VM shows “Running” in Front-End but Actullay its in “Poweroff”

Product Support Operations
1

Hello Team,

I AM observing an inconsistency between the OpenNebula Front-End and the nebula logs. Some virtual machines appear in the “Running” state in the OpenNebula Front-End, but when we check on the corresponding logs in nebula, the same VMs are actually in a “Poweroff” state.

Could you please help us understand in which scenarios this “fake running” state occurs?

  • What typically causes this mismatch between the Front-End and the actual state?

  • Is this issue related to communication between the Front-End and the actual state (e.g., due to monitoring or libvirt sync issues)?

  • How can we detect and correct this situation when it happens?

  • How can we prevent this or fix this?

We would appreciate a detailed explanation or any troubleshooting steps to prevent this mismatch in the future.

Thank you for your support.

2

Hello Senthil,

This usually happens when monitoring data between the Front-End and the host is out of sync.
Typical causes include:

  • Network or timeout issues during monitoring.

  • Manual actions on the host (virsh destroy, shutdown, etc.) outside OpenNebula.

  • libvirt or host reboot while OpenNebula still shows the last known state.

Please check:

  • onevm show and onevm .log on the Front-End.
  • /var/log/one/oned.log and /var/log/one/monitor.log for missed or delayed monitor updates.
  • On the host:

    • virsh list --all | grep one–

    • virsh domstate one–

    • journalctl -u libvirtd and /var/log/libvirt/qemu/.log.

To prevent/figure our the root case:

  • Keep NTP in sync on all nodes
  • Avoid manual virsh actions
  • Check monitor.log regularly for timeouts or unreachable hosts
3

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.