ESXi Provisioned Storage Issue

[tomtas]

Hi All,

I've recently migrated a number of physical servers to virtual machines on ESXi 4.0 but have encountered an issue regarding a large increase in the provisioned storage of certain VMs, I suspect there's a simple solution/configuration I've overlooked.

We have a file server VM (Windows Server 2008 R1 64-bit Datacenter Edition), I provisioned 350GB for the hard/virtual disk and all seemed to be fine for a time. Now when I look at the resource details for the VM it has the following;

Provisioned Storage: 955.84GB (only 1.83GB free in datastore)

Not-shared Storage: 583.19GB

Used Storage: 583.19GB

Additionally there are 3 vmdk files within the VM folder in the datastore; server.vmdk (343GB) server-000001.vmdk (267GB) and server-000002.vmdk (182GB). I've tried moving the additional files (aside from the original .vmdk) but then the VM refuses to start, initially I thought to remove any snapshots via the snapshot manager but there was only 1 snapshot, and that failed to be removed due to insufficient space in the datastore and now the snapshot manager shows no snapshots.

I plan to rebuild the file server after office hours, as I see no alternative and need to implement a solution asap. Is this really an issue to do with snapshots or an more likely an incorrect configuraiton during the initial setup of the VM (thin privisioning, etc)?

Any advice would be greatly appreciated.

-Tom

[danm66]

Tom, don't move those files! You have snapshots on your base disks and those 00000X files are the changes that have been made since the snapshot(s) were taken. You will need to power down the VM and the sooner the better because you are close to crashing the system with a full datastore. Open up the snapshot manager and select delete all. That will roll all of the changes up into the base vmdk. If you go back to the base disk, you ruin the integrity of the snapshot change and are going to lose all of the modifications and changes made since the first snapshot was taken.

The other option is shutdown the VM and clone the drive with a command that will look similar to : vmkfstools -i server-000002.vmdk /vmfs/voluments/otherdatastore/folder/newvmname.vmdk .

[tomtas]

Thanks danm66,

I've backed up all the business data off the VM, powered down the VM, removed from the inventory and readded so that the snapshots were once again visible. As suggested I've run the delete all option and it's now sitting at 99% and has been stuck at 99% for the past 45 minutes, I can't tell if it's doing anything (reaching 1am here) which is a shame as I can't leave it like this for much longer (I need sleep and the server needs to be up before 7am).

[danm66]

The operation may have timed out in the vcenter server, but it will keep working in the background.