Issue with Multus on EKS: Multiple Pods with Same IP Address

[kouamdo]

I am currently implementing the Multus solution on EKS and have encountered an issue that I am struggling to understand. Specifically, I am unsure how multiple pods are able to function with the same IP address.

To provide some context, I have followed the recommended steps 1 and 2 for setting up Multus as per the instructions in your git repository. However, I deviated slightly in that I installed the correct version of the Addons CoreDNS and also proceeded with the installation of the thick plugin instead of the thin plugin that was suggested.

Despite these steps, the issue with multiple pods having the same IP address persists. Additionally, I encountered a version error with “aws-cni” during the process. As a result, I was compelled to update the CoreDNS and install the thick plugin instead of following the suggested steps due to this error.

I am hoping to gain some insight into why this might be happening and how it can be resolved. Any guidance or suggestions would be greatly appreciated. Thank you in advance for your assistance.

there was the step :

Install Multus

• Let's go to Bastion host where we can run kubectl.

• Check CoreDNS addons to compatible version for your cluster :

  aws eks describe-addon-versions --addon-name coredns --kubernetes-version 1.29 --query "addons[].addonVersions[].[addonVersion, compatibilities[].defaultVersion]" --output text

• Update CoreDNS to the compatible version

  aws eks update-addon --cluster-name eks-multus-cluster --addon-name coredns --addon-version v1.11.1-eksbuild.4

• To validate that coredns add-on is running, ensure that both the coredns pods are in Running state

  kubectl get pods  -n kube-system | grep coredns

aws eks describe-addon \
    --cluster-name eks-multus-cluster \
    --addon-name coredns \
    --query "addon.addonVersion" \
    --output text

• Install multus CNI thick plugin(current version).

  kubectl apply -f  https://raw.githubusercontent.com/aws/amazon-vpc-cni-k8s/master/config/multus/v4.0.2-eksbuild.1/multus-daemonset-thick.yml

and there is the result when i created a lot of pod (8 pods ) :

[raghs-aws]

Hi, I believe that you are using multus with ipam as "host-local", in that case IP allocation is local to host and not cluster scoped. you should use different IPAM lik "whereabouts" . that ensures that per cluster multus ip are not duplicate. More details: https://github.com/k8snetworkplumbingwg/whereabouts

[kouamdo]

oh , ... i see , but one error occur

i have tried to join a good IAM role for whereabouts serviceaccount , but i'm trying to check again. Is there another way to fix it ?