Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

user-scheduler: tweak modern configuration #1782

Merged

Conversation

consideRatio
Copy link
Member

@consideRatio consideRatio commented Sep 21, 2020

This is a followup on #1778 which didn't ensure that pods are scheduled on the most used node correctly because there were some remnant competing logic. This PR disables the competition tothe primary logic we want to dictate the scheduling in the NodeResourcesMostAllocated plugin.

See #1778 (comment) for a discussion on this PR.

As this PR is a complementary fix to #1778 recently merged, I'll go ahead and merge this directly if tests succeeds.

@consideRatio consideRatio merged commit d363004 into jupyterhub:master Sep 21, 2020
@consideRatio
Copy link
Member Author

consideRatio commented Sep 21, 2020

Tested to have worked out well on k8s 1.17 woop woop! 🎉 Now no scheduler plugin influence the node-scoring to disrupt the important plugin: NodeResourcesMostAllocated.

InterPodAffinity            scores on pod1 => [{node1 0} {node2 0}]
NodeAffinity                scores on pod1 => [{node1 0} {node2 0}]
NodePreferAvoidPods         scores on pod1 => [{node1 1000000} {node2 1000000}]
PodTopologySpread           scores on pod1 => [{node1 200} {node2 200}]
TaintToleration             scores on pod1 => [{node1 0} {node2 100}]
NodeResourcesMostAllocated  scores on pod1 => [{node1 8} {node2 55}]

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant