Harness CI - Build and Push to ECR - EOF error

Harness Support
1

Hi,

I’m using the Delegate in an EKS cluster, and the AWS Connector is configured using IRSA.

The pipeline is pulling the images from private repositories from ECR, but when I try to push the application image using the “Build and Push to ECR,” I get the following error:

+ /kaniko/executor --dockerfile=/harness/docker/Dockerfile --context=dir:///harness/ --destination=<aws-account-id>.dkr.ecr.sa-east-1.amazonaws.com/hermod:latest --snapshotMode=redo --digest-file=/kaniko/digest-file
error checking push permissions -- make sure you entered the correct tag name, and that you are authenticated correctly, and try again: checking push permission for "<aws-account-id>.dkr.ecr.sa-east-1.amazonaws.com/hermod:latest": Post "https://<aws-account-id>.dkr.ecr.sa-east-1.amazonaws.com/v2/hermod/blobs/uploads/": EOF
exit status 1

I tested the role configured on the service account, and I was able to push an image to the repository. So I don’t think it is a matter of permission.

Looking at the logs, I’ve found the following error message:

commands: "/kaniko/kaniko-ecr"
error: "failed to read artifact file: /addon/tmp/.plugin/artifact: open /addon/tmp/.plugin/artifact: no such file or directory"
errorVerbose: "open /addon/tmp/.plugin/artifact: no such file or directory\\nfailed to read artifact file: /addon/tmp/.plugin/artifact\\ngithub.com/harness/harness-core/product/ci/addon/artifact.ParseArtifactFromFile\\n\\tproduct/ci/addon/artifact/artifact.go:73\\ngithub.com/harness/harness-core/product/ci/addon/artifact.GetArtifactProtoFromFile\\n\\tproduct/ci/addon/artifact/artifact.go:57\\ngithub.com/harness/harness-core/product/ci/addon/tasks.(*pluginTask).execute\\n\\tproduct/ci/addon/tasks/plugin.go:188\\ngithub.com/harness/harness-core/product/ci/addon/tasks.(*pluginTask).Run\\n\\tproduct/ci/addon/tasks/plugin.go:111\\ngithub.com/harness/harness-core/product/ci/addon/grpc.(*handler).ExecuteStep\\n\\tproduct/ci/addon/grpc/handler.go:91\\ngithub.com/harness/harness-core/product/ci/addon/proto._Addon_ExecuteStep_Handler\\n\\tbazel-out/k8-fastbuild/bin/product/ci/addon/proto/ciaddonpb_go_proto_/github.com/harness/harness-core/product/ci/addon/proto/addon.pb.go:490\\ngoogle.golang.org/grpc.(*Server).processUnaryRPC\\n\\texternal/org_golang_google_grpc/server.go:1282\\ngoogle.golang.org/grpc.(*Server).handleStream\\n\\texternal/org_golang_google_grpc/server.go:1616\\ngoogle.golang.org/grpc.(*Server).serveStreams.func1.2\\n\\texternal/org_golang_google_grpc/server.go:921\\nruntime.goexit\\n\\tsrc/runtime/asm_amd64.s:1571\"

Does someone knows what is happening and how to debug the Kaniko container?

2

I’ve found the solution.
I had to configure the service account in the advanced infrastructure configuration:

          infrastructure:
            type: KubernetesDirect
            spec:
              connectorRef: test-connector
              namespace: harness-delegate-ng
              serviceAccountName: **harness-delegate-sa**
              automountServiceAccountToken: true
              nodeSelector: {}
              os: Linux

And another thing I had to do was create a Run step before the “Build and Push to ECR” step to run the command:

aws ecr get-login-password --region <aws_region> | docker login --username AWS --password-stdin <aws_account_id>.dkr.ecr.<aws_region>.amazonaws.com

Now the pipeline is working.

1 Like
3

hey @gleise.teixeira , that’s great! Do let us know if there are any other issues you are facing?Thanks

4

Hey @gleise.teixeira
Can you share the screenshot of the pipeline you created?