Join GitHub today
GitHub is home to over 31 million developers working together to host and review code, manage projects, and build software together.Sign up
Unify priority functions pattern #51455
There are still three priority funcions(the whole one) named
I have completed one refactoring named
If all priority functions refactoring done by using map/reduce pattern, we will benifit from:
I am not list all things, just hopes to initiate analyzing.
Unify the priority functions pattern if we can.
referenced this issue
Aug 30, 2017
Yes - I've always wanted to do that but never really had time for it.
I'm happy to review your PR for spreading, but will not have time before code freeze (and I think we shouldn't rush for 1.8 with that).
The Pod affinity/anti-affinity might be more tricky to do, because they really depend on the whole state of the cluster. But maybe it's actually time to thing and solve pod affinity/anti-affinity more properly.
@wojtek-t , one more input for this is to add an interface for those algorithm to include
Sorry, I think I'm not really following. Can you clarify?
Yes - this one is tricky. I think no matter how hard we try, if we wan to solve if efficiently, we need to consider all nodes at once. It seems that the workaround here might be to actually precompute results in the precomputation phase, and in real map/reduce functions just grab the results from the metadata that is passed there. I can't see any better option for doing this.
Anyway - if you have any better idea, let me know - I'm open to any suggestions.