IldarAlmaty
Contributor
Contributor

vmfs hot copy: any chance to recover?

Jump to solution

Hi!

I got into a trouble.

Being very limited in time I did on the running VMFS:

dd bs=100M if=/dev/sdc1 of=/dev/sda8

Both /dev/sdc1 and /dev/sda8 are of type fb.

After that I had sdc removed.

On next rebootsESX does not see /dev/sda8 as a datastore. I spent a lot of time playing with ESX already but couldn't make it work.

1. Was it really bad mistake to hot-copy a running VMFS?

2. Are there any ways to fsck the VMFS?

3. sda8 is a little bigger then the origin, is it wrong?

Thanks for any advice!

Tags (3)
0 Kudos
1 Solution

Accepted Solutions
BUGCHK
Commander
Commander

My guess is that sda8 is presented on a different LUN address. In that case you have to resignature the VMFS to get access again.

VMKB9453805 - Resignaturing VMFS3 Volumes That Are Not Snapshots

http://kb.vmware.com/kb/9453805

(whether the VMFS and/or its embedded files do not have some kind of corruption is another question)

View solution in original post

0 Kudos
4 Replies
IldarAlmaty
Contributor
Contributor

Just in case:

1. The partition is properly alligned:

hexdump /dev/sda8

0000000 0000 0000 0000 0000 0000 0000 0000 0000

*

0100000 d00d c001

I even resized it well, so that no

WARNING: LVM: 1625: Device size mismatch

(actual 445562648 blocks, stored 445562647 blocks)

appears now.

2. vmkfstools is not very helpful Smiley Sad

vmkfstools -R vmhba3\:0\:0\:8

...

Error: vmkfstools failed: vmkernel is not loaded or call not implemented.

vmkfstools -s vmhba3

vmkfstools -l vmhba3\:0\:0\:8

Invalid ROR type: 'vmhba3:0:0:8'

3. Now I see just a few suspicios log messages:

Dec 23 23:28:35 esx1 vmkernel: 0:00:15:36.510 cpu5:1040)WARNING: ScsiUid: 550: Path 'vmhba3:C0:T0:L0' : supports ANSI version 'SCSI-2' (0x2). In order to be used with ESX a device must support the SCSI 3 protocol.

Dec 23 23:28:35 esx1 vmkernel: 0:00:15:36.510 cpu5:1040)ScsiScan: 516: Path 'vmhba3:C0:T0:L0': No standard UID: Failure

Dec 23 23:28:35 esx1 vmkernel: 0:00:15:36.510 cpu5:1040)SCSI: 861: GetInfo for adapter vmhba3, , max_vports=0, vports_inuse=0, linktype=0, state=0,

failreason=0, rv=-1, sts=bad001f

Dec 23 23:15:30 esx1 vmkernel: 0:00:02:31.216 cpu5:1040)BC: 814: FileIO failed with 0x0xbad0006(Limit exceeded)

and on starting up:

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.096 cpu4:1040)ScsiUid: 754: Path 'vmhba3:C0:T0:L0' does not support VPD Serial Id page.

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.096 cpu4:1040)WARNING: ScsiUid: 550: Path 'vmhba3:C0:T0:L0' : supports ANSI version 'SCSI-2' (0x2). In order to be us

ed with ESX a device must support the SCSI 3 protocol.

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.096 cpu4:1040)ScsiScan: 559: Path 'vmhba3:C0:T0:L0': No standard UID: Failure

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.096 cpu4:1040)ScsiScan: 641: Discovered path vmhba3:C0:T0:L0

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.143 cpu4:1040)SCSI: 1917: Device vmhba3:0:0 has not been identified as being attached to an active/passive SAN. It i

s either attached to an active/active SAN or is a local device.

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.143 cpu4:1040)ScsiPath: 3178: Plugin 'legacyMP' claimed path 'vmhba3:C0:T0:L0'

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.181 cpu4:1040)ScsiDevice: 3479: Adding Device "vml.0200000000600508e0000000000d6d2df83d04dc064c6f67696361" from Plugi

n "legacyMP", Device Type 0

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.181 cpu4:1040)SCSI: 5495: Logical device vml.0200000000600508e0000000000d6d2df83d04dc064c6f67696361 for target vmhba3

:0:0 was registered successfully.

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.181 cpu4:1040)SCSI: 861: GetInfo for adapter vmhba3, , max_vports=0, vports_inuse=0, linktype=0, state=0,

failreason=0, rv=-1, sts=bad001f

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.183 cpu4:1040)Mod: 1436: Initialization for mptscsi_2xx succeeded with module ID 4.

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.183 cpu4:1040)mptscsi_2xx loaded successfully.

Dec 23 23:15:15 esx1 vmkernel:

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.187 cpu4:1040)Mod: 1445: modLoaderHeap avail after: 7808816

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.187 cpu4:1040)Mod: 1473: no private ID set

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.187 cpu0:1024)VMNIX: VmkDev: 1715: mptscsi_2xx

Dec 23 23:15:15 esx1 vmkernel: 0:00:00:15.187 cpu0:1024)VMNIX: VmkDev: 1741: Registered 1 adapters

Any ideas?

0 Kudos
BUGCHK
Commander
Commander

My guess is that sda8 is presented on a different LUN address. In that case you have to resignature the VMFS to get access again.

VMKB9453805 - Resignaturing VMFS3 Volumes That Are Not Snapshots

http://kb.vmware.com/kb/9453805

(whether the VMFS and/or its embedded files do not have some kind of corruption is another question)

0 Kudos
IldarAlmaty
Contributor
Contributor

That's right. Kicking resignature helped. Though:

1. extensive googling showed me that mainly people see snapshots instead of their datastores. In my case ESX didn't see snapshots.

2. Your link as well as many others give VI moves as solutions. I don't have it right now. (VC is in the datastore itself Smiley Happy)

That's what I did:

echo 1 > /proc/vmware/config/LVM/EnableResignature

And now I seem to have snapshots to resolve Smiley Happy

Thank you anyway. Your hint was in the right direction.

0 Kudos
IldarAlmaty
Contributor
Contributor

OK.

Issue resolved. The right procedure can be foud here:

http://communities.vmware.com/thread/149634#961480

Respect to

0 Kudos