Bug #1943

VM state flips between RUNNING and UNKNOWN

Added by Marc Warne about 8 years ago. Updated about 8 years ago.

Status:ClosedStart date:04/23/2013
Priority:NormalDue date:
Assignee:-% Done:

0%

Category:Drivers - VM
Target version:-
Resolution:worksforme Pull request:
Affected Versions:OpenNebula 4.0

Description

I'm using the 4.0 RC and my VM (using the CXM TM) gets deployed just fine.

However, the LCM changes its state to 'UNKNOWN' all the time so it flips between RUNNING and UNKNOWN. From what I can see, the monitoring runs just fine (oned.log shows polls with STATE=a) so I'm not sure why the LCM decides it should be UNKNOWN.

Attached are the logs for the VM and oned.log.

13.log (37.2 KB) Marc Warne, 04/23/2013 09:11 AM

oned.log (160 KB) Marc Warne, 04/23/2013 09:11 AM

History

#1 Updated by Ruben S. Montero about 8 years ago

There are some messages in your log like:

STATE=- USEDMEMORY=131072 USEDCPU=0.0 NETRX=1636 NETTX=0 NAME=oneadmin-Debian13.one

Then there are others with STATE=a. I'm a bit concerned about seeing Monitoring actions for the VM. 4.0 minimizes the number of polls by including VM monitoring info in the Host poll. When a VM that is supposed to running in the host is not found is moved to UNKNOWN. There should be no need to issue specific Monitoring actions for each VM.

There is an issue (#1949) solved related to monitoring when the monitor time is less than the time to execute the probes (~10 sec.) in your case.

So, as far as UNKNOWN to RUNNING flips seems to be related to STATE=- and STATE=a in your logs.

#2 Updated by Ruben S. Montero about 8 years ago

  • Category set to Drivers - VM
  • Status changed from New to Closed
  • Resolution set to worksforme

Closing this one as worksforme, we can reopen it if there is more information about this error.

Also available in: Atom PDF