AWS Thinkbox Discussion Forums

Batch-wide machine limit?

We’re doing some tiled rendering, but when we have 100 tiles it fills up the farm quite quickly. Other jobs submitted can’t get a hold of any of those slaves, and if those tiles still take 1+ hours to render, we have a lot of jobs building up.

Looking at Machine Limits seems promising, however that is still for each job within the batch. Meaning if I set a limit of 50, each job within the batch will only use 50 slaves. Not helpful.

We’re going to look into pools, etc, for prioritizing a little better, but is there a way to tell a batch to not allow more than “x” number of slaves to run on any of its child jobs?

Hello,

What you would want to do here is also use Limit Groups( docs.thinkboxsoftware.com/produ … tml#limits ), not just Machine Limits. Limit groups will apply limits to all jobs in the limit group not just to individual jobs. This would allow you to, say, restrict all jobs to just 50% of your render nodes, leaving the rest for other work.

Privacy | Site terms | Cookie preferences