enable skyDNS round robin A/AAAA records

[wu8685]

After issue #116 fixed, I hope we can enable skyDNS round robin to support load balance for headless service.

[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://github.com/kubernetes/kubernetes/wiki/CLA-FAQ 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 email the CNCF helpdesk: [email protected]

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.

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[MrHohn]

/ok-to-test
/remove-lifecycle stale

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[wu8685]

Shall we consider this PR to enable round robin in order to support LB for multiple pods behind a headless service ?

[thockin]

What is the end result of this? If you have a headless service with 3 endpoints, every query should get all 3 addresses, no? Does this just shuffle them?

[miekg]

[ Quoting <[email protected]> in "Re: [kubernetes/dns] enable skyDNS ..." ]

What is the end result of this? If you have a headless service with 3 endpoints, every query should get all 3 addresses, no? Does this just shuffle them?

I believe so, yes.

[wu8685]

Yes, it dose shuffle:)

However, it is not enough.
When we use the dns only has shuffle to LB the headless svc, the backend server does not be hit equally.

I have done some test to count the backend server hits. The result is listed below:

// enable RR
ip: 10.212.151.227, count: 530
ip: 10.212.151.218, count: 470

// disable RR
ip: 10.212.151.218, count: 873
ip: 10.212.151.227, count: 127

[miekg]

[ Quoting <[email protected]> in "Re: [kubernetes/dns] enable skyDNS ..." ]

Yes, it dose shuffle:) However, it is not enough. When we use the dns only has shuffle to LB the headless svc, the backend server does not be hit equally. I have done some test to count the backend server hits. The result is listed below: ``` // enable RR ip: 10.212.151.227, count: 530 ip: 10.212.151.218, count: 470 // disable RR ip: 10.212.151.218, count: 873 ip: 10.212.151.227, count: 127 ```

this just randomizes, has no concept of load or any other external factors.

[wu8685]

Is it a good ideal to export the round-robin as a kube-dns configuration ?
Users can decide themselves.

[thockin]

Shuffling seems OK. I assumed that clients would randomly pick from a response set, but maybe they tend to take the first item. Low cost to shuffle, I guess. The real issue is whether there's actually going to be another kube-dns release or not.

…

On Fri, Jul 6, 2018 at 3:15 AM Kan Wu ***@***.***> wrote: Is it a good ideal to export the round-robin as a kube-dns configuration ? Users can decide themselves. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#167 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFVgVP3_fg-qkWQtGQsL_DVvpS0cWKPNks5uDzjMgaJpZM4QkJkM> .

[miekg]

[ Quoting <[email protected]> in "Re: [kubernetes/dns] enable skyDNS ..." ]

Shuffling seems OK. I assumed that clients would randomly pick from a response set, but maybe they tend to take the first item. Low cost to shuffle, I guess.

No, sadly, most clients this pick the first.

The real issue is whether there's actually going to be another kube-dns release or not.

Doesn't hurt to open issue here: https://github.com/coredns/deployment/issues and ask for the loadbalance plugin to be enabled. /Miek

…

-- Miek Gieben

[thockin]

I see zero reason not to do this for CoreDNS (unless it has some cost, but it seems a random shuffle is fine). We need to decide if there will be another kube-dns release ever, other than security...

…

On Fri, Jul 6, 2018 at 9:05 AM Miek Gieben ***@***.***> wrote: [ Quoting ***@***.***> in "Re: [kubernetes/dns] enable skyDNS ..." ] >Shuffling seems OK. I assumed that clients would randomly pick from a >response set, but maybe they tend to take the first item. Low cost to >shuffle, I guess. No, sadly, most clients this pick the first. >The real issue is whether there's actually going to be another kube-dns >release or not. Doesn't hurt to open issue here: https://github.com/coredns/deployment/issues and ask for the loadbalance plugin to be enabled. /Miek -- Miek Gieben — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#167 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFVgVHuNs9vWjQ1avcJApgBYIQtbitLwks5uD4qsgaJpZM4QkJkM> .

[wu8685]

I have used a forked kube-dns with the RR enabled in my k8s environment.
It is ok to consider this PR till there is a release.

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

[fejta-bot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

[k8s-ci-robot]

@fejta-bot: Closing this PR.

In response to this:

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

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.

[bboreham]

Came across this by accident and puzzled to see no mention of the "happy eyeballs" algorithm which defeats round-robin via DNS.

https://daniel.haxx.se/blog/2012/01/03/getaddrinfo-with-round-robin-dns-and-happy-eyeballs/