CGRU Forum

Is it possible to resolve the job name once a job has been submitted? Alternatively, is it possible to have dependency masks that are not strings (job id, or something of that nature)? In Houdini we often build a reusable network of dependencies to execute on the farm. if we resubmit the network bef...

I could have swore I got this to work before -- but for some reason, i'm unable to do so right now.

What settings do I need to set on the afanasy rop in order for it ONLY generate ifds, and not perform a mantra render?

G

Thanks Timur,

And thanks for pointing me at GitHub. I keep meaning to learn about it but never seem to find the time... As you might have guessed from the hacky code, software development is not really my forté - I'm more of a motion blur nerd

Cheers,

G

Heya Timur, We've recently had to modify render.py in the nuke plugins folder, line 166 to: if writenode.Class() != 'Write' and writenode.Class() != 'DeepWrite': errorExit('Node "%s" class is not "Write".' % xnode, True and a few small tweaks to afanasy.py in $GCRU_LOCATION/plugins/nuke/scripts/cgru...

You are right -- it's possible.

It might be convenient (read: more discoverable for new users) if you allowed it all to happen in one node since you DO allow the 'separate render' from inside one node.

Just a suggestion.

G

Heya Timur, It would be very beneficial to be able to control the "frames per task" parameter separately for the ifdgen stage, and the mantra stage. As you know, the ifdgen stage is relying on engine licenses, while mantra itself is free. I'd like to be able to submit a 1 frame /task job for mantra ...

Heya Timur,

Since each job knows the name of the user that submitted it, is possible to assign ownership/group of the final frame to that user once the task is completed?

At the moment all renders are owned by either root or nonrootuser (if specified).

G

So after all this noise... I was completely off track. After checking the logs in /var/tmp I realized that what was crashing was the convert post-process (for thumbnails), and not the actual mayatovray job. I've updated ImageMagick to the current version, and the crashes went away. What tipped me of...

Timur, would you be open to the idea of also checking the process exit code in conjunction with parsing the logs to flag errors?

G

Upon investigating further, I was wrong my my assumption that it wasn't tripping the parser. While the mayatovray parser isn't tripped, it seems that the error is flagged by the parent class parser.py. line 109: for string in self.str_error: if lower.find(string.lower()) != -1: self.error = True Thi...