VMs and LXCs are frequently 'unknown status'

[ddos127]

Running 9.1.9, and I'm seeing the status of all VMs and LXCs are frequently reporting 'unknown status', I would say it happens 5 to 6 times a day

Temp Fix

systemctl restart pvestatd pveproxy pvedaemon

Has anyone experienced this issue and found a permanent resolution ?

 

[longer]

I have not experienced this issue before. are you still observing this problem?
The journalctl log during the issue time might contain a clear reason for the problem.

 

[kawano-y]

I agree that the journal around the time of the issue would be useful. In particular, I would look at pvestatd first.

For example:

journalctl -u pvestatd --since "YYYY-MM-DD HH:MM" --until "YYYY-MM-DD HH:MM"
pveversion -v
pvesm status
cat /etc/pve/storage.cfg

Since restarting pvestatd temporarily fixes it, I would first check whether pvestatd is crashing or getting blocked while polling storage.

 

[Kingneutron]

Make sure your local time is synced across all instances. You can setup chronyd as the main NTP server on the proxmox host and point all clients to that source.

I have PVE running on a 2018 Intel Mac mini and it goes for a good 20 minutes after reboot with Unknown Status icons, but then it's dead stable.

If you have Ansible running in your environment, I have helpful scripts:

https://github.com/kneutron/ansitest/blob/master/ansible/ansible-query-date--known-hosts.sh

You can do things like check the time on instances with just ssh and keys copied over, but Ansible makes automation easier as long as you can put up with YAML restrictions on whitespace.