Datastore is missing after SAN power failure but LUN is presented and visible on ESX.

Resolution:

To resolve this issue:

1. To identify the devices presented to the ESX Server, run the command:
   
   esxcfg-vmhbadevs -q

2. To identify datastores and the associated devices, run the command:
   
   esxcfg-vmhbadevs -m
   
   For the datastore that is missing, you see a message similar to:
   
   cannot open volume /vmfs/volumes/4569-08764-eb45-2347a/

3. Compare the list from the output of step 1 with the output of step 2 to identify the device which is missing.
   
   For example, if you see /dev/sdl in the output of step 1, but it does not appear in the output of step 2, /dev/sdl is the missing device.

4. To verify the disk is set to VMFS, run the command:
   
   fdisk -lu device
   
   For example:
   
   fdisk -lu /dev/sdl

5. If this is correct, verify that the magic number is correct by running the command:
   
   hexdump -c /dev/sdl

6. After verifying that it is correct, run a rescan by running the command:
   
   esxcfg-rescan vmhbaX
   
   Where X is the appropriate number in your environment.

7. Check the /var/log/vmkernel log for a Failed to reserve volume error. If the error exists, perform a LUN reset.
   
   For example:
   
   vmkfstools -L lunreset /vmfs/devices/disks/vmhba1\:1\:0\:0

8. Check the /var/log/vmkernel log to verify that the LUN reset succeeded by searching for the device reset message.

9. Mount the datastore by running the command:
   
   vmkfstools -V

Notes:

• If the datastore still does not mount, this indicates that there is a lock on the Storage Processor, as none of the ESX hosts have the datastore presented. Since the lock will not be from another ESX host at this point, try trespassing the LUN from one storage processor to another, and then rescan the HBA on the host. The datastore mounts successfully after the rescan.

• These steps are only valid if one device is missing after an event.