-
Notifications
You must be signed in to change notification settings - Fork 4.7k
New issue
Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? # to your account
Canal on K8s v1.8 with chain append mode no longer respects networkpolicies #4037
Comments
@KashifSaadat just to confirm, this is only an issue when the |
Yes I believe so, because the flannel rules are then at the bottom of the |
See my comment here. Essentially, Calico is behaving as expected for the given config. There's probably room for a change in flannel to play nicer in this situation, but if possible I'd recommend using the default |
Automatic merge from submit-queue. Downgrade Flannel in Canal deployment to v0.9.0 Flannel v0.9.1 introduces a single change to add 2 iptables rules to the `FORWARD` chain, permitting traffic in/out of the pod network (introduced to improve compatibility with newer versions of Docker). This change is unnecessary for Canal deployments for the following reasons: - Calico's `DefaultEndpointToHostAction` is set to `ACCEPT` in the manifest deployed by kops, allowing traffic by default once all other Calico rules are processed. - If Calico's `ChainInsertMode` is set to `APPEND`, the flannel rules will be processed before the Calico rules, accepting traffic by default, and so Kubernetes network policies will not take effect This change is temporary until a more permanent resolution is available with Flannel, such as providing a configurable option to disable the addition of these rules when deployed with Calico. Related to #4037
@KashifSaadat Could you raise an issue against flannel to make this configurable |
@tomdee sure, I've raised the following issue to track this: flannel-io/flannel#938 |
Issues go stale after 90d of inactivity. If this issue is safe to close now please do so with Send feedback to sig-testing, kubernetes/test-infra and/or fejta. |
/remove-lifecycle stale I've raised this PR to address the issue within flannel: flannel-io/flannel#978 |
The information around this issue is documented here: projectcalico/canal#115
Depending on the outcome from the above issue, we will likely need updates to the following files:
Versions
Cluster Spec (networking block)
CC @chrislovecnm
The text was updated successfully, but these errors were encountered: