Request: mirror defaultbackend image needed by nginx-ingress to registry.gitlab.com

Summary

This is not a bug, it's a special request to upload the defaultbackend nginx-ingress image that serves 404 pages for a very large client. This image is stored in k8s.gcr.io/defaultbackend which is not accessible from the client's internal network. The client only allow access through its proxied docker repository. Placing this image inside gitlab-org/** would allow the referred image to be pulled from the internal image repository.

Steps to reproduce

Deploy the chart: helm install very-large-deploy -f custom-values.yaml

Configuration used

This is the modified values.yaml file (excerpt)

nginx-ingress:
  controller:
    image: 
      repository: client.imagerepository.internal/kubernetes-ingress-controller/nginx-ingress-controller
      tag: "0.21.0"
    podLabels:
      sdr.appname: gitlab-poc
    priorityClassName: p3
  defaultBackend:
    image: 
      repository: client.imagerepository.internal/gitlab-org/some-directory/defaultbackend
      tag: "1.4"
    priorityClassName: p3

Current behavior

Image is being pulled from k8s.gcr.io that is not accessible from internal network

Expected behavior

To be able to pull the image from the allowed registry.gitlab.com/gitlab-org/** domain.

Versions

  • Chart: 4.7.1
  • Platform:
    • Self-hosted: Other
  • Kubernetes: (kubectl version)
    • Client Version: version.Info{Major:"1", Minor:"19", GitVersion:"v1.19.4"
    • Server Version: version.Info{Major:"1", Minor:"18", GitVersion:"v1.18.9"
  • Helm: v3.4.2

Relevant logs

Failed to pull image "k8s.gcr.io/defaultbackend:1.14": rpc error: code = 
 Unknown desc = failed to pull and unpack image "k8s.gcr.io/defaultbackend:1.14": failed to resolve reference "k8s.gcr.io/defaultbackend:1.14": failed to do request: Head "https://k8s.gcr.io/v2/defaultbackend/manifests/1.14": dial tcp [2607:f8b0:4023:c03::52]:443: connect: connection refused
Edited by Jason Plum