-
Notifications
You must be signed in to change notification settings - Fork 748
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
Calico updates to v3.0, add auto-scaler, add typha #34
Conversation
@tmjd |
The testing I did was to bring up a small 3 node cluster (1 master, 2 nodes), deploy with the manifest and then ran through the Simple Policy Demo which deploys pods and a service then uses another pod to test that network policy will block traffic and then that traffic can be allowed. I also verified that the traffic worked across hosts. |
I am getting following error:
|
I think it's because you're running an older version of Kubernetes. @tmjd we may need to use the v1beta1 apiVersion in order to support earlier k8s versions. |
@liwenwu-amazon I've pushed changes so they all use the v1beta1 apiVersion, I've verified it still works with my v1.9.3 cluster. |
@tmjd thanks Eric. I no longer see any errors when applying aws-k8s-cni-calico.yaml . But I can NOT get simple policy demo working. Is it due to the fact my cluster is 1.7.4? thanks |
Off the top of my head I think that should work. Which step specifically is not working? If it is the isolation that is not working try using the following from a previous version of the demo (notice it doesn't have
If the part that is not working is |
Issue-1: With version 1.7.4, I am NOT seeing Isolation policy being enfoced. Issue-2: With version 1.93, "allow Access using NetworkPolicy" ONLY affects the new Pods created after the Network Policy configuration. It does NOT work for the existing running Pods. |
For clarification, do you mean pods created before applying the policy? Or pods created before installing Calico? |
Pods created before applying the policy. |
I just deployed a K8s v1.7.4 cluster and the Simple Policy demo (the Calico v3.0 version) worked as expected. (Though my kubectl client was v1.7.11.) When you had that deployed did you take a look at the logs from calico-node? My instinct is that calico-node was unable to connect to the kubernetes api or policy/pod configuration wasn't correct like they were in different namespaces, is there anyway it could have been that? Issue-2: That's new behavior. Did you have to re-create all pods before the policy worked, both the nginx and access pods or just the access pod? |
I've rebased the changes on latest master. |
I am seeing following on my 1.9 cluster, it looks like the policy is NOT working.
if I delete all policy, i can see it works again
Here is kubectl and cluster version
|
This problem sounds like what you would see if you were having an issue like #35. There are a couple things you can check:
|
@tmjd @caseydavenport , my setup is different than issue #35, where VPC ONLY have 1 CIDR.
Here is the iptable dump after enable NetworkPolicy
|
Rebased onto master |
A prior installation is not upgradable from the previous version. It would be necessary to first update Calico to v2.6.6+ and then upgrade to this version should be possible.