So I don't know why, but my server is locking up when running tdarr.

It can take 30 minutes, it can take 90 minutes, but whenever it is it does seem to eventually happen.
My RAM recently failed and I replaced it but that didn't fix it.  I also upgraded from Unraid 6.9.2 to 6.10.3 hoping maybe it was some kind of issue with the Nvidia driver (Which was behind I think due to the old version).  None of that has helped.
The next thing I'm trying is removing the GPU Statistics plugin by b3rs3rk because I saw something about it maybe causing issues?   I'll report back if that works.

If someone who knows how to read the system diagnostics better than I can figure it out I would greatly appreciate it!

tower-diagnostics-20220815-2344.zip

Edited August 18, 20223 yr by Daxter

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

 

13 minutes ago, JorgeB said:

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

 

I'll look into this.
What did you see that led you to link me to this?
Is it that my RAM is overclocked to 3600MHz?  My old kit was 3200MHz so I could see that being the issue.

1 hour ago, Daxter said:

Is it that my RAM is overclocked to 3600MHz?

That and also check that the power supply idle control is correctly set

12 minutes ago, JorgeB said:

That and also check that the power supply idle control is correctly set

The PSU idle control was set to auto, I changed it to "typical current idle" and will test it when I wake up tomorrow.

If that fails it looks like I should set the RAM to 2667MHz or maybe 3200MHz since that worked before?  I checked and I have dual rank RAM and a 2nd gen Ryzen CPU.

I would recommend setting the RAM do the max officially supported by that config.

13 minutes ago, Daxter said:

f that fails it looks like I should set the RAM to 2667MHz

I would start with this initially as the recommended value for your CPU as any instability is bad in a server.   If that proves stable you can always later try a slight overlock to see if the system remains stable.

consider adding some Extra Parameters to avoid completely maxing out the docker use of your resources. For my tdarr I have the following in Extra Parameters in the docker:

 

--cpu-shares=256 --cpus=8

 

which puts the docker at lower priority than pretty much everything else (cpu-shares numbers range from 0 to 999 I think?? Default is 999 or 1000 I don't remember. Point is this number is lower) and cpus sets the max number of cpus that can be used at any point in time. Personally I prefer this over pinning the cpus as unraid just uses whatever it wants up to a maximum equivalent of 8 total CPUs. In my system that's 25% of available so that the system always has resources to do its thing.

 

The other thing I found when I hit a certain threshold of number of dockers, read write into them, etc. is that having appdata and the docker image on the same SSD actually caused me a world of issues. Splitting appdata and system (docker) onto separate physical disks using separate cache pools completely eliminated my system going unresponsive.

On 8/18/2022 at 6:28 AM, Daxter said:

The PSU idle control was set to auto, I changed it to "typical current idle" and will test it when I wake up tomorrow.

If that fails it looks like I should set the RAM to 2667MHz or maybe 3200MHz since that worked before?  I checked and I have dual rank RAM and a 2nd gen Ryzen CPU.

So far so good, ~20 hours later and no issues after setting the PSU idle control to "typical current idle".

 

Edit: ~43 hours now and everything is running perfectly fine, through Tdarr transcodes and thorough health checks on the CPU and GPU.

Edited August 22, 20223 yr by Daxter
Update