I Need to Use REISUB Too Often

General system Kernel, boot, graphics & hardware
1

I seem to have to need alt-sysreq-reisub more often than I would like to. By more often, I mean every other week … occasionally, weekly. Most of the time when sitting down after the computer has been idle but, yesterday it did happen in the middle of working on stuff in the browser (Chrome).

Are there system logs somewhere that I can look for errors that might tell me what’s going wrong?

I’ve been using Linux since the 90s but I’ve never dove into this area deeply. This is too frequent for “reboot and get back to work”, though. It’s time to learn how issolate and fix.

2

You can view the systemd journal from different boots:

journalctl --list-boots

Then view the specific IDX. The following command should always show the last boot, so that’s what is important here for you:

journalctl -b -1

3

Just FYI: Currently, I have to REISUB quite often, as my system currently is crashing semi-regularly. Always with this message:

amdgpu 0000:03:00.0: [drm] ERROR [CRTC:79:crtc-0] flip_done timed out

I googled around and found that it is an issue with AMD and current state MESA. There are workarounds, but package pinning and downgrading are not my kind of thing for an occasional crash. Not unless the crashes appear more often.

Take a look at your system logs, maybe this is the problem you have too.

4

I’ll roll with that and learn what I can.

Question … Are those logs specific to booting up or will they tell me things about after boot and login?

5

How did you see that message? Is there a specific command you can point me to as a starting point?

I have a Nvidia card but, my mind was already gravitating towards a video card issue.

6

If the computer is idle and goes into sleep mode it could be your nvidia gpu causing the issue.

https://download.nvidia.com/XFree86/Linux-x86_64/510.39.01/README/powermanagement.html

7

I saw it right after the REISUB reboot using journalctl -b -1 (the logs of my last boot) as the last message(s).