Ever since updating from 6.12.10 to 6.12.14 I've had frequent server crashing. I finally figured out how to capture the syslog and the most relevant lines seem to be 

Dec 18 23:06:32 StrongBad kernel: amdgpu 0000:0c:00.0: amdgpu: SMU: I'm not done with your previous command: SMN_C2PMSG_66:0x0000001C SMN_C2PMSG_82:0x00000007
Dec 18 23:06:32 StrongBad kernel: amdgpu 0000:0c:00.0: amdgpu: Failed to export SMU metrics table!

 

Basics of the system:

• AMD Ryzen 7 5700G
• ASRock X570 Taichi

 

Everything was working fine prior to the update. I've made no other changes since then. Including my anonymized diagnostics, let me know if you need me to add anything else. Thanks in advance!

strongbad-diagnostics-20241219-0011.zip

Edited December 19, 20241 yr by barakthecat
Removing random link

Try blacklisting the GPU driver:

 

https://docs.unraid.net/unraid-os/release-notes/6.10.0#linux-kernel

Thanks, I'm trying it now. I had another crash a few hours ago, but nothing was captured in the log this time - last entry was at 8:38 am, but the crash happened at 9:19 am.

That didn't work, I've had several crashes since then. Here's an updated diagnostics.

strongbad-diagnostics-20241220-1709.zip

1 hour ago, barakthecat said:

That didn't work, I've had several crashes since then. Here's an updated diagnostics.

strongbad-diagnostics-20241220-1709.zip 161.43 kB · 0 downloads

I would look to remove the vfio mappings for the internal usb controller to see if that fixes stability.

Just wanted to give an update here in case anyone has similar issues. I've been working with George R on the Unraid Support forum.

 

• I tried blacklisting the GPU driver with no improvement
• I followed JorgeB's recommendations on Ryzen servers here - https://forums.unraid.net/topic/46802-faq-for-unraid-v6/page/2/#findComment-819173
  • I already had C-States disabled
  • I updated my BIOS from 4.6 (dated 2021/8/5) to 5.6 (dated 2024/1/22), which is the latest released BIOS at this time. The crashes became even more frequent after this.
• I cleared the VFIO mappings per SimonF's recommendations (had no choice as the BIOS update changed my IOMMU groupings), with no improvement.
• George R noted I am running a mismatched RAM setup and recommended I try removing some RAM to see if stability improves.
• Prior to trying this I updated the BIOS to beta 5.63 (dated 2024/8/26) and at the same time turned on XMP RAM profiles to run the RAM at its full rated speed (I had not been doing this previously). Since then the server has been up for over 3 days without issue.
  • I don't know if the fix was in the 5.63 beta BIOS or in turning on the XMP RAM profile

This weekend I'll try restarting the server and turning off the XMP profile and see if it becomes crashy again, just so I know which of the two changes made a difference.