RHEV-M: VM's are "Status Unkown"
I came in on monday to find rhev-m very sluggish to respond and not listing any of the assets attached to it (vms, hosts etc) I ran 'top' on the rhev-m machine and found java eating up a lot of CPU%. I tried restarting 'jbossas' but the sluggishness continued, I rebooted the rhev-m box itself. I had to fiddle about with the hosts after that to get them to 'Activate' again (putting them in Maintenance, rebooting, clicking Activate). At last the hosts are connected again, status Up, one of the two hosts has the SPM Status.
However some virtual machines that were previously up now have the question mark and 'Unknown Status' displayed in the image I have attached. There seems to be no way to reset them, or even remove them. Does anyone have any suggestions at all to either restart these virtual machines or remove them? For now it looks like virtual machines running on RHEV3 can become 'jammed' in Status Unknown if it's necessary to restart services on the Manager or the entire boxes (Manager or Hosts) and impossible to restart or remove.
Responses
Hi,
It sounds like you might have been hit by the leap second issue on your RHEV-M. Please refer to https://access.redhat.com/knowledge/solutions/154793 for details and a solution.
As for the current state, "Unknown" means the VM was running on a host RHEV-M stopped being able to reach, and could not fence, so it does not know whether the VM is up or not, and will not start that VM on another host, for fear of splitbrain.
The easiest solution in this case is to fence the host, and if power management is not available, powercycle the host manually, and in the RHEV admin GUI click "Confirm Host Has Been Rebooted", to let the system know it can now start the VM elsewhere.
Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.