The other day I got a call about a VMware Virtual Machine (VM) not turning on after it had crashed. I setup a remote connection to work with the administrator to see what exactly was going on.
The Virtual Machine was not powering on and giving the error that it failed to extend the swap.
What this usually means is that the datastore is too small for the VM.
First thing you want to try in this situation is to edit the virtual machines memory. Lower it then try booting the VM. In this particular case, this failed to work because this did not provide enough free space for the VM to power on.
Looking at VMware’s KB article: KB1004082, it gives steps to change the swap file location, which could fix our issue and give us more space to work with when powering on the VM.
-This VM is in a cluster and would involve reconfiguring the cluster settings for virtual machine swap file location. This just seemed like more work than is necessary and what if moving the swap file does not provide us enough space?
Move the VM to a different LARGER datastore.
-The reason for this will likely allow for more room to power on VM with swap file – also being that the datastore is likely out of space.
After the VM was moved it powered on without issues.
I believe the VMware’s KB article would have fixed the issue temporarily but we would have had still had a space issue. The problem was that the SAN was setup for 200GB Thin Provision and the VM was setup for 300GB. Which would explain why we didn’t see any warning on the Windows Server letting us know that we were running out space. Windows thought it had another 100GB of space!
Needless to say the Exchange Admin was not too happy with the Storage Admin!