I suffer from this error when I deploy a pod.

• the image lies on google container registry within the same project as the cluster

• i can pull the image from the registry on my local computer

• i cannot pull the image if I ssh into the instance

From the docs it states that this should work out of the box. I checked and storage read access is indeed there.

Here's the config:

apiVersion: v1
kind: ReplicationController
metadata:
  name: luigi
spec:
  replicas: 1
selector:
  app: luigi
template:
  metadata:
    name: luigi
    labels:
      app: luigi
  spec:
    containers:
    - name: scheduler
      image: eu.gcr.io/bi/luigi/scheduler:latest
      command: ['/usr/src/app/run_scheduler.sh']
    - name: worker
      image: eu.gcr.io/bi/luigi/scheduler:latest
      command: ['/usr/src/app/run_worker.sh']

Describe gives me:

Failed to pull image "eu.gcr.io/bi/luigi/scheduler:latest": rpc error: code = Unknown desc = Error response from daemon: repository eu.gcr.io/bi/luigi/scheduler not found: does not exist or no pull access