Closed KernAttila closed 3 weeks ago
Can i try please??
Sure, I'll give you the proper branches to try it out when I'm back home tonight. You'll need to build cuebot and use my updated cuesubmit branch to test it out. So far we've been successful with this implementation
You will need to build cuebot against this branch : https://github.com/Wolf-Pipeline/OpenCue/tree/cuebot-reserve-all-cores instructions here: https://www.opencue.io/docs/getting-started/deploying-cuebot/#option-4-build-from-source
And pull this branch to run cuebot + cuesubmit : https://github.com/Wolf-Pipeline/OpenCue/tree/cuesubmit-dev
Let me know if you need some assistance.
Is your feature request related to a problem? Please describe. Our park is made of machines with different core count. We can't make dispatching rules based on a fixed number of cores to have only one job per machine, some machines will still pick up 2 jobs.
Describe the solution you'd like As in many render engines, we should be able to set a negative core requirement.
Describe alternatives you've considered As an alternative, we can put hosts in different allocation groups according to their resources. But it makes handling jobs quite tedious.