Follow-up to Flavor Fungibility Prefers Fit over NoBorrow#6181
Follow-up to Flavor Fungibility Prefers Fit over NoBorrow#6181k8s-ci-robot merged 3 commits intokubernetes-sigs:mainfrom
Conversation
| if len(candidates) == 0 { | ||
| return preemptioncommon.NoCandidates, borrowAfterPreemptions | ||
| borrow, _ := classical.FindHeightOfLowestSubtreeThatFits(cq, fr, quantity) | ||
| return preemptioncommon.NoCandidates, borrow |
There was a problem hiding this comment.
It actually matters what value of borrow we return with NoCandidates because it is later used to order entries in scheduling.
There are tests that were failing whenever I was trying to return some fixed value to this actually has to be correctly calculated. For that reason I can't think of any new testcase needed for this.
For example the new unit tests for capacity blocking (capacity blocked when lending clusterqueue not guaranteed to reclaim (ReclaimWithinCohort=Never)).
There was a problem hiding this comment.
Awesome, thank you for checking that we had unit tests already.
I like unit tests to be our first line of defense whenever possible (especially for scheduling), cause they are much easier to debug than integration or e2e.
Co-authored-by: gabesaba <15304068+gabesaba@users.noreply.github.com>
✅ Deploy Preview for kubernetes-sigs-kueue canceled.
|
|
/lgtm thanks! |
|
LGTM label has been added. DetailsGit tree hash: 5450ae3e52986518fd3688dd1bdaf05a58cedf35 |
|
/lgtm |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: gabesaba, mimowo, pajakd The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
race condition :) |
…-sigs#6181) * Followup * Update pkg/scheduler/preemption/preemption_oracle.go Co-authored-by: gabesaba <15304068+gabesaba@users.noreply.github.com> * Fix whitespaces --------- Co-authored-by: gabesaba <15304068+gabesaba@users.noreply.github.com>
What type of PR is this?
/kind cleanup
What this PR does / why we need it:
Follow-up to #6132 addressing the comments #6132 (comment)
Which issue(s) this PR fixes:
Fixes #
Special notes for your reviewer:
Does this PR introduce a user-facing change?