Unraid Becomes Unresponsive During GPU Transcoding / Randomly

[LukePOLO]

As the title states, unraid randomly becomes unresponsive either randomly or during transcoding. Although its been happening for a long time, i was able to go for up to 3 months without a crash (limited transcoding) and just dealt with it.

 

Sometimes the system stays up for hours sometimes for minutes. The more intense transcoding I do (either through plex or unmanic) its more reliable to become unresponsive. Both plex and unmanic use `/tmp` to make sure memory doesn't get smashed. 

 

Troubleshooting: 

* memtest (pass)

* switch to using /tmp to make sure it could only consume 1/2 of the total memory
 

Hardware on the Way:

* New Memory 

* New USB Drive

* New Power Supply (350 w -> 500 w) 

 

Attached :

* Diagnostics 

* Img of Last Crash (last-crash.jpg) : some odd shit in there

* Remote SysLog for past 2 days (when i started todo constant transcoding)  : last crash `Aug 9 12:21:2` ish 

* Memory Graph During Crash (Ram Usage.png)

 

Any Help would be Greatly Appreciated! 

 

 

 

 

Update #1:

Replaced PSU , from a 350 -> 500. 

 

 

I noticed that my UPS load was hovering around 360~ , but i have other things connected to it as well (some switches and other things) , so I always thought it was ok and I had some headroom. Will report back after a few hours of testing heavy transcoding. 

 

 

Update #2: 

 

It crashed. Waiting for memory now. 

 

 

 

 

tower-diagnostics-20220809-1411.zip unraid.log

Edited August 10, 2022 by LukePOLO
crashed with new USB and Power Supply