So when launching Nuke from SG Desktop I get an alert “The Frame Server has encountered an error.” and I just click on the continue and I can keep working but it would be nice to know what is going on.
Complete Alert:
The Frame Server has encountered an error.
Nuke 12.1v5, 64 bit, built Sep 30 2020.
Copyright (c) 2020 The Foundry Visionmongers Ltd. All Rights Reserved.
Loading - init.py
Traceback (most recent call last):
File “/Applications/Nuke12.1v5/Nuke12.1v5.app/Contents/Resources/pythonextensions/site-packages/foundry/frameserver/nuke/workerapplication.py”, line 18, in
from util import(asUtf8, asUnicode)
ImportError: cannot import name asUtf8
cannot import name asUtf8
Hi @dschwarz
That’s not an error I’ve seen before, so it is likely something specific to your setup.
I presume it works fine when not launched from Shotgun?
It looks like it is failing to import the asUtf8
function from /Applications/Nuke12.1v5/Nuke12.1v5.app/Contents/Resources/pythonextensions/site-packages/foundry/frameserver/nuke/util.py
Maybe there is something in your environment that is causing an issue here. Are you for example including standard Python modules or site-packages from a different Python instance?
Thanks Phillip I’ll look into that.
We did figure out a work around which also lead us to discover that this is what was causing a pause on every 18th frame in Nuke when playing back. The work around is to launch Nuke first not using SG Desktop and then launch Nuke with SG Desktop. No error and we do not get the every 18th frame pause happening now.
<snark>I’m sure the error the Frame Server is reporting is that it realizes its own existence, and how it totally sucks.</snark>
FWIW, we’re been disabling the frameserver since its intro (it interferes with other tools) with the flag --disable-nuke-frameserver
under launch arguments in tk-multi-launchapp without any noticeable issues.
3 Likes
Thanks, I’ll try out disabling and see if that makes us happy.
Can either of you point me to documentation on what the frame server is and does?
Again, thanks!
So we actually found the culprit on our end. There was a dev path still on the config. Our bad!
Thanks guys for helping!
2 Likes