🐛 ensure envtest stops the whole process group

[dnwe]

Previous PR #2560 attempted to prevent envtest from leaking orphaned etcd / kube-apiserver etc. processes by adding a SysProcAttr of Setpgid: true. However, it made a few mistakes:

• It incorrectly described the change as causing the process to "inherit the parent's process group id", which is the opposite of what it does. In fact it causes a brand new process group to be created for the child process, which is the actual behaviour we want, so correct the comment.
• Whilst it added a new process group, it didn't update the signalling to send SIGTERM (and eventually SIGKILL) to the process group (by using a - prefix) and was still just sending the signal to the process itself [1].
• It didn't add a unittest to assert on the desired behaviour before and after the changes.

While with this PR the whole process group is stopped when stopping envtest, this doesn't really result in behavioral changes because the processsgroup consists of one process in all circumstances (either kas or etcd).

[1] https://medium.com/@felixge/killing-a-child-process-and-all-of-its-children-in-go-54079af94773

[k8s-ci-robot]

Welcome @dnwe!

It looks like this is your first PR to kubernetes-sigs/controller-runtime 🎉. 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/controller-runtime 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. 😃

[k8s-ci-robot]

Hi @dnwe. 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.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[alvaroaleman]

/ok-to-test

[alvaroaleman]

While your updated description is correctly describing what already happened, I am not sure I understand why that actually matters. Unless I am missing something, we directly start the apiserver and etcd binary without an intermediate like a shell so it shouldn't make a difference if we send a signal to them directly or to the processgroup?

I've seen the stale etcd and apiserver processes too, but my theory would be that happens because the test dies and Stop() doesn't run, not because the cleanup is buggy. But maybe I am missing something here?

[dnwe]

Ah yes sadly you're right I think that this won't help the etcd/apiserver case. I hadn't actually checked how they were being started, I'd just spotted the leaking processes and then found #2560 which looked a bit wrong.

As you say, it seems like in pkg/internal/testing/controlplane/plane.go we just start them one after the other as singleton processes. So I think the PR is still right, in that it corrects the behaviour when a group of processes is spawned, but it's more of a housekeeping thing rather than actually improving the enduser experience when running envtest 😔

[sbueringer]

Would it make sense to slightly change the PR description (and PR title) to surface that this won't change anything for envtest?

[alvaroaleman]

Thanks!

[k8s-ci-robot]

LGTM label has been added.

Details

Git tree hash: 3916cadbe6f68a45ee0daef4f00af73c88272893

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: alvaroaleman, dnwe

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

The pull request process is described here

Details

Needs approval from an approver in each of these files:

• OWNERS [alvaroaleman]

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