When submitting a job, Conductor's plugins automatically detect required data dependencies from the currently open scene file and uploads them as part of the job submission process.
For some jobs, this may represents multiple gigabytes of data and will require considerable time to upload before rendering can begin. To get your rendering to begin faster, consider uploading data ahead of time so they will be already accessible. Once uploaded, the file's fingerprint will remain available for all jobs sent to your account. As long as the pathing or file doesn't change, it will not need to be uploaded again.
There are two methods of uploading files
- Upload Only submissions
- Upload daemons
Once data is uploaded, it will stay available for that job until deleted. If any dependencies were updated between initial upload and the rendering submittal, the Conductor plugin always does a quick comparison and uploads the newer representations or missing files.
Of course, upload only jobs still can tie up your artists as dependencies are located and uploaded. We recommend for most customers is to run a separate daemon process which handles the uploading of assets and frees up your DCC session on job submission. To do this, you'd need to follow a few steps:
- In your local config.yml (pointed to by the CONDUCTOR_CONFIG environment variable), you would add the line:
This will defer the upload to this separate daemon process. Ideally, this would be run on a separate machine or VM that has access to all of the various assets used in your scene file. The daemon can also be run on your local machine to offload the uploading process from the DCC application to the uploader. In either scenario, to run the daemon process you would execute in a terminal:
To get more help for the uploader command type in:
conductor uploader --help
With most of your files pre-uploaded, sync's are much quicker and the startup times can be dramatically shortened.