(solved) server keeps rebooting randomly once a day

[zeighy]

Seems my UnRaid is acting weird. Once a day, it would randomly restart and kick in a "unclean shutdown" parity check.

 

I can't figure out what's wrong. Tried memtest 24 hr, all clean. Nothing weird with the power supply either. All hardware seems fine. So, I started to enable mirror syslog to flash... And it stopped randomly restarting. Waited it out 1 weeks, and after no more incident, I turned off mirror syslog to flash assuming it somehow got resolved. And just today, it restarted on its own again. So, I enabled mirror syslog to flash again just so I have a record again.

 

Any idea? Is it potentially the flash drive is dying and I need to replace it?

 

I've also recently replace one drive after UnRaid kept complaining it had problems. SMART reports all good, and parity checks reported no problem either. I still replaced it, just to be safe. Checked the drive on a different machine, and seemed fine. No sure if relevant, but wanted to mention just in case. Only after I replaced the drive when the random restarts started happening, not before.

Edited June 5, 2021 by zeighy
mark as solved

[JonathanM]

Have you done the ryzen tweaks?

[zeighy]

3 hours ago, jonathanm said:

Have you done the ryzen tweaks?

Ryzen tweaks? I do have a Ryzen 1700X on a X370 board. It's been running on this system for a couple of months no issues, so I didn't think I need to do any tweaks.

 

Looking it up seems it's about the C states. I'll do some more digging and will report back. Thanks for the suggestion

[John_M]

It's the Power Supply Idle Control setting. You'll find it under the extensive AMD CBS section of the BIOS. The default setting is Low Current Idle (or maybe Auto). Change it to Typical Currrent Idle and leave the C states options at their default values.

 

However, this is only effective against freezes when idle, not re-boots.

 

Edited May 27, 2021 by John_M

[zeighy]

On 5/26/2021 at 9:27 PM, John_M said:

It's the Power Supply Idle Control setting. You'll find it under the extensive AMD CBS section of the BIOS. The default setting is Low Current Idle (or maybe Auto). Change it to Typical Currrent Idle and leave the C states options at their default values.

 

However, this is only effective against freezes when idle, not re-boots.

 

I'd like to report that it's been 4 days since I implemented this change, as opposed to doing any ryzen tweak for c-states and I have not had any random reboot even when syslog is not mirrored to flash. Seems stable.

 

I'll give it a few more days before I would close this and call it solved, but I just wanted to report back.

 

( I haven't tested the c-state change, yet.... But the typical current idle change seems the less drastic change of the two, so it was best tried first.)

[zeighy]

Just a quick update that the change to "typical current idle" for power supply control seems to have worked and I have not had any random shutdown/reboot. No other changes to configuration or bios settings were done.

 

Still untested is the c-states, but seems unnecessary.

[JorgeB]

8 hours ago, zeighy said:

but seems unnecessary.

It should be, another common issue with Ryzen is the RAM speed, good to check you're withing max official specs.

 

 

[hybrdthry911]

On 5/26/2021 at 9:27 PM, John_M said:

It's the Power Supply Idle Control setting. You'll find it under the extensive AMD CBS section of the BIOS. The default setting is Low Current Idle (or maybe Auto). Change it to Typical Currrent Idle and leave the C states options at their default values.

 

However, this is only effective against freezes when idle, not re-boots.

 

So far so good. This appears to be a solid fix for my recent reboots.