k8s-dns-node-cache doesn't support iptables-nft

[champtar]

Right now k8s-dns-node-cache only support iptables-legacy
It would be nice to select iptables-legacy / iptables-nft automatically
see kubernetes/kubernetes#82966 for how kubeproxy does it

[caruccio]

@bowei Is it possible to give some attention to this issue? This is a show stopper for centos 8.

[champtar]

I rebuilt with

BASEIMAGE?=k8s.gcr.io/debian-iptables-$(ARCH):v12.0.1

And tested on CentOS8 / nft it works fine
Just need to fight with the Makefile a bit and I'll try to send a PR

[caruccio]

Thanks! It's working now.

[champtar]

@caruccio you rebuilt the image yourself ? because I don't see a new release with #355 in it

[caruccio]

@chadswen I was able to run it with gcr.io/google-containers/kube-proxy:v1.16.9-beta.0

[caruccio]

@champtar where are you looking at for the builds? Is this a public read-only CI?

[champtar]

I was just looking at the tags:
https://github.com/kubernetes/dns/releases

[caruccio]

OH ... maybe google is building images from master?

[champtar]

You are talking about kube-proxy, this is k8s-dns-node-cache*
https://console.cloud.google.com/gcr/images/google-containers/GLOBAL/k8s-dns-node-cache-amd64?gcrImageListsize=30

[caruccio]

Even so, this is the same issue in kube-proxy. I will wait for this fix to be release, until then, kube-dns-node-cache is disabled for me. Thanks and sorry for the mess :)

[champtar]

For kube-proxy this is fixed in 1.17+

[caruccio]

For whoever makes the same mistake as I did, the PR for kube-proxy is this kubernetes/kubernetes#82966

[champtar]

#355 was reverted, this will be fixed by #367

[justaugustus]

/reopen
(until #367 merges)

[k8s-ci-robot]

@justaugustus: Reopened this issue.

In response to this:

/reopen
(until #367 merges)

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.

[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

[bowei]

/remove-lifecycle stale