not automatically recover (or connect back) on affected hosts after the PDL event remediation.
• Unmount or detaching vVols datastore fails.

Environment

VMware vSphere ESXi 8.0.0

Cause

During a PDL on vVol datastore, the vVol datastore may fail to recover or connect back to the host even post recovery from PDL of the datastore.

Resolution

Since VMCP is not supported for vVol datastore the Virtual Machines in the vVol datastore may not get terminated. This non-terminated Virtual Machines may hold a reference on the vVol datastore failing the recovery of vVol  datastore even post recovery of PDL event. In such case the Virtual Machines has to be terminated manually in order to recover the vVol datastore post PDL event.

Workaround:

If a vVold datastore fails to recover post a PDL event.

1. Check whether any worlds is holding a reference on the vVol datastore.

Note down the world ID's from the below command.
naa.624a9370b3a54a07e99f44760016b55e being the PE which is backing the container or the vVol datastore that failed to recover.
esxcli storage core device world list -d naa.624a9370b3a54a07e99f44760016b55e (World ID : 3002128)
Ex :
[root@w2-hs4-m0403:~] esxcli storage core device world list -d naa.624a9370b3a54a07e99f44760016b55e
Device                                World ID  Open Count  World Name
------------------------------------  --------  ----------  ----------
naa.624a9370b3a54a07e99f44760016b55e   3002128           1  vmm0:vvol1-10.208.4.153-T_scsi_vvol_fc_32tb_Pure11_1-rhel7-2020-vmwpv-lc-0034

2. Kill all the worlds using esxcli command that are holding a reference on the PDL vVol datastore. (All the worlds one by one if more than one world holding the reference).

esxcli vm process kill -t soft -w <World ID holding lock on this device>
(Ex:
[root@w2-hs4-m0403:~] esxcli vm process kill -t soft -w 3002128
)

3. Make sure that no more world is holding reference anymore on the device.

esxcli storage core device world list -d naa.624a9370b3a54a07e99f44760016b55e
(Ex :
[root@w2-hs4-m0403:~] esxcli storage core device world list -d naa.624a9370b3a54a07e99f44760016b55e
Device  World ID  Open Count  World Name
------  --------  ----------  ----------
)

Post this the device should get refreshed\recovered and get connected back to the hosts if the PDL is recovered for that device in that specific host.