kubernetes-sigs / descheduler

Descheduler for Kubernetes
https://sigs.k8s.io/descheduler
Apache License 2.0
4.23k stars 645 forks source link

fix(podlifetime): fix failed unittest #1334

Closed xujihui1985 closed 6 months ago

xujihui1985 commented 6 months ago

correct desired pod creation time

"Two pods in the dev Namespace, 1 created 605 seconds ago. 1 should be evicted.", the create time stamp should be 605 s before current time

k8s-ci-robot commented 6 months ago

Welcome @xujihui1985!

It looks like this is your first PR to kubernetes-sigs/descheduler 🎉. Please refer to our pull request process documentation to help your PR have a smooth ride to approval.

You will be prompted by a bot to use commands during the review process. Do not be afraid to follow the prompts! It is okay to experiment. Here is the bot commands documentation.

You can also check if kubernetes-sigs/descheduler has its own contribution guidelines.

You may want to refer to our testing guide if you run into trouble with your tests not passing.

If you are having difficulty getting your pull request seen, please follow the recommended escalation practices. Also, for tips and tricks in the contribution process you may want to read the Kubernetes contributor cheat sheet. We want to make sure your contribution gets all the attention it needs!

Thank you, and welcome to Kubernetes. :smiley:

k8s-ci-robot commented 6 months ago

Hi @xujihui1985. Thanks for your PR.

I'm waiting for a kubernetes-sigs member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available [here](https://git.k8s.io/community/contributors/guide/pull-requests.md). If you have questions or suggestions related to my behavior, please file an issue against the [kubernetes/test-infra](https://github.com/kubernetes/test-infra/issues/new?title=Prow%20issue:) repository.
xujihui1985 commented 6 months ago

/retest

ingvagabund commented 6 months ago

@xujihui1985 nice catch!!!

I wonder how "Two pods in the dev Namespace, 1 created 605 seconds ago. 1 should be evicted." test has been passing until now when it should not.

Also, "Two pods in the dev Namespace, 1 is new and 1 very is old. 1 should be evicted." and "Two pods in the dev Namespace, 1 created 605 seconds ago. 1 should be evicted." tests are functionally equivalent wrt. maxLifeTime. The same holds for "Two pods in the dev Namespace, 2 are new and 0 are old. 0 should be evicted." and "Two pods in the dev Namespace, 1 created 595 seconds ago. 0 should be evicted.".

xujihui1985 commented 6 months ago

@xujihui1985 nice catch!!!

I wonder how "Two pods in the dev Namespace, 1 created 605 seconds ago. 1 should be evicted." test has been passing until now when it should not.

Also, "Two pods in the dev Namespace, 1 is new and 1 very is old. 1 should be evicted." and "Two pods in the dev Namespace, 1 created 605 seconds ago. 1 should be evicted." tests are functionally equivalent wrt. maxLifeTime. The same holds for "Two pods in the dev Namespace, 2 are new and 0 are old. 0 should be evicted." and "Two pods in the dev Namespace, 1 created 595 seconds ago. 0 should be evicted.".

I happened to run the test in my ci system, and can't pass this test, so I look into the test and find out the test condition might be wrong, but I haven't check why this test used to be success.

ingvagabund commented 6 months ago

Running the test from the current master:HEAD it passes. Though only because the computation overflows:

    podFilter = podutil.WrapFilterFuncs(podFilter, func(pod *v1.Pod) bool {
        podAgeSeconds := uint(metav1.Now().Sub(pod.GetCreationTimestamp().Local()).Seconds())
        return podAgeSeconds > *podLifeTimeArgs.MaxPodLifeTimeSeconds
    })

The code does not assume a pod gets created in the future :) Dropping uint the case fails.

ingvagabund commented 6 months ago

@xujihui1985 would you please also update the filter code as well? E.g.

    maxPodLifeTimeSeconds := int(*podLifeTimeArgs.MaxPodLifeTimeSeconds)
    podFilter = podutil.WrapFilterFuncs(podFilter, func(pod *v1.Pod) bool {
        podAgeSeconds := int(metav1.Now().Sub(pod.GetCreationTimestamp().Local()).Seconds())
        return podAgeSeconds > maxPodLifeTimeSeconds
    })

@a7i we should probably ignore any pod created in the future :)

xujihui1985 commented 6 months ago

sure, glad to do that

ingvagabund commented 6 months ago

/approve

k8s-ci-robot commented 6 months ago

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: ingvagabund

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: - ~~[OWNERS](https://github.com/kubernetes-sigs/descheduler/blob/master/OWNERS)~~ [ingvagabund] Approvers can indicate their approval by writing `/approve` in a comment Approvers can cancel approval by writing `/approve cancel` in a comment
xujihui1985 commented 6 months ago

@xujihui1985 would you please also update the filter code as well? E.g.

  maxPodLifeTimeSeconds := int(*podLifeTimeArgs.MaxPodLifeTimeSeconds)
  podFilter = podutil.WrapFilterFuncs(podFilter, func(pod *v1.Pod) bool {
      podAgeSeconds := int(metav1.Now().Sub(pod.GetCreationTimestamp().Local()).Seconds())
      return podAgeSeconds > maxPodLifeTimeSeconds
  })

@a7i we should probably ignore any pod created in the future :)

I see, covert to uint make the podAgeSeconds a large number because of the nagtive sign, I have update the MR according to your suggestion, but actually this future pod doesn't make sense in real life, so either we compare the creation seconds with convert it to int or if (podAgeSeconds < 0), we just return false

xujihui1985 commented 6 months ago

/retest

a7i commented 6 months ago

/lgtm