You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I agree with you that capacity and allocatable may not be same for batch and mid; besides, I think we need to give the definition about the gap between capacity and allocatable(maybe batch reservation?)
In the beginning, I want to use the capacity to solve Cluster Autoscaler(CA) problem.
when pod pending CA will scale out new node to satisfy user requests, it will select one node randomly from the cluster, use the allocatable of the node to calculate how many new node needed.
Because the allocatable is always changing, so I want to make capacity be stable to solve the CA problem.
So I think the capacity of batch should be this: node capacity * reclaim threshold - reserved, which means the max resource can be used for batch.
This issue has been automatically marked as stale because it has not had recent activity.
This bot triages issues and PRs according to the following rules:
After 90d of inactivity, lifecycle/stale is applied
After 30d of inactivity since lifecycle/stale was applied, the issue is closed
You can:
Mark this issue or PR as fresh with /remove-lifecycle stale
Close this issue or PR with /close
Thank you for your contributions.
What is your proposal:
Change capacity of batch/mid resource to node capacity * reclaimThreshold to make use of capacity field.
https://github.com/koordinator-sh/koordinator/blob/main/pkg/slo-controller/noderesource/plugins/batchresource/plugin.go#L147
Why is this needed:
Is there a suggested solution, if so, please add it:
The text was updated successfully, but these errors were encountered: