After updating to 8.3.1, PVE host shuts down within a minute after boot

[stooovie]

Working for two years. Now, the first boot after installing 8.3.1, the host shuts down after every boot.

- it boots successfully, I can log into the web GUI
- it starts to boot the LXCs and VMs
- ...and then it shuts down completely. Power LED off.

BIOS is set to never sleep, it should not be a power management issue. It happens so quickly, I don't have time to investigate. When I connect HDMI and keyboard to the host, I get to the login, and then it just shuts down.
- tried previous kernel, didn't help
- No error message.

Any tips? Thanks.

 

[leesteken]

Can you manually select an older kernel from the boot menu? Then you could look at the logs from the previous problematic boot/shutdown with journalctl -b -1.

 

[stooovie]

yes, didn't help. I'm trying recovery mode now, it's doing the same but I see the console. will try catching it with a phone

 

[leesteken]

yes, didn't help. I'm trying recovery mode now, it's doing the same but I see the console. will try catching it with a phone

Maybe try booting the system with something else like a Ubuntu installer. That could rule out Proxmox.
Maybe this is the first reboot in a long time? Maybe the BIOS settings are wrong because the battery is empty?

 

[stooovie]

It does NOT shut down when I enter the "maintenance mode" during Recovery. I can issue commands there. Here's what I could catch, just as photos (can not ssh into the machine during this time):

https://imgur.com/a/HRYHXqd

 

[stooovie]

Maybe try booting the system with something else like a Ubuntu installer. That could rule out Proxmox.
Maybe this is the first reboot in a long time? Maybe the BIOS settings are wrong because the battery is empty?

I rebooted a couple of days ago (after like four months), but it worked as expected. I rebooted again today after I saw new kernel in the 8.3.1 update, and the issue started after that.

 

[stooovie]

I see there's a "UPS: forced shutdown in progress". I do have it on an UPS (by APC, connected to a Synology that works as a NUT server). Is it possible the system thinks the power is out (it is not)? Any idea how to reset this?

 

[leesteken]

https://imgur.com/a/HRYHXqd

Looks like the shutdown command runs automatically. Did you add it yourself somewhere?
It looks like you have a UPS, maybe that is sending something because the battery is low and power is missing (or because of a bug)?

I see there's a "UPS: forced shutdown in progress". I do have it on an UPS (by APC, connected to a Synology that works as a NUT server). Is it possible the system thinks the power is out (it is not)? Any idea how to reset this?

Yes it looks a lot like that. Maybe disconnect the USB or serial connection from the UPS to the system?

 

[stooovie]

That was really it! PVE up and running. Whew!

Thanks Lee!

 

[mslatter]

I was scrambling to determine what was shutting down my Proxmox and this was the solution!

Man, I was so proud of myself for setting up NUT on ProxMox so that when Synology sees the electric go down, it gently shuts down ProxMox and tonight that caused some serious stress. Thank you SO much for the tip. I unplugged the USB from my APC to Synology and now it's working again.

I'll need to figure out what is causing the false "on battery" status...but ProxMox is running again. Thanks so much.