Nvidia-smi: Unable to determine the device handle for GPU: Unknown Error (Docker container)

Hello all. I am having issues with running crypto mining software inside a Docker container. I am successfully able to run the software outside of the container for days on end, but within the container my hashrate drops to 0 on one GPU (so far it has been either the first or last cards on the bus) and nvidia-smi produces the error in the title. nbminer gives: “GPU X hung detected!”. The clues seem to point to a hardware issue, but considering that it is a variable GPU that hangs as well as the fact that I can mine without issues outside of the container, this seems to suggest other issues. I wanted to get a more experienced set of eyes on this, as I have had a lot of trouble interpreting the bug report and I’m unsure how to continue troubleshooting.

Thanks in advance!
nvidia-bug-report.log.gz (2.7 MB)

You’re getting a XID 79, fallen off the bus. Most common reasons are overheating or lack of power. Monitor temperatures, reseat power connectors/the card in its slot, check/replace PSU.