Wednesday, 24 June 2009

Vmware issues

Not so much a dedicated security issue (though Availability is in the CIA triangle that should be drummed into everyone by now), but something I thought I would blog about as I found it no where else.
I was having an interesting issue with a guest on one of the ESX clusters I manage, looking at the ESX host, none of its other guests were having issues. The guest in questions came up as disconnected, not powered on. But I could RDP to it.
I logged into the host and checked esxtop and noticed the Guest was in the list.
Checking the tasks of the Guest in the Vmware client I noticed its VCB backup last night died, and that the error I was getting on the guest was Unable to communicate with host, since it is disconnected, I got this same message when trying to power on the guest.I quickly checked the vmware.log and dmesg on the ESX host that was hosting the guest, nothing obvious, googling around gave me no answers. It was then I noticed the last entry in the vmware.log was early this morning, to do with CD rom errors. I thought it could be a simle management disconnect, so I ran

/etc/init.d/mgmt-vmware restart
The whole ESX server disconnected from the VMware client as you expect, then it came backup, the problem host came backup too, no downtime, no mess. Ran a quick manual backup and all done.