feat(source): optional exclusion of unschedulable nodes

[Hayajiro]

Description

This fixes a behavioral regression introduced in #4761, where nodes that were previously added to DNS are removed when they are considered unschedulable, for example due to automated maintenance tasks.

This change will introduce a new flag called exclude-unschedulable, which defaults to true in order to keep in line with the current behavior. However, it would also be reasonable to restore the initial behavior before #4761, which would mean to not exclude unschedulable nodes by default.

Fixes #4991

Checklist

• Unit tests updated
• End user documentation updated

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: Hayajiro / name: Markus (9da75f6, 24c45af, 8f61c0f)

[k8s-ci-robot]

Welcome @Hayajiro!

It looks like this is your first PR to kubernetes-sigs/external-dns 🎉. 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/external-dns 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 @Hayajiro. 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. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[mloiseleur]

@Hayajiro you need to sign the CLA in order for this PR to be reviewed.

[Hayajiro]

@Hayajiro you need to sign the CLA in order for this PR to be reviewed.

I'm fully aware of this, but CLA signing is still in process.

[Hayajiro]

CLA is finally signed. 🎉 Also rebased with master and fixed the conflicts.

So I guess this is ready for review now?

[ivankatliarchuk]

/ok-to-test

[ivankatliarchuk]

Can we add tests that capture debug logging?

external-dns/internal/testutils/log.go

Line 36 in e64e536

func LogsToBuffer(level log.Level, t *testing.T) *bytes.Buffer {

You could search for example usages in the code.

[Hayajiro]

Sorry for the late response, but I've been pretty busy the last couple weeks.

I've now added an example of how this could be implemented in the existing test method (which would allow the other testcases to check for logs, as well. Not quite ideal, as this requires disabling parallelism for these testcases, but the alternative would be having to duplicate more of the test logic, which I'm not really a fan of, either.

[Hayajiro]

Not sure why there are apparently data races, as I can't really reproduce them locally...

[ivankatliarchuk]

Feels like our logger is not implemented correctly for tests. Could you try something like here

external-dns/source/gateway_httproute_test.go

Line 1222 in e456a45

if len(tt.logExpectations) == 0 {

and here

external-dns/source/gateway_httproute_test.go

Line 1252 in e456a45

b = testutils.LogsToBuffer(log.DebugLevel, t)

If not going to work, will test logging next time, when I have a more stable framework

[Hayajiro]

When running with make test PARALLEL=64 ITERATION=1000, I was able to reproduce the CI failure, allowing me to take a closer look.

Turns out the problem lies with the fact that, after calling to LogsToBuffer, we don't reset the output buffers, which means we're carrying around stale references. These, in turn, seem to sometimes cause these data races (which seem to only occur if tests are run with enough parallelization and iterations).

When calling to {,k}log.SetOutput(io.Discard), these races seem to no longer occur.

@ivankatliarchuk How do you want to proceed with this? While the fix itself seems to be fairly simple, I'm not quite sure if this is still in scope for this PR.

Edit: Okay, now I can't reproduce this locally anymore and the CI is also green. I feel like this is definitely a flake.

[ivankatliarchuk]

Yeah, is another bug that not related to current work/fix.

[ivankatliarchuk]

/lgtm

cc: @mloiseleur

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: ivankatliarchuk
Once this PR has been reviewed and has the lgtm label, please assign raffo for approval. For more information see the Code Review Process.

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

Needs approval from an approver in each of these files:

• OWNERS

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

[k8s-ci-robot]

New changes are detected. LGTM label has been removed.

[ivankatliarchuk]

Hi @Hayajiro any chance not to force push? I'm not sure now what was reviewed

[ivankatliarchuk]

/label tide/merge-method-squash

[Hayajiro]

Hi @Hayajiro any chance not to force push? I'm not sure now what was reviewed

Well, it was stated that the MR needs rebasing, sooo....

[mloiseleur]

/retitle feat(source): optional exclusion of unschedulable nodes

[mloiseleur]

@Hayajiro PR LGTM. Documentation was updated in #4761. Before we merge it, would you please update node source documentation accordingly, explaining the use case ?