Hello,
I'm wondering how input and output files are transferred to and from workers. For instance, how does the maya rendering tool get the input to the workers for rendering, and then collect the output back to the user who spawned the job? Is there a way to transfer the files through Qube or is this functionality that needs to be built externally?
I've run into some trouble regarding mounting the user's environment for the worker to work out of. I would like to potentially have dozens of WinXP Pro PC's submitting jobs to the system, and potentially hundreds of workers executing jobs. One way I was able to get the input files visible to the workers was to set up a drive mapping on my local PC that pointed to my own network share through my network name, then have the worker access the files through that drive letter. This doesn't seem like a scalable solution, as when hundreds of workers try to do bits of work and try to mount my network share, it will easily exceed windows' connection limit and deny them access. Is it typical for Qube to work by having someone that's submitting a job first upload the data to be worked on to a proper file server, and then have the workers all work off of it? Or is there a way I can send the input data to the worker, and get the output back without using windows file sharing?
Thanks in advance for any tips!
