Unable to transcode, orchestrator exits

[pabloromeo]

Unable to transcode, orchestrator exits

[sjdaws]

So spot on about the strategy being the problem, I removed it and it's now passing through great.

Thanks for your help!

[sjdaws]

I have this weird set up because I have 3 hypervisors with GPU passthrough (to the worker), but plex/orchestrator need to be able to be moved via vmotion, which can't happen with GPU passthrough.

[pabloromeo]

Do you get the same error if you remove or comment out the environment variable for WORKER_SELECTION_STRATEGY="LOAD_RANK" ? LOAD_RANK is the default anyways.

You can also try removing the double-quotes, don't think they are needed here and it could be causing issues.
My logs shows the following, so I believe removing the quotes should get it to work again:

Initializing orchestrator
Using Worker Selection Strategy: LOAD_RANK
Stream-Splitting: DISABLED
Setting up websockets
Ready
Server listening on port 3500

[pabloromeo]

One other thing you might need to do later once transcoding starts to work, is to possibly mount /tmp on both PMS and the Workers on a shared location. Just doing the transcode location might not work, you may need the entire /tmp to be shared across.
I can't recall exactly but without it some users faced issues with plex's EAE encoder which does stuff in /tmp.
I updated the docs in the README recently, but YMMV.