I’m looking into creating scripts to rsync scene dependencies, generate low-rez JPEGs, etc. if a job is a cloud job, but how do I know a job is cloudy and not local? Is there a property to the job, or can I somehow tag anything going through the balancer to have the pre and post jobs appended to a job?
Sorry for being vague. The whole thing is still very unclear in my head as I work this use case out.
Thank you,
-ctj
ok. so, if I have something that needs to happen on cloud VMs and not locally, then I need to
- have a group for cloud machines (‘cloudMachineGrp’)
- have a pre-render task system that checks for the group membership of the task’s assigned slave and execute accordingly
Sound about right?
-ctj
Yeah that’s exactly how I’d do it. If you submit the job to the cloud group you won’t even need to check the group because only slaves in that group will pick up the job.
Eric
the issue I’m dealing with is path remapping. since we’re using maya binary (.mb) files, path mapping doesn’t work for paths internal to the .mb file, so if a render is going to the cloud group, then I’ll hand it the ‘cloudy’ version of the file, with internal paths remapped to suite the cloud, but if it’s in any other group, I need it to keep the local file system paths.
So far so good. Local remapping is working locally and cloud regions are mapping the their cloudy stuff, except for the paths internal to the maya binary files, as expected, so I have a submit script that creates a “cloud” version of the binary scene with the paths for the cloud drives mapped internal to the binary file as well as populating the asset dependencies so crawler knows what to crawl for. Seems to be working so far with simple files. Today we test more production worthy files.
Thanks for the help.
-ctj