Invalid configuration for device '0' when creating VM from Template

[EclipseAgent]

Not sure why this started happening, but I am unable to create new VM's from a template.

It gets to about 90 - 92% then errors with the following error:

Invalid configuration for device '0'

I can convert the template to a VM and run the VM, then convert the VM back to template and still get this same error.

[acr]

You could convert your Template back to a VM, clone it using vmkfstools -i (or VC), create another VM from that Clone, then converter new VM to a Template..

Deploy from the new Template see if it works.. You could then delete the old Template..

[EclipseAgent]

Thank you for the response. I changed my VMKernel IP's to private IP's (they weren't for some reason).

That let me get farther, but now I get " Thespecified key, name, or identifier already exists"

I am going to try your suggestion and get a new template.. because I think that one may be bunk.

[acr]

Its more a work arround i know, but at the end of the day, it could be quickest solution..

Good Luck..

[EclipseAgent]

Ok.. back to step one.. I attempted to clone the machine and got "Invalid configuration for device '0'" again..

[EclipseAgent]

Ok problem was with floppy disk pointing to an image in the vm that i converted to a template. checked to client machine and seems to work fine

[acr]

Nice One..!

[skywalkr]

Hi,

We were seeing this same ERROR when creating a new VM and not using a template and yes it refuses to attach a floppy image from a CDSTORE on the SAN. It will work if attaching a floppy locall We isolated the cause to one of the recent patches applied to ESX 3.01. We get this error on systems which have been patched with all the current ESX patches thru 12/31/06 (2158032, 1410076, 1006511, 9986131, 8173580, 6921838, 2066306). We DO NOT get this error on systems at the default ESX 3.01 CD install.

Something in the patches has broken something?

Later, GC Mobley

[waldorfo2]

Has anyone opened support request with VMware for this problem?

I had experienced this problem in 3 different environments while creating VM from template. All servers were running ESX 3.01 and were fully patched at the time (latest patch set from 1/31/2007 was not included)

[rri]

Had the same problem, ESX 3.01 with all fixes applied, deploying from a template worked fine on one machine, didn't work on the others.

Turned out to be quite 'explainable': the template had a ISO-image attached as CD-Rom. This image was on a local drive on the server where the deployment did work. So ... simply disconnecting the iso from the image did the trick.

[jduriez]

i got the same problem with cold migration to relocate VMDK :

ESX 3.0.1 patched i tried between cluster , cluster and stand alone host, same problem "Invalid configuration for device '1'" , when i retry i get another message:"The specified key, name, or identifier already exists"

Message was edited by:

jduriez

[kkelley]

Convert the template to a Virtual Machine. Edit the settings to ensure that no floppy or CD-ROM devices are pointing to datastore images. I have found setting the device to "Client Device" and making sure the "Connected at Power on" check box is not checked has worked to resolve this issue every time. I have not investigated whether 'host' devices cause the same problem.

Device '0' I think is the floppy drive, if one is attached. Otherwise, it would be the CD-ROM.

[freefall]

i got the same problem with cold migration to relocate VMDK :

ESX 3.0.1 patched i tried between cluster , cluster and stand alone host, same problem: "Invalid configuration for device '1'" , when i retry i get another message: "The specified key, name, or identifier already exists"

I tried disconnecting ISO image from the CDROM and got the same problem : "The specified key, name, or identifier already exists"

Has this been raised by anyone as a support request?

[jduriez]

i installed patch 2 for VC and all my ESX are now member of cluster and i haven't the problem anymore ( for the moment )

For each VM to migrate i set the CDROM and the Floppy as "host device"or client device, i re-build new template.

if it can help !!

[bwilliamsdc]

Wow, this was an odd one. That'll teach me to actually apply patches! Like other posters, I don't have this problem in the original 3.0.1 installation -- only after I applied patches for the first time.

<br/><br/>

I fixed this error by disconnecting the ISO image from the VM. I next got the "specified name, key, or identifier" error which I fixed by doing this.

<br/><br/>

Thanks vmware forums.

[Sandakan]

Greetings,

My error was "Invalid configuration for device '1' ", I got it fixed by disconnecting the CD ISO.

Thank you everyone^^

[dgrace]

I had the same error today and found this thread (thanks!).

I did notice that after the failed cloning, the disk was not cleaned up so I had three 51G directories leftover. Had to go clean them up manually. Trying again after restarting the services...

[italiandutch]

I also had the same error "invalid configuration for device '0' ".

I have esx 3.0.2 and latest patches (march 2009).

In my case, I can see why it's complaining and that's because hard disk 1 on bus 0 is pointing to a vmdk disk location which no longer exists.

In fact it's pointing to an old datastore name, which I then deleted and now it shows something like[] templatewin2k3/new virtual machine.vmdk and it should be templatewin2k3/new virtual machine.vmdk

the only problem is that when I edit the settings of the vm, it does not allow me to remove this wrongly confiugured hard disk, not even if I add the good one on good location first.

It's really annoying and there is no way out. Clone will fail with "unable to access file. [] templatewin2k3/new virtual machine.vmdk".

convert to vm and back to template won't help.

it's really frustrating!

Domenico

[italiandutch]

Oh YES YES YES !!!!

after 3 hours working on this, I finally found a way out without having to rec reate the VM.

Basically the Virtual Client GUI just WON'T let you get out of this situation.

You need to go to the ESX server/host console (login via sshwith root) and then go into the subdirectory under the vmfs/volumes dir which contains the vmx file named as the VM which you are trying to repair.

Just open another ssh window and do a cat of a vmx file of another VM which you know is working fine.

Change the path to the scsi0:0.filename parameter, sched.swap.drivedname parameter and change eventually any other parameter which shows the wrong path.

Also I changed the uuid.location to whatever the other good VM was showing.

Whenyou save the file, remove the VM from inventory, browse the datastore, make sure you refresh the whole directory structure with all VM dirs twice, and then re-add the vm to the inventory by right clicking on thevmx file for thatvm with the problem....

Try power back on the VM and it will ask you if you want to create a new UUID, say yes.

et voila'!

It all works for me now.

[Paul_Knight1]

We found that we received this error whenever we tried to edit the settings during the initial deployment from a template. The solution was to deploy the vm from template with the default settings and change them after the vm is created.