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

Descheduler profiles #548

Merged

Conversation

damemi
Copy link

@damemi damemi commented Nov 24, 2020

No description provided.

@damemi
Copy link
Author

damemi commented Nov 24, 2020

/cc @soltysh @ingvagabund

last-updated: 2020-11-23
status: provisional
see-also:
- "https://github.com/openshift/enhancements/pull/541/"
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't expect the other PR not merging, so I'd just link to the file like below.


## Release Signoff Checklist

- [ ] Enhancement is `implementable`
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Check this, we are implementing this right away.

type KubeDeschedulerSpec struct {
Profiles []string `json:"profiles"`
...
}
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Another alternative is a list of enums.

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

are there any opinions on preference for either one? I'm not attached either way but curious if anyone else had any insight

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

let's go with a list of enums, we can then re-use the kubebuilder validation for enums for that type. The only thing we'll have to ensure is to de-duplication or call out what happens and validation if contradicting profiles are applied.

Ideally, this field will immediately replace the existing `strategies` config as the initial GA option for
configuring the descheduler.

However, since the current API is version `v1beta1` we may be required to continue supporting the
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'd clearly express the deprecation timeline, which is Beta: 9 months or 3 releases (whichever is longer), for beta-level APIs. For conversions you'll need to have both (strategies and profile) added to both APIs, but v1beta1 version should be deprecated right away, same with strategies field in v1, and point to using the new profiles.

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ok, so for the v1 API we should still provide strategies with a note that it's deprecated and will be removed? Or just provide the v1 API with only profiles and still support the v1beta1 api with strategies until that API is removed according to that timeline?

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@damemi yeah, for v1 provide deprecated strategies. We should also announce deprecation of v1beta1 too.

@damemi damemi force-pushed the descheduler-profiles branch 2 times, most recently from 2f6b9c0 to 27735e8 Compare November 30, 2020 19:16
Copy link
Author

@damemi damemi left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

updated with feedback and a couple more questions

type KubeDeschedulerSpec struct {
Profiles []string `json:"profiles"`
...
}
Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

are there any opinions on preference for either one? I'm not attached either way but curious if anyone else had any insight

Ideally, this field will immediately replace the existing `strategies` config as the initial GA option for
configuring the descheduler.

However, since the current API is version `v1beta1` we may be required to continue supporting the
Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ok, so for the v1 API we should still provide strategies with a note that it's deprecated and will be removed? Or just provide the v1 API with only profiles and still support the v1beta1 api with strategies until that API is removed according to that timeline?

@damemi
Copy link
Author

damemi commented Dec 1, 2020

To clarify on what we decided for this, we will add both fields to v1beta1 and remove functionality for the old field (only providing it for transition support). When v1beta1 is removed in favor of v1, those fields will be removed and not present in v1. I have updated the enhancement to include this.

Copy link

@soltysh soltysh left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This was discussed earlier today during our call and this was agreed between @damemi, @ingvagabund, @deads2k and myself

/lgtm
/approve

@openshift-ci-robot openshift-ci-robot added the lgtm Indicates that a PR is ready to be merged. label Dec 1, 2020
@openshift-ci-robot
Copy link

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: damemi, soltysh

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

@openshift-ci-robot openshift-ci-robot added the approved Indicates a PR has been approved by an approver from all required OWNERS files. label Dec 1, 2020
@openshift-merge-robot openshift-merge-robot merged commit 81bfc8d into openshift:master Dec 1, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
approved Indicates a PR has been approved by an approver from all required OWNERS files. lgtm Indicates that a PR is ready to be merged.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants