ESXi 4.1 Lost Datastore - restored partition table and datastore is wrong

[pjonesMG]

Hi.  I have 13 ESXi machines that share access to several datastores that are connected via iSCSI.  Just today, they all lost access to one of the datastores rather spectacularly.  After getting into the console and looking at 'fdisk -l <device>', I found that the partition table was missing.  Google turned up this KB article: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=100228... which I followed.  I saw the datastore connect again, but it was blank, and its size and other parameters were messed up.  It should have been 2TB in size, mostly full, with a 2M block size.  It was detected as 2TB, but only 158GB of formatted space, and the block size was 1M.

Has anyone seen this before?  Any ideas?

I have ghettoVCB backups of the VMs that really matter, but the rest were dev machines that will take awhile to get put back together.  I'd really love it if I could just get the partition detected properly again.

Thanks,

-Paul

[pjonesMG]

Replying to myself with more research - if nothing else than to have all of this in one place.

Looking through this post here:

http://virtuallyhyper.com/2012/09/recreating-vmfs-partitions-using-hexdump/

I see that I do have the VMFS partition right where I would expect it to be, 128 sectors in.

That post refers to this post here:

http://blog.laspina.ca/ubiquitous/understanding-vmfs-volumes

Which is a little over my head, but I remember seeing the UUID, and some other parts, but not all of them.  I'm going to go pour through some hex dumps again.  Would still appreciate any help anyone has.

[depping]

Contact Support. That is if you need your data back!

[continuum]

>  Has anyone seen this before?  Any ideas?

Yes - I get one of this cases about once a month. Let me have a guess - do you use OpenE iscsi ?

Follow Duncans advice and call VMware support.
I also highly recommend to create a dd-dump of the whole LUN and store it in a safe location.

I hope you do not use an unsupported iSCSI-system or unsupported Raid-controllers ... that would give support an easy excuse to reject your case.

Good luck - and come back for plan B if necessary