Hello all!  Been experiencing stability issues since upgrading to a AMD7900x and supporting MOBO and DDR5.  Thought I had it figured and had an uptime around 20 days when my system started crashing again with kernel panics and kernel BUGs.  The only thing I can think that might be causing this is I'm running a Quadro P2000 and a driver update got pushed a few weeks back.  Anyways I will attach my syslog and hopefully one of you can give me some insight as to what's causing the problem.  Also, in an effort to check if it's the driver i've tried to figure out what my old driver was so I could revert back however I can't find anything that lists old driver versions.  Any help is greatly appreciated, Thanks!

 

Full kernel output is:

kernel BUG at drivers/md/unraid.c:522!

invalid opcode: 0000 [#1] PREEMPT SMP NOPTI

syslog_3_11_2023

So I found a list of the old gpu drivers and reverted back to an older one as a test.  Unfortunately my system crashed within 30 minutes again.  I will be uploading a picture I took of the full kernel trace in an effort to provide more information for anyone reading this thread.

There have been other users reporting stability issues with Ryzen 7xxx, possibly still very new for the kernel, there have been report that disabling C-States, BIOS update or safe mode helps in some cases.

I disabled C-States months ago and have disabled any overclocking done in the BIOS.  I've done a 24 hour memtest and everything passed with 0 errors so I don't think it's a hardware issue.  I've noticed that occasionally I'll get a kernel time sync error or random memory errors but nothing is ever consistent.  I believe it is a kernel problem and the hardware being too new.  Is it possible to compile a newer version of the kernel and apply to unRaid?  or when is the next update of unRaid supposed to be released?

v6.12-rc1 should be out very soon, most likely later this week, it users a much newer kernel, try that one when available.