VM Gets "A file was not found" Error

[jfmorales]

We're using the VMWare Infrastructure Client 2.5.0 with VMWare ESX Server 3.5.0 GA release. For one of my virtual machines, when I power it up, I get the error "A file was not found." The VM was originally deployed from a template and has several snapshots. The vmware.log file shows the error

Feb 07 14:01:12.116: vmx| http://msg.disk.fileNotFound VMware ESX Server cannot find the virtual disk "/vmfs/volumes/4734b0eb-55066770-ea4f-001c23c113f3/W2K3 SP2 IE7 IIS SNMP/W2K3 SP2 IE7 IIS SNMP.vmdk". Please verify the path is valid and try again.

Feb 07 14:01:12.116: vmx| http://msg.disk.noBackEnd Cannot open the disk '/vmfs/volumes/4734b1aa-55aaedb0-2d50-001c23c113f3/jfm-mom-2005b/W2K3 SP2 IE7 IIS SNMP-000004.vmdk' or one of the snapshot disks it depends on.

I attach the vmware log files and a GIF showing the list of the files in the folder, which is jfm-mom-2005b. The snapshot file mentioned in the second message is actually present. However, the virtual disk file referred to in the first message seems to have the wrong name and location. The actual virtual disk is jfm-mom-2005b.vmdk in the same folder as the other files. The virtual disk file referred to in the message actually is named after the template that the VM was based on. Also, the message seems to indicate that the file would be at the global level on the volume, rather than in the jfm-mom-2005b subfolder. Unfortunately, the original template no longer exists. But a VM shouldn't depend on the continued existence of the template it was based on, should it??

Thanks for any suggestions on how to fix and/or prevent this situation.

[ponpalani2001]

Hi,

Edit the vmdk file by using nano or vi, and correct the path to the orginal vmdk file. And then try to power on the vm

Thanks

[jfmorales]

Thanks, ponpalani. Unfortunately, I have no Linux skills and no familiarity with nano or vi. So I downloaded the vmdk file for the first snapshot to my Windows system. After the download, I found two files on my Windows system:

W2K3 SP2 IE7 IIS SNMP-000001.vmdk

W2K3 SP2 IE7 IIS SNMP-000001-delta.vmdk

The first (and smaller) file seems to be the descriptor. Anyway, I edited this in Wordpad and uploaded it back to the VMWare ESX Server. The next time I fired up the VM, the original error went away and I got a "parent virtual disk has been modified since the child was created" error instead. So I went and found an article on CID chain repair at . Based on this info, I again modified the W2K3 SP2 IE7 IIS SNMP-000001.vmdk file so the parentCID value matches the actual one in the base disk. Then I fired up the VM again. This time it powered up OK, but when I went to the VM console, it said "Windows could not start because the following file is missing or corrupt: \WINDOWS\SYSTEM32\CONFIG\SYSTEM. You can attempt to repair this file by starting Windows Setup using the original Setup CD-ROM. Select 'r' at the first screen to start repair."

This sounds pretty bad, like maybe the base disk is corrupt now. Is there anything else I can try or should I assume this VM is unrecoverable? Note that I still have a backup copy of the original W2K3 SP2 IE7 IIS SNMP-000001.vmdk file that I've been modifying.

Thanks, Joseph

[ponpalani2001]

Hi,

while exporting a vmdk disk out of vmfs folder it should be properly exported and imported vice versa you should use a VMware convertor to do this. Do you still have your original vmdk files and snapshots associated with that ?

[DCasota]

Hi,

are you sure that the LUN-GUID (4734b0eb-55066770-ea4f-001c23c113f3) is still correct? If there was a resignaturing of the LUNs this might be the problem. You can check the GUID by console: ls /vmfs/volumes

If you don't need your old snapshots, you could try following

- Backup the .vmx-File (Use the Datastore Browser und the Cut/Paste option to backup the file to another directory if you're not familiar with console commands like cp).

- Go to the VM-Settings and remember the SCSI-IDs of the VM hard disks, the SCSI Controller Typ and its SCSI Bus Sharing settings.

- Remove the VM harddisks witht the option "Remove from virtual machine" (NOT the other option which deletes the hard disk files from disk)

- Save the VM-Settings

- Add the hard disk again with the corresponding SCSI-Settings

- Save the VM-Settings again

. Try to boot the VM

If it doesn't work, recover the original .vmx.

Hope it helps.

Daniel

[jfmorales]

>> while exporting a vmdk disk out of vmfs folder it should be properly exported and imported vice versa you should use a VMware convertor to do this.

Thanks, I'll do that next time. As near as I can see in the VMWare Converter, it would only export the VM to a VMWare ESX Server, not to my PC workstation per se. If this is so, I suppose I'd still have to learn Linux and nano or vi to edit the descriptors. Does this sound correct?

>> Do you still have your original vmdk files and snapshots associated with that ?

I still have a backup copy of the vmdk files for the first snapshot. Unfortunately, I don't have a complete backup of all the files associated with the virtual machine. (When I looked at downloading them all to my workstation, it turned out that the base disk image was too large to fit in the available space on my workstation.) And the last time I tried to power up the VM, it appears that the base disk got modified. So I can't really go back to square one anymore.

At this point, I think I'll probably have to create a new VM to replace the old one. Thanks for your suggestions, which I think will be helpful if this ever happens again.

[jfmorales]

>> are you sure that the LUN-GUID (4734b0eb-55066770-ea4f-001c23c113f3) is still correct? <<

No, that GUID had changed, but I already fixed the first snapshot to use the new GUID. I think that's one reason the VM was able to at least power up the last time I tried it. Unfortunately as I said, Windows seems to have a file corruption problem now.

>> If you don't need your old snapshots, you could try following <<

Unfortunately, the snapshots would have to be included for this VM to have any value.

Thanks for your help. Per my discussion with ponpalani, I will make a complete backup of all the VM files if this ever happens again.