aws/amazon-vpc-cni-k8s

Setting MTU on ipv4 egress on a ipv6 only cluster

umutuzgur opened this issue · 4 comments

umutuzgur commented

What happened:

We are trying to change the MTU values on all interfaces on our ipv6 only cluster with ipv4 egress. When we set the AWS_VPC_ENI_MTU value, this only affects the eth0 interface but not the v4if0 interface where ipv4 egress is supported. Which values do we need to set so the MTU values are persisted on both interfaces?

Environment:

  • Kubernetes version (use kubectl version): v1.30.0-eks-036c24b
  • CNI Version: v1.18.1-eksbuild.3
  • OS (e.g: cat /etc/os-release): Amazon Linux 2
  • Kernel (e.g. uname -a): 5.10.213-201.855.amzn2.x86_64
umutuzgur commented

This might become a feature request or a bug depending on how it is interpreted but we found the culprit to be this here https://github.com/aws/amazon-vpc-cni-k8s/blob/master/misc/10-aws.conflist#L18. When that line is turned to __MTU__, everything works as expected

  • 👍1
orsenthil commented

This may be a bug. The environment variable AWS_VPC_ENI_MTU is picked up and assigned to MTU
Since the value was set to 9001, I assume this wasn't getting set for egress-cni property for IPV4 Egress on IPV6 Cluster.

benben commented

@orsenthil thanks for accepting this as a bug. I sent the change on how we fixed it with more context in #2951

Please let me know in case you need anything else. Thanks!

github-actions commented

This issue is now closed. Comments on closed issues are hard for our team to see.
If you need more assistance, please either tag a team member or open a new issue that references this one.