Server goes unresponsive

[Dundee]

I installed unraid a few days ago and now it always run fine a few hours then it goes unresponsive. Its not accessible on my network and i cant see it on my unifi server.

Luckily i have a monitor and keyboard/mouse attached to the server so i was able to check what was going on. I was able type "root" but after i pressed "enter" it stuck there (picture 1)

 

I waited 2 mins, nothing. I tried a few things and i think ctrl-alt-del is what finally triggered something and it went into shutdown.

Picture 2.

 

Now ~30 mins later is still stuck at "starting diagnostic collection".

 

Any idea?

 

 

 

 

 

[trurl]

Use the power switch.

 

Get Diagnostics after reboot and attach to your NEXT post. 

 

Setup syslog server 

[Dundee]

Thanks trurl!

 

Yes, i was about to edit my post to add that i had setup syslog on my old synology but you have been faster

 

I attached the diagnostics and also a screenshot of the syslog.

 

unraidnas-diagnostics-20220819-2208.zip

[Dundee]

I found something about ryzen cpus and c-states. I'll try something and update tomorrow.

[ChatNoir]

3 hours ago, Dundee said:

I found something about ryzen cpus and c-states. I'll try something and update tomorrow.

If that does not solve your issue, post a .txt of the syslog after the next freeze/crash.

[Dundee]

It crashed during the night again after being up 2 hours.

Almost nothing in the syslog. The TIME_ERROR came after ~30mins, but the server disappeared from the network 90mins later.

 

I had the 3.0 bios version for my asrock motherboard. I saw that i should update, so i did 3 updates (3.0 --> 3.60 --> 5.30 --> 6.60).

After a few problems, i now have the latest version.

Hopefully it will fix everything

 

All_2022-8-20-6_31_2.csv

[eagle470]

Did this work for you? I'm running a ryzen 9 3900x and I'm having wierd issues.

[JorgeB]

2 minutes ago, eagle470 said:

I'm running a ryzen 9 3900x and I'm having wierd issues.

Start here:

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=819173

 

[Dundee]

Update: issue is still there, even with newest bios and disabled c-states.

 

I noticed today that even if the server was powered on (power button light, fans, etc), periphericals such as my keyboard werent powered anymore. No power to usb ports.

Does this help pinpointing the issue? 

[ChatNoir]

On 8/20/2022 at 3:31 AM, trurl said:

Setup syslog server 

then post this log after the next crash/freeze.

[Dundee]

Here's the syslog.

Almost nothing there.

 

I started the server on the 20th, then it crashed while i was on vacation.

Restarted the server when i came back on the 26th, crashed again during the night.

 

 

All_2022-8-29-12_30_11.csv

[Dundee]

Quick update.

I discovered that since the beginning of this thread, one my ram stick died. In the first diagnosis i posted here, i had 2x16GB. Now only 1 is detected (yes i swapped them to make sure its not a MB slot problem).

 

Now its been 2 days and everything seems fine.

So maybe at the beginning it was a bios / c-state problem, but it might have become a RAM problem.

 

I'll post a final status in a week.