xe vm-list shows halted - but vm is running and working - XS71ECU2

[Goumltz Reinicke]

Hi,

 

after a raid crash, which killed a NFS SR we still clean up the dead parts which are left over.

 

During that clean up I noticed a VM, which is still up and running (I'm ssh-ing to it, it is running the installed services, ...) but in XenCenter ist is shown red white square (powered off) and  xe vm-list shows halted.

 

Now I m confused. Any suggestions to debug or "fix" that? The VM should be up and running and the tools should show the correct status :)

 

Regards . Götz

[Alan Lantz]

Oh, thats weird. I would blame the poolmaster. Maybe do a xe-toolstack-restart on it and see if it picks up the correct 

state of the VM.

 

--Alan--

 

[Goumltz Reinicke]

HI Alan,

 

I did a xe-toolstack-restart, but no effect. VM is marked halted, but still up and running 

 

More suggestions?

 

What could happen in the worst case if I just restart / halt the VM from within the vm .... ? :) getting curious.

[Tobias Kreidl]

Try resetting the VM's power state with "xe vm-reset-powerstate uuid=<UUID-of-VM> force=true".

 

-=Tobias

[Goumltz Reinicke]

Hi Tobias,

 

no luck. Still get power-state halted. 

 

/Götz

[Alan Lantz]

Any chance the VM you see running really is running somewhere else? I've just never seen a VM 

report incorrectly like that. What happens if you try to start that VM marked down ? 

 

--Alan--

 

[Tobias Kreidl]

^^^ Me, either. Very odd.

[Goumltz Reinicke]

14 hours ago, Alan Lantz said:

Any chance the VM you see running really is running somewhere else? I've just never seen a VM 

report incorrectly like that. What happens if you try to start that VM marked down ? 

 

Hm ... as the VM is quit critical testing a reboot or "second start" makes me feel a bit uncomfortable.

 

Any suggestions for a procedure (doing a data backup is done)

 

/Götz

[Goumltz Reinicke]

I tried to start it and I do get the message: The VDI is not available. 

[Tobias Kreidl]

Seems like something is corrupted. I'd look in the various log files for more clues.