Just realizing now our Pipeline guy and I were a bit lazy and actually labeled it as Nuke 12? LOL yeah its not Nuke 12 but here is what Ive got…
Category=Nuke 12 Render Executables
Label=Nuke 11.1 Render Executable
Description=The path to the Nuke 11.2 executable file used for rendering. Enter alternative paths on separate lines.
Our Nuke Upgrade Path was very similar to yours. Nuke 10.0v1—>Nuke11.0v4---->Nuke 11.2v3
We found some bugs along that way actually that the Foundry have fixed in .2v4 but havn’t taken the plunge yet with upgrading. Keeping those versions consistent was important for Published Files in our Shotgun Pipeline.
Anyway once that is there in the parm file, make sure it is reflecting and showing in your deadline monitor under Tools > Configure Plugins > Nuke. Then any Nuke Job sent from 11.2v3 should know where to render once it hits your farm. I could be forgetting something so by all means if someone at Thinkbox thinks otherwise please let us know. Once we added that we off to the renders/races. Thats more or less how we figured it out, submitting jobs at first were erroring out and it was pretty obvious why from the logs.
Going back to your question on keeping configs consistent etc. We have taken over the default deadline submissions script, and have it shared on our local NAS. It actually doesn’t give artists many options, just have good job submission arguments we defined within the script. For EG 2 simultaneous concurrent tasks, and a specific pool for jobs requiring specific Nuke plugins such as the Neat Video OFX plugin.
Then just have an env variable of NUKE_PATH and point it to the respective operating systems path on disk to your scripts on storage. WIN Z:\settings\nuke NIX:/mnt/server/settings/nuke OSX:/Volumes/server/settings/nuke
Hope this helps you!