(6.9.2) Ryzen server keeps becoming unreacheble, require hard reboot

[ChloeAAApossum]

Hi, I have a Ryzen 7 2700X on a ASRock B450M Pro4 (bios version P3.50). I have disabled c-state in the bios and added rcu_nocbs=0-15 to Unraid OS in the flash key syslinux configuration.

The server keep becoming unreachable (no access to the webUI, it is pingable but rebooting via SSH does not work and thus require an hard reboot). Anyone have a similar problem and know how to fix it?

 

The diagnosis zip is after doing the hard reboot, array isn't started.

melody-diagnostics-20210601-2233.zip

[John_M]

2 hours ago, ChloeAAApossum said:

I have disabled c-state in the bios

 

Don't do that.

 

2 hours ago, ChloeAAApossum said:

and added rcu_nocbs=0-15 to Unraid OS in the flash key syslinux configuration.

 

Don't do that either.

 

The BIOS setting you're looking for is Power Supply Idle Control. You should find it buried in the extensive AMD CBS submenu. The default setting is either "Low Current Idle" or "auto". You can change it to "Typical Current Idle" and it will prevent the CPU from entering the very lowest power mode from which early first generation Ryzens sometime fail to wake up.

 

However, I don't think this is your problem, for two reasons. Firstly, you don't have a first generation CPU. The BIOS change won't do any harm though. Secondly, your server responds to pings, which I don't think it would do if the CPU was frozen.

 

Your diagnostics don't reveal very much because they cover less than a minute of up-time. But I notice you have a very strange and invalid memory configuration: 3 x 8 GB plus 1 x 16 GB, giving a total of 40 GB. Try running with just a pair of 8 GB DIMMs for a while.