Hi all -
I’m experiencing a strange issue on my Threadripper box that started out as a weird power-on issue (that I haven’t seen since).
Having just started the box, I’m SSHed in and it’s idling. After a duration (I haven’t been able to measure this), the box seems to turn off.
I’ve looked at dmesg
, journalctl -b
but I just can’t seem to find any logs of the previous crash/shutdown.
Thoughts? CC @wendell
journalctl --boot=1
gets me the last log, and that’s a good while ago when I did a clean shutdown… hmm!!
Baz
September 29, 2017, 3:00pm
3
If it’s a sudden hard shutdown chances are there’s not gonna be any logs anywhere I think… what if you make a script having dmesg write (overwrite) say the last 20 lines to a file every second? Manual logging you know.
Yeah, I’ll try that. I should also check default Power config in Fedora (UI). Been running headless all this time.
Nothing out of the ordinary there…
I found info on my memory related crash in /var/log/kern.log on Debian and I imagine fedora uses syslogd as well maybe. Worth a look
Honestly - not had much time on the UI side. Been doing stuff mostly over SSH. Waiting on grabbing a 4K IPS display for the Threadripper box…
There’s no /var/log/kern.log
– everything is tracked by journalctl
Eden
September 29, 2017, 4:54pm
9
journalctl --list-boots
to get a specific boot, specify thwe ID of the boot from the list you got
journalctl --boot=-1
to get just error specify the log level
journalctl --boot=-1 -p err
or for example to get log levels 1 - 4
journalctl --boot=-1 -p 1..4
specify -x
to add explinations where it can, specify -r
for reverse order
2 Likes
w00t, figured it out!
Earlier today, I had loaded the it87
module to check my CPU temp – this seems to be rather unstable.
After all, the module is loaded with modprobe it87 ignore_resource_conflict=1
. Seems I’m not alone in this .
Glad to at least figure out the potential cause…
1 Like