System instability on unstable version - how do I get logs?

trinity

Unstable system is unstable. File to the "no way!" category.

I upgraded to the latest unstable version of NetBSD and my system is locking up / crashing A LOT. Which is pretty much as advertised, so it's hard to complain. Specifically it's locking up whenever I try to play any sort of video - the most I've gotten is 54s through an Invidious video (usually much less) before it stops working and the system becomes unresponsive to keyboard input (mouse still works but isn't very useful).

I was wondering how to get logging enabled / find logs that would help me figure out where the problem is. I'm getting this issue pretty consistently and as far as I can tell it started when I upgraded the kernel so perhaps it's an issue with video drivers? I don't know.

Thanks for any help.

trinity uname -a ?
There were a few broken HEAD images a few weeks ago.

How did you install it and which binaries are you using?

I run HEAD myself and have no issues right now.

trinity

I fetched the CVS sources and compiled the userland and kernel, following NetBSD guide section 35.1.

I've updated my sources and re-compiled generic and custom kernels so I'll update this if it goes away (seems to be good right now...), will also probably re-compile userland at some point.

trinity Good to know it's working now 🙂

trinity

Oops, didn't even answer the uname -a question!

NetBSD nagisa 9.99.87 NetBSD 9.99.87 (X200T) #2: Thu Jul 22 07:24:40 EDT 2021 root@nagisa:/usr/obj/sys/arch/amd64/compile/X200T amd64

X200T is my host (a Thinkpad X200 Tablet) and also the name of the custom kernel configuration.

trinity I suspect something wrong with yesterday's base sets. I've just upgraded to the 22/07 9.99.87 and all is fine.

trinity

Yeah, now it's working perfectly!

I can't rule out that it's a hardware issue (I had lockups that behaved slightly differently on a couple different Linux distros and on GhostBSD) but changing the kernel seemed to have introduced it and now changing the kernel seems to have fixed it? Or perhaps something just started working again with the [unreasonable number] reboots.

Anyway: does anyone know how to store logs that would maybe tell me what's going on if this starts happening again?

nia

To save a crash dump in /var/crash instead of rebooting on kernel panic set ddb.onpanic=1 in sysctl.conf and type "sync" if you get a kernel debugger prompt.

trinity

I've set ddb.onpanic and I found the dmesg logs at /var/log/messages.*.gz. This should give me a good idea of what's going on. Thanks everyone!