Closed mimowo closed 1 week ago
Name | Link |
---|---|
Latest commit | 985b955410c864d405428a25176037f85ec5547b |
Latest deploy log | https://app.netlify.com/sites/kubernetes-sigs-kueue/deploys/673b81cb9fcc4c000804500c |
/assign @PBundyra cc @tenzen-y
/test pull-kueue-test-tas-e2e-main
LGTM, good job! I'll hold for @tenzen-y to take a look /hold /lgtm
LGTM label has been added.
[APPROVALNOTIFIER] This PR is APPROVED
This pull-request has been approved by: mimowo, tenzen-y
The full list of commands accepted by this bot can be found here.
The pull request process is described here
/hold cancel
Do we have integration testing for this situation?
no, but we have unit testing which is quite useful here, because it is relatively high-level (for the exported function), which makes it similar to TestSchedule which proved to be very useful over time.
So, personally I'm using unit tests for debugging, the e2e tests are useful to confirm the feature works.
I considered adding integration tests, but they are actually tricky, because TopologyUngater operates on pods, but we don't enable the k8s Job controller in the integaration tests (no kube-controller-manager, only API server). So, we would need to create the pods ourselves, just like in unit tests, so I don't see much gain.
What type of PR is this?
/kind cleanup
What this PR does / why we need it:
Which issue(s) this PR fixes:
Part of #3450 Part of #3533
Special notes for your reviewer:
rack(pod0)==rack(pod1)
andrack(pod2)==rack(pod3)
, but I think it will be harder to debug when the assert fails. With the exact assignment check we can easily see the full assignment.I ran it a 4 times and it failed all repeats.
Does this PR introduce a user-facing change?