Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

🌱 Add dualstack e2e test #8517

Merged
merged 1 commit into from
May 16, 2023

Conversation

killianmuldoon
Copy link
Contributor

@killianmuldoon killianmuldoon commented Apr 12, 2023

Adds a dualstack e2e quick start test for CAPI

/area testing

@k8s-ci-robot k8s-ci-robot added area/testing Issues or PRs related to testing cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. labels Apr 12, 2023
@k8s-ci-robot k8s-ci-robot added the size/M Denotes a PR that changes 30-99 lines, ignoring generated files. label Apr 12, 2023
@killianmuldoon killianmuldoon changed the title 🌱 Add dualstack e2e test [WIP] 🌱 Add dualstack e2e test Apr 12, 2023
@k8s-ci-robot k8s-ci-robot added the do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. label Apr 12, 2023
Copy link
Contributor Author

@killianmuldoon killianmuldoon left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Q: should this be implemented with ClusterClass? Should the ipv6 test be on ClusterClass?

@killianmuldoon killianmuldoon force-pushed the pr-e2e-dual-stack branch 4 times, most recently from 80385dd to d32c5c8 Compare April 12, 2023 19:39
@killianmuldoon killianmuldoon changed the title [WIP] 🌱 Add dualstack e2e test 🌱 Add dualstack e2e test Apr 12, 2023
@k8s-ci-robot k8s-ci-robot removed the do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. label Apr 12, 2023
initConfiguration:
nodeRegistration:
kubeletExtraArgs:
node-ip: "0.0.0.0" # Note: The control plane node should use an ipv6 address as an internalIP. The worker node should use an ipv4 address as an internalIP.
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Kubernetes (kubeadm) doesn't allow setting this field as unspecified - i.e. 0.0.0.0 / :: - for dualstack. The compromise here is to mix the nodes between ipv4 and ipv6.

Copy link
Contributor

@lubronzhan lubronzhan Apr 13, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Be careful here kubernetes/kubernetes#116305
There is problem with Dual stack node-ip with external cloud-provider, kubernetes/enhancements#3705
It's fixed in k8s 1.27, added here
Before that, node will only show single IP stack IP with external cloud-provider. Please correct me if I'm wrong @christianang

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If I understand the issue correctly - setting this config on Kubelet means that the node object will only ever advertise one node address. Is that roughly right @lubronzhan?

I think, if we need to have explicitly dualstack nodes here, I can remove this flag and patch the node in the docker provider similar to what a cloud provider would do.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi @killianmuldoon the issue is only with external cloud-provider. if it's baremetal it's fine. Will you set cloud-provider=external on kubelet when using docker provider? If not setting to external it's probably ok

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah I think you both have it covered. As Lubron said it probably depends on if you set cloud-provider to external because in the end both affect the node-ips that are set in the node status.

Copy link
Member

@chrischdi chrischdi left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

/lgtm

Just wondering if there's something so we could extend the QuickStartSpec to check if the nodes actually have the dualstack / ips (something like via kubectl get nodes (of course via client) or so 🤔

@k8s-ci-robot k8s-ci-robot added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Apr 13, 2023
@k8s-ci-robot
Copy link
Contributor

LGTM label has been added.

Git tree hash: 1d315f6a3fe0d95b713d536a6a64bbebc0f6df75

@killianmuldoon
Copy link
Contributor Author

Just wondering if there's something so we could extend the QuickStartSpec to check if the nodes actually have the dualstack / ips (something like via kubectl get nodes (of course via client) or so thinking

Probably a good idea - we could us a hook like we have in the upgrade test to run this check. Unfortunately though this test doesn't give us dual-stack node IPs, at least not in its current form. It results in mixed IPs - some 6, some 4.

Copy link
Member

@sbueringer sbueringer left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Overall looks good.

I'm a bit confused why this seems to work while we needed a separate job where we run kind with IPv6 to get the IPv6-only stuff working. But if it just works that's fine, although we should have a follow-up to check if we still need a separate job for IPv6 and if we do figure out why IPv6 only doesn't work while DualStack works.

But maybe it's just that things changed either in kind or in our test framework and the problems we had in the past have been resolved

@k8s-ci-robot k8s-ci-robot removed the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Apr 13, 2023
@christianang
Copy link
Contributor

You could deploy dual stack with different “primary” ip families, where the CIDRs and listen addresses, node-ip, etc are either IPv6 or IPv4. It could be worth testing both families as the “primary” ip family to get additional coverage on if a field can work with both IPv6 or IPv4.

Related, it is interesting that you configure the control plane as IPv6 family and the worker as IPv4, which I assume is to get similar coverage as what I’m suggesting above. Probably fine to do and probably gets some of the coverage I’m suggesting above.

@k8s-ci-robot k8s-ci-robot added size/XL Denotes a PR that changes 500-999 lines, ignoring generated files. and removed size/M Denotes a PR that changes 30-99 lines, ignoring generated files. labels May 4, 2023
@killianmuldoon killianmuldoon force-pushed the pr-e2e-dual-stack branch 2 times, most recently from 2fb6ede to cdf9a42 Compare May 4, 2023 19:20
@killianmuldoon killianmuldoon force-pushed the pr-e2e-dual-stack branch 5 times, most recently from 9268327 to 7be3410 Compare May 14, 2023 21:26
@sbueringer
Copy link
Member

sbueringer commented May 15, 2023

@killianmuldoon I did another retry and it's green now. What is the current state of this PR?
(+/- I should do another review)

@killianmuldoon
Copy link
Contributor Author

I've got a couple of fixups pending, should be ready for another round later today.

@killianmuldoon killianmuldoon force-pushed the pr-e2e-dual-stack branch 2 times, most recently from b3bc55a to 3325f45 Compare May 15, 2023 12:08
@killianmuldoon
Copy link
Contributor Author

/retest

@killianmuldoon killianmuldoon force-pushed the pr-e2e-dual-stack branch 3 times, most recently from eb9627d to 26a9598 Compare May 15, 2023 12:45
test/e2e/config/docker.yaml Show resolved Hide resolved
test/e2e/data/kubetest/conformance.yaml Show resolved Hide resolved
test/e2e/quick_start_test.go Outdated Show resolved Hide resolved
test/infrastructure/docker/internal/docker/machine.go Outdated Show resolved Hide resolved
test/infrastructure/docker/internal/docker/machine.go Outdated Show resolved Hide resolved
test/infrastructure/docker/internal/docker/util.go Outdated Show resolved Hide resolved
test/infrastructure/docker/internal/docker/util.go Outdated Show resolved Hide resolved
Signed-off-by: killianmuldoon <kmuldoon@vmware.com>
@sbueringer
Copy link
Member

Thank you!

/lgtm

/assign @fabriziopandini

@k8s-ci-robot k8s-ci-robot added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label May 15, 2023
@k8s-ci-robot
Copy link
Contributor

LGTM label has been added.

Git tree hash: 2795c6f201d9f2b5d7b0ad07c3ccadfe8fd2c106

Copy link
Member

@fabriziopandini fabriziopandini left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

great work!
/lgtm
/approve

@@ -70,7 +70,15 @@ func WithDockerSockMount() KindClusterOption {
// the new kind cluster.
func WithIPv6Family() KindClusterOption {
return kindClusterOptionAdapter(func(k *KindClusterProvider) {
k.ipFamily = clusterv1.IPv6IPFamily
k.ipFamily = clusterv1.DualStackIPFamily
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can we add a note why are we setting DuaStack for WithIPv6Family

@k8s-ci-robot
Copy link
Contributor

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: fabriziopandini

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:

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 k8s-ci-robot added the approved Indicates a PR has been approved by an approver from all required OWNERS files. label May 16, 2023
@k8s-ci-robot k8s-ci-robot merged commit ae6832c into kubernetes-sigs:main May 16, 2023
@k8s-ci-robot k8s-ci-robot added this to the v1.5 milestone May 16, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
approved Indicates a PR has been approved by an approver from all required OWNERS files. area/testing Issues or PRs related to testing cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. lgtm "Looks good to me", indicates that a PR is ready to be merged. size/L Denotes a PR that changes 100-499 lines, ignoring generated files.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

7 participants