Unable to mount Secondary Image LUN to ESX host

[jprior]

I have two EMC CX3-20's mirroring using Mirrorview Synchronous via Eclipse 1620 SAN routers.

I break the mirror in Navisphere by doing an Administrative Fracture. Then I add the LUN to my ESX storage group. When I rescan the storage adapters, the LUN appears, but when I add the storage to the host, I get the error 'Unable to read partition informatoin from this disk.'

Any ideas on how to access this LUN? The source LUN is a vmfs formatted LUN hosting VMs, mirrored from 1 rack to another.

Any ideas? Am I doing anything obviously wrong?

[cheeko]

Have you seen this thread?

http://www.vmware.com/community/thread.jspa?messageID=518083

[admin]

If you're adding the mirrored LUN to ESX hosts than can see the original LUN it will see the mirrored LUN as a snapshot and refuse to mount it by default. You need to enable resignaturing in order to mount it.

Loads more info in this thread: http://www.vmware.com/community/thread.jspa?messageID=421670

and this one: http://www.vmware.com/community/message.jspa?messageID=364352#364352

[jprior]

Thank you both for the links, I will read those shortly.

No, the hosts are different, in fact connected to different SANs. The source LUN is on a different SAN than the mirror, and the hosts do not share storage (other than the mirrored LUNs).

[jprior]

I ran the commands:

echo 1 >/proc/vmware/config/LVM/EnableResignature & vmkfstools -V

and was still unable to mount the LUN in Virtual Center, same error message. Checking the host vmkernel log I see the following error messages:

Jun 1 03:12:53 fmesx4 vmkernel: 10:19:07:07.879 cpu2:1033)[b]Config: 384: "EnableResignature" = 1, Old Value: 0, (Status: 0x0)[/b]

Jun 1 03:12:53 fmesx4 vmkernel: 10:19:07:07.881 cpu3:1034)[b]SCSI: 8043: vmhba1:0:4:0 status = 2/0 0x5 0x25 0x1[/b]

Jun 1 03:12:53 fmesx4 vmkernel: 10:19:07:07.881 cpu3:1034)[b]SCSI: 8120: vmhba1:0:4:0 is a deactivated snapshot.[/b]

Jun 1 03:12:53 fmesx4 vmkernel: 10:19:07:07.881 cpu3:1034)[b]WARNING: SCSI: 5637: status SCSI LUN is in snapshot state, rstatus 0xc0de00 for vmhba1:0:4. residual R 999, CR 80, ER 3[/b]

Jun 1 03:12:53 fmesx4 vmkernel: 10:19:07:07.881 cpu3:1034)[b]SCSI: 6624: Device vmhba1:0:4 is a deactivated snapshot[/b]

[jprior]

Setting DisallowSnapshotLun to 0 didn't help.

[jprior]

Does it/Should it make a difference how the source LUN was formatted? It is currently VMFS3. Should it be RDM?

[jprior]

The problem was I needed to promote the secondary image to primary once the mirror was broken. After that the datastore appeared without needing the LUN to be added (had EnableResignature=1). Browsing the datastore and adding the VM's to inventory allowed full start up of VM's, without issue.

[kvogel]

>Browsing the datastore and adding the VM's to inventory allowed

full start up of VM's, without issue.

And when the VM had an second vmdk? The path in the vmx is now incorrect, while have a new signature.

is there an better solution? There is an presentation fron VMWORLD 2006, there is the solution:

LVM.EnableSignature 0

LVM.Disallowsnapshotlun 0

The desaster recovery ESX must not have in the Primary Storage Group, only in the Secondary.

Kai