Lost SAN storage

[Steve_Forbes1]

I have a 4 way ESX 3.0.2 cluster with IBM DS8100 storage. New LUNs were being presented to the zone used by the servers when we suddenly lost sight of everything.

I can log directly into the ESX servers using VC. However the datastores that were on SAN are missing. In the service console under /vmfs/volumes/ the soft link to the datastore device is flashing (missing link) as the volume is missing.

The disks (LUNs) are still there, the datastores are still under /vmfs/devices/lvm.

How do I get my datastores back?

/var/log/vmware/hostd-8.log;

Result:

(vim.fault.PlatformConfigFault) {

text = "Cannot open volume: /vmfs/volumes/45534e5f-3f14764d-a558-0015605382c6"

msg = ""

}

Steve

[Steve_Forbes1]

Under /vmfs/devices/disks, the vmhbaX:0:0:0 devices for each of my datastores has gone. Can this be recovered somehow? I think this is the root of my problem.

Steve

[EGRAdmin]

Does it give you the ability to add the LUN's as new volumes?

Were they part of an "extent" Volume?

Your partition tables may have been deleted.

execute this command from the service colsole.

fdisk -l | more

does it show your SAN LUN's but does not show them as formatted?

you should see something like;

Device Boot Start End Blocks Id System

/dev/sdb1 * 128 164055779 82027826 fb Unknown

That's the standard partition information for our VMFS volumes.

You can check your's to see what file system it's formatted under.

[Steve_Forbes1]

If I go to create a new Datastore then yes, my LUNs are available although I have not obviously clicked to create the datastore.

The datastore was an extent datastore

Each LUN is much like this;

Disk /dev/sdl: 44.0 GB, 44023414784 bytes

255 heads, 63 sectors/track, 5352 cylinders

Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System

/dev/sdl1 1 5352 42989876 fb Unknown

OR

Disk /dev/sda doesn't contain a valid partition table

So I guess the partition tables may have been deleted. Is there a way to recover them?

Thanks,

Steve

[EGRAdmin]

Broken extents are nasty things.

It can happen for a bunch of different reasons that's why I only use them for secondary backup storage.

This powerpoint I think has a section where it describes the restore process to possibly rebuild an extent.

I've never tried as I had major problems with extents during testing and decided not to use them.

(top support issues part2)

Or this direct link;

[Steve_Forbes1]

Thanks for the help.

Problem was resolved by "re-adding" the first set of disks from the SAN storage gui. Server was rebooted and vmktools run to rescan the volumes. Everything looked ok. The VMS in the VC's inventry had to be re-added to the inventry but that was no great shake.

All working now. I need to build a new datastore though for new VM'. Obviously I'm wary of using extents. Is there a better way? How do I implement it?

Thanks,