dial tcp 127.0.0.1:8080: connect: connection refused AWS CodeBuild EKS

Question

I'm encountering following error message when I'm trying to deploy to EKS Cluster even I've already added CodeBuild IAM role to aws-auth.yaml like

- rolearn: arn:aws:iam::<AWS_ACCOUNT_ID>:role/codebuild-eks
  username: codebuild-eks
  groups:
    - system:masters

error: unable to recognize "deployment.yml": Get http://localhost:8080/api?timeout=32s: dial tcp 127.0.0.1:8080: connect: connection refused

And here is my codebuild command:

version: 0.2
phases:
  install:
    commands:
      - curl -o kubectl https://amazon-eks.s3-us-west-2.amazonaws.com/1.11.10/2019-06-21/bin/linux/amd64/kubectl
      - chmod +x ./kubectl
      - mkdir -p $HOME/bin && cp ./kubectl $HOME/bin/kubectl && export PATH=$HOME/bin:$PATH
      - kubectl version --short --client
      - curl -o aws-iam-authenticator https://amazon-eks.s3-us-west-2.amazonaws.com/1.13.7/2019-06-11/bin/linux/amd64/aws-iam-authenticator
      - chmod +x ./aws-iam-authenticator
      - mkdir -p $HOME/bin && cp ./aws-iam-authenticator $HOME/bin/aws-iam-authenticator && export PATH=$HOME/bin:$PATH
      - aws-iam-authenticator help
  pre_build:
    commands:
      - echo Entered the pre_build phase...
      - echo Logging in to Amazon EKS...
      - mkdir -p ~/.kube
      - aws s3 cp s3://ppshein-eks/config ~/.kube/config
      - export KUBECONFIG=$KUBECONFIG:~/.kube/config
      - aws eks --region $AWS_DEFAULT_REGION update-kubeconfig --name $AWS_CLUSTER_NAME
  build:
    commands:
      - echo Entered the build phase...
      - echo Change directory to secondary source
      - cd $CODEBUILD_SRC_DIR
      - echo List directory
      - ls -la
      - kubectl get pods --kubeconfig ~/.kube/config
      - kubectl apply -f deployment.yml

Problem is when CodeBuild run this kubectl apply -f deployment.yml statement, I've got error message but above one kubectl get pods --kubeconfig ~/.kube/config is working fine.

Please let me know which area I've missed to add or configure. Thanks.

I'm no CodeBuild expert, but it looks like those phases are independent of one another; thus, you'll need - export KUBECONFIG=$KUBECONFIG:~/.kube/config in your build: commands: also, or do the right thing and set that as a variable for your whole pipeline, if CodeBuild is smart enough to allow that — mdaniel

shariqmaws shariqmaws · Accepted Answer · 2020-08-04T09:11:40

These errors indicates that kubectl was unable to reach the kubernetes server endpoint at 127.0.0.1:8080, or the local host. Since you have configured kubeconfig with the command update-kubeconfig, it seems multiple configs are being merged 1 due to this command:

- export KUBECONFIG=$KUBECONFIG:~/.kube/config

To see the resultant config that kubectl sees, run this command before the failing command:

- kubectl config view                      # Add this
- kubectl apply -f deployment.yml

To fix, I recommend to change as follows in pre_build phase:

- export KUBECONFIG=~/.kube/config

Or, use '--context' flag with kubectl to select the correct context.

- export KUBECONFIG=file1:file2
- kubectl get pods --context=cluster-1
- kubectl get pods --context=cluster-2

dial tcp 127.0.0.1:8080: connect: connection refused AWS CodeBuild EKS

1 Answers