I thought I understood how VMs live in the datastores, but maybe not...

[SteveInTallyFl]

My Guest VM (Prod01) says that it has resources in Datastore01 and Datastore04.

When I look at the Hard Drives assigned to the VM (under Edit Settings) I see that all three Hard Drives are on Datastore04/Prod01/.

When I browse the Datastore01 I do not find anything related to the VM Prod01. Why, then does this datastore show up under the resources for this VM? When I browse the Datastore04 I find under /Prod01/ all the expected files associated with the VM and its defined Hard Drives. Th vmware-##.log files are numbered 52-57.

Then the really unexpected....when I browse Datastore05 I discover under /Prod01/ a partial set of files for this VM but they are dated January 20 of last year. There is no vswp or prod01_3.vmdk. The vmware-##.log files are numbered 19-23.

Can someone explain how/why this VM has a presence on Datastore05? Is it appropriate to delete these files and, if so, is there a proper process for removing them?

[Linjo]

No ISO or Floppy images mounted from that datastore?

// Linjo

[SteveInTallyFl]

Nope, no ISO or other images are mounted from here.

[a_p_]

If the second datastore shows up, there's got to be a reason for this other than just files with the same names in the datastore. As a first step, I'd suggest you take a look at the VM's latest vmware.log file to see whether it contains references to this datastore.

André

[SteveInTallyFl]

Event Viewer on Datastore05 shows all these /prod01/ files being moved from Datastore04 on January 20, 2012.  Event viewer on Datastore04 shows these /prod01/ files being  deleted an hour after the files were moved to Datastore05/prod01/.

This looks like a Storage vMotion event, does it not?

What I do not understand is, the Event Viewer on Datastore04 shows nothing about moving the files back from Datastore05, but the VM is currently on Datastore04 and with current date/timestamps, more current logs, etc.

I cannot find any reference to Datastore05 in vmware-57.log (the hghest numbered log, with the most recent date of 11/19/2012).

[a_p_]

Ok I read your post another time. It looks like there are two issue with this VM. The first one is why Datastore01 shows up although the active files are located on Datastore04 and the second one are the files on Datstore05. Since Datstore05 does not show up in the VM's summary, it's hard to say what happened, so let's concentrate on the first issue. Do you see a reference to this datastore in the VM's current vmware.log file on Datstore04?

André

[SteveInTallyFl]

there are no references to Datastore05/prod01 in the current vmware.log (dated today, first entry is Nov 18, 2012).  All references to datastores are in the following style: /vmfs/volumes/9ce3d5ea-a8ec8b9f/prod01/<file>. I assume there are mapping abstractions. I have scanned these references for anything OTHER THAN "9ce3d5ea-a8ec8b9f". There are none.

[a_p_]

Can you please compress/zip the vmware.log and attach it to a reply post to see whether it contains any other hints which might explain the presence of Datatore01?

André

[SteveInTallyFl]

Attached

[a_p_]

Earlier you mentioned that there's no ISO mounted. However the log file shows

Mar 05 09:15:49.919: vmx| DICT           ide1:0.fileName = /vmfs/volumes/80eb1d95-4024e007/ISOs/ISO/OracleLinux-R5-U7-Server-x86_64-dvd.iso

Please edit the VM's settings and set the CD-ROM to e.g. "Client Device" and disconnect it to see whether this removed Datastore01 from the list of in-use datastores.

André

[SteveInTallyFl]

Changing the User Settings on CD/DVD to Client Device removed the Datastore01 reference on the Guest.  Still do not know the correct action to follow regarding the existance of /Prod01/ on Datastore05. Can it safely be simply deleted through the Datastore Browser interface?

[Linjo]

So there was an ISO on the VM? I though thats what I suggested in the first response... 😉

If the files have not been accessed in a long time I would (after doing a backup of course ) remove them and monitor the vm to make sure that is not affecting it in any way.

// Linjo

[a_p_]

I'd suggest you run RVTools (http://www.robware.net/) to see whether the files on Datastore05 are reported as Zombies in the vHealth tab. If yes, either follow Linjo's advice or at least rename the folder to see whether everything continues to run as expected and delete the files later.

André

[SteveInTallyFl]

Thanks for advice. Downloading  RVTools now....will run.

I am finding another VM with odd folders and files on datastores.

The VM is named "emgr01" and has a disk defined on Datastore01 in a folder named "emgr01_1".

Datastore01 has the "emgr01_1" folder containing the emgr01.vmdk file provisioned for 40GB and in use. This datastore also has a folder named "emgr01" with a 0 byte file named "emgr01_1.vmdk" provisioned for 60GB.

The VM also has a  hard disk defined on Datastore02 in a folder named "emgr01". In this folder is a file named "emgr01_1.vmdk" provisioned at 128GB and in use.

This VM has no snapshots under snapshot manager.

I believe it will be safe to delete the 0 byte vmdk file and its folder. I will use RVTools to verify that this is a zombie situation.