cni plugin not initialized on nodes created by Karpenter on EKS cluster with VPC-CNI add on

0

I've a EKS cluster with node groups, when I sent request it get processed without any error. Some times based on demand Karpenter provisions a new eks node dynamically and processes a request , on karpenter eks node in logs it showing error message - "Error syncing pod, skipping" err="network is not ready: container runtime network not ready: NetworkReady=false reason:NetworkPluginNotReady message:Network plugin returns error: cni plugin not initialized" .

Already cluster is aws managed cluster and it has vpc-cni add on . Karpenter is installed on same cluster hence it should not show the cni plugin error. Any suggestion much appreciated .

Srujan
已提問 6 個月前檢視次數 1206 次
1 個回答
0

The error message you're encountering points to a problem with the Container Network Interface (CNI) plugin on the dynamically provisioned EKS node by Karpenter. This indicates that the network plugin necessary for container networking isn't initialized or ready.

To troubleshoot and resolve this issue, consider the following steps. First, ensure that there is a aws-node pod corresponding to each worker node, and that the pods are in a "Running" state. If any pods are in a different state or are missing, it could indicate an issue with the AWS VPC CNI addon on those nodes. This can provide insights into what might be causing the network initialization problem.

Additionally, ensure compatibility between the version of the VPC CNI addon and your EKS cluster version. AWS regularly updates services, and a newer EKS version may require an updated VPC CNI addon. Refer to AWS documentation for supported versions [1], and if necessary, update the VPC CNI addon to a version that aligns with your EKS cluster.

Review the configuration of Karpenter to ensure it aligns with your EKS cluster. Verify that Karpenter is configured with the correct networking settings and is aware of the VPC CNI. Karpenter utilizes a configuration file, typically named karpenter.yaml, to specify various settings, including cluster configuration. Confirm the presence and correctness of this configuration file.

By systematically examining CNI plugin logs, verifying VPC CNI addon compatibility, and reviewing Karpenter configuration, you can identify and address the issues contributing to the network plugin initialization problem on your dynamically provisioned EKS node.

Reference: [1] https://docs.aws.amazon.com/eks/latest/userguide/managing-vpc-cni.html

AWS
已回答 5 個月前
  • Thank you for response. Yes I've verified that my node is in ready state and pod is running and application does not have any impact.

    Already verified that vpc CNI addon compatible with cluster.

    I'm using Karpenter helm chat to place it pods on cluster, both Karpenter pods are running and Karpenter provisions a node. How to Verify that Karpenter with correct networking settings and it awre of VPC CNI? Any specific configuration need to add to Karpenter ?

您尚未登入。 登入 去張貼答案。

一個好的回答可以清楚地回答問題並提供建設性的意見回饋,同時有助於提問者的專業成長。

回答問題指南