Creating server from template fails with disk corrupted or unsupported erro

[RobBuxton]

Hi All,

ESX 3.0.1, VC 2.0.1 plus latest patch.

We had several instances where attempts to deploy a server from a template fail with the error; "Failed to deploy template: The virtual disk is either corrupted or not a supported format"

On one occasion several attempts failed, but then when I selected the check box to temporarily adjust the specification it worked. Deploying to the same disk and not actually changing any options.

The supported format part is definitely wrong as the device is vmfs-3, the corrupted part worries me as this is a SAN based array on our EVA8000. The same array is already home to a number of Virtual Servers.

But it is very odd that it will successfully create the device after a couple of goes, usually with just a twist to the deployment options selected.

So, has anyone seen similar or know of any reasons why this can occur.

TIA

Rob.

[boydd]

I've seen something similar to this in VC 1.x (corrupt template). Try converting template to regular vm and then back to template again.

DB

[chimera]

I have just had the exact same error message - also on an HP SAN (EVA4000) when trying to deploy from a template. Trying a second time it worked (same as you, I slightly changed customisation options selected) The fact it eventually works would tend to rule out corruption on the SAN disks themselves, I'd say it's more likely to be either an initial data transfer or timeout issue from the SAN (about to check for any firmware upgrades for the SAN or HBA's) or a VMWare bug with templates.

[RobBuxton]

I think that's encouraging!

Hopefully it does just mean that it's some kind of timing issue.

Our VMWare hosts are all HP DL385 G1's with QLogic HBA's.

Software on the EVA 8000 is reported as CR0BF7xcsp-5100

Any other similarities in your configuration?

[Oletho]

I just started having the same problem on and off, not during deploy of template but relocation of vm's.

Patched ESX3.01/VC2.01 this weekend.

Relocated a handful with no errors, then one vm gave me this error. Tried several times on different LUN's until it suddenly succeded, and after that it could be relocated to LUNs that previously failed.

Running HA/DRS cluster with HP BL25p and HP EVA4000.

Ole Thomsen

[Oletho]

Are you guys running HP VMM?

It seems that the agent is locking access to files now and then, we had a huge problem with our VCB which seems to be much better (or even cured) after stopping the VMM agents and disabling the VC connection from HP SIM.

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1099600

Ole Thomsen

[RobBuxton]

Ole,

I hadn't thought to relate the two issues. I'd seen the HP Forum one regarding VMotion and errors being reported.

Also this was before I installed HPSIM 5.1 and the latest version of the VMM.

I've not had the issue reported since the upgrade to HPSIM 5.1 and we have created a few new servers from templates.

I might need to go back and recheck when things were changed to see if I can identify a pattern.

[kreisbote]

having the same troubles two. Here we have:

- VC 2.0.1 (32042)

- 4x ESX-Server 3.0.1 (35804)

- 2x HP DL380 G3, 2x HP DL380 G4, all using QLA4052c connected to iSCSI Storage.

It seems to occure mostly on migrating VM from one storage to another (local -> iscsi)

[RobBuxton]

I've not seen the template issue for a while and we have done the following;

Updated the QLogic Firmware to the latest available

Applied the more recent ESX Patches to 3.0.1 - but not the latest January release as none of those appear to apply here.

We have seen some errors on VMotion though, I suspect that may be a separate issue possibly involving HP's VMM which we also use.

I think they may be separate issues.

Interesting that kreisbote's post uses iscsi which may rule out the EVA and imply it's just an HP Server, QLogic HBA, ESX issue.

[bssitton]

I just got the error when clone a regular virtual machine. I am not running HP SIM, but I am running VCB. I think someone was deleting a snap-shot at that exact moment, but I don't know if it was the same VMFS volume.

[Eldron]

I am running 3.0.1 and 2.0.1, and have experienced this as well. DL585's and an EVA5000. I have seen this happen where I can deploy a template, and another admin cannot. Same options, and same target server. It seem to be hit and miss. I cannot reproduce this problem at will, so I am having a hard time troubleshooting it.

I do run HPSIM, so I am going to look in to a possible connection. I will try stopping the VMM service the next time this happens.

[RobBuxton]

What kind of HBA's? At the moment this seems to be restricted to QLogic HBA's. Another poster has indicated it happened on an iSCSI storage array, and so it may be just an HP Proliant / QLogic / ESX 3 / VC2 combination.

The issue with HPSIM / VMM I think is a separate issue.

[Eldron]

I would agree that the SIM issues is separate.

I use Emulex HBA's on 3.0.1 and 2.0.1. I have not had this issue in a while. Is anyone able to reproduce this error?

[jeffwilkinson]

I had been having the tempate deployment issue and random VMotion failures which crashed VMs. I had these problems since January and after 3 separate tickets with HP-VMware support in which they could not figure out the problem, I eventually removed the VMM agent and all the problems disappeared.

[TristanT]

I'm enjoying the random "Virtual disk is either corrupted or not a supported format" errors on a couple of ESX 3 production hosts.

ESX - v3.0.1 build 42829

Server Hardware - HP ProLiant DL385 G1

SCSI AC - Smart Array 6i

SAN - EMC Symmetrix

HBA - Emulex LP9800 FCA

HP mgmt agents - SIM only, not using VMM

I get the error randomly when doing a cold migration or when deploying a new VM from template. We are in the process of moving all of the VMs and templates off of attached SCSI to our SAN volumes. The random errors are very frustrating.

Has anyone in this thread had any updates/revelations since the last post on May 26, '07? I'm opening a case with VMTS later today.

Thanks much!

[ThomasV]

To keep this thread alive to find a solution, at my latest project i'm having this problem aswell.

setup:

HP C-class blades BL 460

Qlogic QMH2462 HBA's

HP EVA 4000

[RobBuxton]

I've not seen this here for a while. But that doesn't mean it's been fixed.

If I see the event again I'll go through the process of opening a call and providing the diagnostics data.

[TristanT]

In addition to the "Virtual disk is either corrupted or not a supported format", I'm now getting random "General fault caused by file" errors when I attempt to deploy from template or clone a VM. About 30% of time I am able to deploy successfully.

In preparing to call VMTS, I updated the firmware and drivers on my Emulex FC HBAs. I'm going to operate for a few more days and see if things improve. I'll post an update soon.

[PerryM]

I'm getting the same problem. I have a host on my SAN (HP EMA12000 with HSG80 controllers), using Emulex LP8000 cards patched with the latest firmware that HP published for these cards.

Yes, my SAN isn't officially supported but it works.

I was trying to do a cold migration (relocation of a volume) of one of the vmdk's from the SAN to local storage, using an HP DL585 w/emulex LP8000 card. The vmdk was 102gig. About 1/3 way through copying the vmdk, I got the same error "disk corrupted or unsupported format.

[TiBoReR]

Are you guys using the same server for VirtualCenter and VCB ?