I ran into an interesting situation on my Debian trixie servers the other day that others might want to be aware of.
I have a nut server running attached to a battery backup, and I had triggered a battery calibration.
Then a few days later I rebooted the nut-server following some software updates, and my debian trixie servers which monitored the server all shut themselves down.
That was not at all what I expected would happen, so I did some digging and found it might be related to this bug:
Debian Bug report logs - #1056190 /sbin/upsmon: CAL flag in UPSMON never cleared, shutdown procedure will not be triggered
That got me thinking - what else am I running that is Debian based? Am I crazy that Proxmox VE might be impacted?I have not wanted to test because I don't want to take down a cluster node. That said, I did find a github issue where another person reports running into a similar issue in PVE.
https://github.com/networkupstools/nut/issues/2794
I do notice that the bug tracker says this is fixed in debian nut/2.8.3-1 but the PVE server lists the latest version as nut/2.8.1 so I guess it would make sense that this might still impact PVE.
For now I'll probably work around the issue by manually installing a later version.
Edit: It appears like the fix has been pushed for Debian trixie, but it's just waiting on acceptance from the release team. I missed that in the issue chat.
I have a nut server running attached to a battery backup, and I had triggered a battery calibration.
Then a few days later I rebooted the nut-server following some software updates, and my debian trixie servers which monitored the server all shut themselves down.
That was not at all what I expected would happen, so I did some digging and found it might be related to this bug:
Debian Bug report logs - #1056190 /sbin/upsmon: CAL flag in UPSMON never cleared, shutdown procedure will not be triggered
That got me thinking - what else am I running that is Debian based? Am I crazy that Proxmox VE might be impacted?I have not wanted to test because I don't want to take down a cluster node. That said, I did find a github issue where another person reports running into a similar issue in PVE.
https://github.com/networkupstools/nut/issues/2794
I do notice that the bug tracker says this is fixed in debian nut/2.8.3-1 but the PVE server lists the latest version as nut/2.8.1 so I guess it would make sense that this might still impact PVE.
For now I'll probably work around the issue by manually installing a later version.
Edit: It appears like the fix has been pushed for Debian trixie, but it's just waiting on acceptance from the release team. I missed that in the issue chat.
Last edited: