-
Notifications
You must be signed in to change notification settings - Fork 1.6k
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
kindnetd remove wrong routes #2941
kindnetd remove wrong routes #2941
Conversation
i was not aware capi uses kindnet for e2e. in any case, unclear if this is desired. /cc @aojea |
81cede9
to
3d1376f
Compare
3d1376f
to
ab13987
Compare
Thank you very much for your review and guidance @aojea 👍 I also inlined the func again because as you outlined correctly, we can't just do an early return. |
ab13987
to
5f0e3c3
Compare
/lgtm because we have to build the image to run it in the CI, this code is not being exercised right now on the CI @BenTheElder can you PTAL |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
kind is generally -1 on adding code paths to support things we don't support (a tautology if you will!) but this change is small and seems reasonable and useful for multiple projects in-org.
+1 to antonio's review thus far :-)
/approve
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: BenTheElder, chrischdi 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 |
/test pull-kind-e2e-kubernetes-1-23 |
/assign |
5f0e3c3
to
eb00dc0
Compare
OK, all of the prow CI jobs (not github actions) should run with this now. |
/lgtm |
This adds logic to kindnetd to remove routes which do not match the expected gateway address for an (Pod)CIDR.
I'm currently working at
Cluster API uses kindnetd as CNI for lots of e2e tests.
As of today, kindnetd only adds new routes and does never delete orphaned routes.
This could lead to the following situation in the workload Cluster created by Cluster API when kindnetd is used as CNI:
Broken pod to pod communication may include kube-apiserver to webhook communication, which is the case I detected during my tests.