Array dropping offline / MCE errors

[Fiffty]

Hi everyone,

 

I am having a few issues with my unraid server since I moved the disks into a new platform. The new system is X370/Ryzen 1st gen based and initially I had a problem with random system freezes, but since setting rcu_nocbs=0-5 (Ryzen 5 1600), as per this post the system does not hang anymore.

 

However, now the array is randomly dropping offline every few hours (without the server hanging) and I am clueless as to why. I had a look in the logs and apparently the CPU is throwing some MCE codes which is of course rather disconcerting, but with it appearing after the above mentioned fix I am wondering if it is connected?

 

Aug 20 06:14:45 VAULT kernel: mce: [Hardware Error]: Machine check events logged

Aug 20 06:14:45 VAULT kernel: mce: [Hardware Error]: CPU 7: Machine Check: 0 Bank 5: bea0000000000108

Aug 20 06:14:45 VAULT kernel: mce: [Hardware Error]: TSC 0 ADDR 1ffff816560e2 MISC d012000200000000 SYND 4d000000 IPID 500b000000000

Aug 20 06:14:45 VAULT kernel: mce: [Hardware Error]: PROCESSOR 2:800f11 TIME 1566274467 SOCKET 0 APIC 3 microcode 8001137

 

System:

Gigabyte Aorus X370 G5

Ryzen 5 1600

16 GiB DDR4-2666

3x Hdd / 2x SATA SSD / 1x nvme SSD

 

unraid 6.7.2 2019-06-25

Plugins: Community Applications, Fix common problems, Dynamix SSD trim, Dynamix File Integrity, PreClear Disks, Nerdtools

Docker: Plex, Transmission

 

Diagnostics: here

 

Edit: I just realized I had the rcu_nocbs=0-5 set for cores instead of threads. Now changed it to rcu_nocbs=0-11, will report back if problem goes away.

Edit 2: Problem persists

Edited August 20, 2019 by Fiffty