Add ability to provide custom CoreDNS tolerations and affinity

[hierynomus]

No description provided.

[k8s-ci-robot]

Thanks for your pull request. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

📝 Please follow instructions at https://git.k8s.io/community/CLA.md#the-contributor-license-agreement to sign the CLA.

It may take a couple minutes for the CLA signature to be fully registered; after that, please reply here with a new comment and we'll verify. Thanks.

---

• If you've already signed a CLA, it's possible we don't have your GitHub username or you're using a different email address. Check your existing CLA data and verify that your email is set on your git commits.
• If you signed the CLA as a corporation, please sign in with your organization's credentials at https://identity.linuxfoundation.org/projects/cncf to be authorized.
• If you have done the above and are still having issues with the CLA being reported as unsigned, please log a ticket with the Linux Foundation Helpdesk: https://support.linuxfoundation.org/
• Should you encounter any issues with the Linux Foundation Helpdesk, send a message to the backup e-mail support address at: login-issues@jira.linuxfoundation.org

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/test-infra repository. I understand the commands that are listed here.

[k8s-ci-robot]

Welcome @hierynomus!

It looks like this is your first PR to kubernetes/kops 🎉. 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/kops 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 @hierynomus. Thanks for your PR.

I'm waiting for a kubernetes 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/test-infra repository.

[hierynomus]

CLA has been signed.

[hierynomus]

@mikesplain @olemarkus Is there something more needed for this PR? Or you just haven't gotten around to it yet?

[olemarkus]

Hi, sorry it is on my list. Have you tested the PR locally? I'd be surprised if this actually works.

Further, offering this kind of customization on a core addon is unlikely to be added. We'd at least have to understand the usecase. Kops will not really work on a cluster where there aren't worker nodes workloads cannot be scheduled on by default.

[hierynomus]

Hi Ole, No problem!
Yes, we've tested this on our own environment using a source-built kops executable in our provisioning pipeline. The change itself does what it needs to do.

The usecase for us is that our cluster is configured in such a way that by default we want everything in our k8s clusters to run on AWS spot nodes. But CoreDNS needs to be scheduled on the master and on-demand nodes only.

With this change the default behavior is unchanged, but it is now allowed to specify custom tolerations and affinity such as this (taken from our kops cluster definition):

  kubeDNS:
    provider: CoreDNS
    tolerations:
      - effect: NoSchedule
        operator: Exists
    affinity:
      nodeAffinity:
        requiredDuringSchedulingIgnoredDuringExecution:
          nodeSelectorTerms:
          - matchExpressions:
            - key: kops.k8s.io/instancegroup
              operator: In
              values:
              - master
              - ondemand-nodes
      podAntiAffinity:
        preferredDuringSchedulingIgnoredDuringExecution:
        - podAffinityTerm:
            labelSelector:
              matchExpressions:
              - key: k8s-app
                operator: In
                values:
                - kube-dns
            topologyKey: kubernetes.io/hostname
          weight: 100

[hierynomus]

@olemarkus Is the usecase that I described clear?

[olemarkus]

So this is essentially to avoid coredns from running on spot instances? CoreDNS should work fine on spot instances. Even more so if the NTH addon is enabled. I would specifically avoid running anything on master nodes that absolutely does not have to run there.

[hierynomus]

I think it's more generic, that in the case of setting up a kubernetes cluster with multiple instance-groups, that you would want to be able to control where certain pods land. We see the same at our customers who want to target specific workloads to specific instance-groups.

[olemarkus]

Brought this up in office hours today and we decided this feature looks good.

/ok-to-test
/assign

[hierynomus]

/retest

[hierynomus]

@olemarkus What is still needed to proceed with this PR? I see that there are a few CI jobs which require an approval before they're run (build-*-amd64)

Other than that, I think it needs a code-review/approval?

[hierynomus]

I took the liberty to create a test for the change this PR adds: 238670b
It can be a bit tricky for first-time contributors to create these, so I thought this may save some time.
Can you cherry that commit into this PR?

It would also be nice if you could squash some of the commits.

Thanks a lot! That would've indeed cost me considerably more time. I've added your commit and squashed them all into one.

[hierynomus]

/retest

[olemarkus]

You need to run ./hack/update-expected.sh to fix those failures

[hierynomus]

You need to run ./hack/update-expected.sh to fix those failures

That's indeed what I tried locally before doing the retest, but it didn't show any changes after that command. I'll re-run it

[hierynomus]

@olemarkus If I just urn ./hack/update-expected, it executes a bunch of tests, but at the end there is no difference to any of the files. Am I missing something in how to execute this?

[olemarkus]

Not sure if you noticed, but hack/update-expected.sh fails because of the following error:

bootstrapchannelbuilder_test.go:174: error from BootstrapChannelBuilder Build: error reading manifest addons/coredns.addons.k8s.io/k8s-1.12.yaml: error opening resource: error executing resource template "addons/coredns.addons.k8s.io/k8s-1.12.yaml": error executing template "addons/coredns.addons.k8s.io/k8s-1.12.yaml": template: addons/coredns.addons.k8s.io/k8s-1.12.yaml:121:32: executing "addons/coredns.addons.k8s.io/k8s-1.12.yaml" at <8>: wrong type for value; expected string; got []v1.Toleration

[hierynomus]

Missed that in the build logs, weird thing is that doesn't happen locally. Does that behave differently when run on the CI env?

[olemarkus]

It should happen to you locally as well. Do you see this command exiting successfully?

[hierynomus]

Ok, after merging in master, I had the same error locally. The template now uses a newly added ToYAML TemplateFunction to render the v1.Tolerations and v1.Affinity objects as yaml.

[hierynomus]

@olemarkus Any idea about why that test fails? Is it something I changed, or is it flaky somehow?

[olemarkus]

Looks like a flake

/retest

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: olemarkus

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 [olemarkus]

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

[hierynomus]

Thanks @olemarkus! Two tests now failing on that cni-kuberouter test, looks like flakes indeed. /retest

[hierynomus]

/test pull-kops-e2e-cni-kuberouter

[hierynomus]

Thanks for the guidance @olemarkus!