Guest machine disk locates on two volumes. Can not correctly migrate it.

[Puzen]

Guys, good afternoon. Question to the pros. In General, a couple of weeks ago, one of the hosts (5..1) in the cluster died. And one of the guest workstation didn't move.(BSS_IIS)

I created a new guest machine (w_BS_IIS), hooked up a disk to it , everything went up ,everything works.

Yesterday, I noticed, that the disk in the guest machine now rests on two volume.(VOL2 / VOL3) And I can't correctly migrate machine disk to one of them.

I didn't find anything in the logs either. I understand that the problem is in the name of the disks themselves.

I don't know how it turned out. Restore from a backup? Or how can it be solved correctly?

Thank you!

[scott28tt]

First point, any versions older than 6.5 are no longer supported or updated - so if you are running anything like a production workload you should consider an upgrade to at least 6.5.

To your problem, it looks like your original VM was on VOL2, you created a new VM on VOL3 but added the virtual disk from the original VM to it - right?

What do you mean by "can't correctly migrate machine disk to one of them"? What action were you trying to carry out and what errors or messages did you see when you tried?

[Puzen]

Thank you for this information and your time!

1.How I made it. I created new Guest VM and use existing virtual disk (on VOL2).

2.What I mean. I think that it is not all right, that Guest VM disk locates on 2 volumes. And I want migrate it on VOL2 (use vmotion).

But I see error.

[scott28tt]

Do you have another datastore you could migrate to?

As well as the log files you also have some other files (eg. vswp) leftover from your original VM in the VOL2 folder.

[continuum · Accepted solution]

First things first: look at the same problem with WinSCP.

You need to check timestamps of the flat.vmdk and timestamps of the descriptor.vmdk

Just inspecting it with the datastorebrowser does not do it.

Once you figured out if the two vmdks are valid different disks there are 2 options:

a: both disks are identical and a strange operation just dublicated them

b: disks are different and a strange operation messed up the names

If A - just delete one of them while the VM is powered off

If B - power off the VM and unregister it for a minute

use WinSCP embedded editor and rename one of them (means rename the descriptor.vmdk, rename the flat.vmdk and fix the reference to the flat.vmdk in the descriptor vmdk.

When done, edit the scsi0:*.fileName parameter in the vmx-file so that it fits the renamed file.

Save vmx-file and register the VM again.

Should do the trick and should not take more than 3 minutes of a downtime.

And important: you must realize that Datastorebrowser is not good enough to be used for troubleshooting.

The earlier you get that lesson the better !!!

[Puzen]

Hello!

Thank you all for your help and time. I think than your suggestions were right. And now I know about this ways.

My problem guest machines were static. So, I cloned it and deleted old. Today all work!

Thank you!