Server shut down – how to figure out why?

I just noticed that my server had shut down 50 minutes ago. When I powered it on again, everything seems to have come back as it should and working as expected now.

My server is built on a A2SDi-4C-HLN4F, 64 GB ECC RAM, a SSD for boot, a NVMe mirror for apps, and a HDD mirror for SMB shares. Running 25.04.2.1 at the moment. Connected to a Eaton 5SC UPS, but should it have been the UPS that caused the shut down I would have expected a notification about that. This was the first time this happened since I installed the server in early June.

I guess there are logfiles that could give a hint on what happened. But being new to both TrueNAS and Linux, I don’t know where to start looking.

So, three questions:

  1. Are there any tests I should run to confirm everything is okay with the server?
  2. Where should I start digging to find the cause of the shut-down?
  3. Is it safe to use the server as usual (backup target for a couple of Macs + some app containers) or should I turn containers and backups off until I know what caused this?

EDIT: The Event Log in the motherboard’s IPMI interface is clean.

EDIT 2: Well, ising journalctl to have a look at the logs revealed the cause…

Sep 09 20:52:10 truenas systemd-logind[2679]: Power key pressed short.
Sep 09 20:52:10 truenas systemd-logind[2679]: Powering off...
Sep 09 20:52:10 truenas systemd-logind[2679]: System is powering down.

I was putting some boxes on the same shelf where I have my server around (or more probably exactly) the time when it shut down. So, user error of the more unexpected kind.

Pressed the power button to confirm, and the server shut down nice and controlled once again.

But this leads to a new question: Is there away to prevent TrueNAS to shut down on a short press on the power button?


I also noticed this in the logs – not related to the shutdown, but is that something to look into?

Sep 09 20:49:09 truenas systemd-coredump[1685538]: [🡕] Process 1667726 (python3) of user 0 dumped core.
                                                   
                                                   Module [dso] without build-id.
                                                   Stack trace of thread 198:
                                                   #0  0x00007fb6c6b24921 n/a (n/a + 0x0)
                                                   #1  0x00007fb6c448b060 n/a (n/a + 0x0)
                                                   #2  0x00007fb6c4446406 n/a (n/a + 0x0)
                                                   ELF object binary architecture: AMD x86-64