You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We are building the K8S with CNI, and we are not using the EKS node AMI, it is build from centOS AMI.
After everything is setup, pod is allocated with the IP associate with the ENI, however, pod is not able to ping another pod.
Take a long time for investigation, and it turns out it is caused by FORWARD policy is DROP by default in iptables, after update it as ACCEPT, everything goes well. In EKS AMI, the policy is update to ACCEPT already.
We are building the K8S with CNI, and we are not using the EKS node AMI, it is build from centOS AMI.
![Screen Shot 2019-09-18 at 12 52 33 PM](https://user-images.githubusercontent.com/17717672/65133155-bc129600-da34-11e9-81fb-9abc1719df18.png)
After everything is setup, pod is allocated with the IP associate with the ENI, however, pod is not able to ping another pod.
Take a long time for investigation, and it turns out it is caused by FORWARD policy is DROP by default in iptables, after update it as ACCEPT, everything goes well. In EKS AMI, the policy is update to ACCEPT already.
I think everyone try to setup the CNI without EKS AMI may meet the issue, and it is not easy for the trouble shooting, would be helpful if add issues like this to the trouble shooting documents.
https://github.com/aws/amazon-vpc-cni-k8s/blob/master/docs/troubleshooting.md
The text was updated successfully, but these errors were encountered: