ESXi host cannot mount existing datastore although the devices are seen by the host.

Products

• VMware vSphere ESXi

Issue/Introduction

• The host used to have the datastore mounted before it was rebooted.
• The host is well connected to storage and the devices are seen correctly under storage devices.
• Other hosts that weren't rebooted have the datastore well-mounted and operational.
• When querying the partition table of the device, the partition table is not well presented as for example:

[root@localhost:~] partedUtil getptbl /vmfs/devices/disks/naa.############################
gpt
715159 255 63 11489037516

Note: for a healthy datastore we would have an additional line with the information on the VMFS datastore, as the example below:

[root@localhost:~] partedUtil getptbl /vmfs/devices/disks/naa.############################
gpt
715159 255 63 11489037516
1 2048 6442448000 AA31E02A400F11DB9590000C2911D1B8 vmfs 0

Cause

The datastore is failing to mount because the VMFS partition table is missing.

Resolution

The partition table needs to be re-created.

To have assistance in checking if the partition table can be re-created, please open a ticket with Broadcom support.

 

Additional Information

• Please note, if other hosts can access the same datastore that has a missing VMFS partition table, they will continue using the datastore while they don't have to read the partition table  - this happens normally during the reboot process or in the manual mounting process. So refrain from rebooting any additional hosts while this issue is being addressed.
• Please check lun with hexdump command  "hexdump -C /vmfs/devices/disks/naa.############################ " and validate metadata information. sometime LUN is overwritten by secure wipe or other OS install process.