Hi,
Typically, clients start Maya via [MayaBatch.exe/maya] or [Render.exe/Render]. The advantage here is that the running executable process can have env variables injected into just this instance of the running executable instead of injecting the env variables permanently onto the system executing the job. Both options have their own merits. I see that you have already discovered that a job post-submission, can have 1 or more environment variables defined via the right-click job, “Modify Job Properties…”, “Environment” section - “add” to add a KEY=VALUE pair. Notice the checkbox at the bottom to inject only into the currently rendering process (one of the Maya exe’s as mentioned above).
You can inject environment variables at submission time via adding this entry to the job_info_file as described here (together with all the other options):
thinkboxsoftware.com/deadlin … nt_Options
Note the other options which may be of interest here as well:
“IncludeEnvironment”
“UseJobEnvironmentOnly”
Depending on how you are currently submitting your Maya jobs to your queue, you would just need to modify the appropriate submission script (all open code - mel, py) to include the additional functionality you require:
thinkboxsoftware.com/deadlin … Submission
Alternatively, you could hook into the Event Plugin system and write a py event which reacts to jobs of a certain type such as Maya being submitted into the queue and react accordingly with regard injecting environment variables. See here for more info:
thinkboxsoftware.com/deadlin … eventssdk/
As a side note, GPU rendering doesn’t work if you are running Deadline Slave as a service. See here for an additional feature we added recently for Redshift:
thinkboxsoftware.com/deadlin … Render_Job
Search for: “If rendering with Redshift, there will be an additional Redshift Options under the Maya Options:”
Regards,
Mike