Does Google Container Registry undergo issues?

Question

I am facing an issue while using container optimized VMs. I launch an instance with following command:

gcloud compute instances create "$instance_name" \
    --tags "http-server" \
    --image container-vm \
    --scopes storage-rw,logging-write \
    --metadata-from-file google-container-manifest="m2.yml" \
    --zone "$my_zone" \
    --machine-type "$my_machine_type"

where m2.yml is:

version: v1beta2
containers:
  - name: nginx
    image: nginx

When I ssh to the instance and look into /var/log/docker.log I see:

time="2015-06-17T07:42:59Z" level=error msg="Handler for GET /containers/{name:.*}/json returned error: no such id: kubelet" 
time="2015-06-17T07:42:59Z" level=error msg="HTTP Error: statusCode=404 no such id: kubelet" 
time="2015-06-17T07:42:59Z" level=info msg="GET /version" 
time="2015-06-17T07:42:59Z" level=info msg="+job version()" 
time="2015-06-17T07:42:59Z" level=info msg="-job version() = OK (0)" 
time="2015-06-17T07:42:59Z" level=info msg="GET /containers/docker-daemon/json" 
time="2015-06-17T07:42:59Z" level=info msg="+job container_inspect(docker-daemon)" 
no such id: docker-daemon
time="2015-06-17T07:42:59Z" level=info msg="-job container_inspect(docker-daemon) = ERR (1)" 
time="2015-06-17T07:42:59Z" level=error msg="Handler for GET /containers/{name:.*}/json returned error: no such id: docker-daemon" 
time="2015-06-17T07:42:59Z" level=error msg="HTTP Error: statusCode=404 no such id: docker-daemon" 

There is no containers running on the instance and docker images -a says:

REPOSITORY                       TAG                 IMAGE ID            CREATED              VIRTUAL SIZE
gcr.io/google_containers/pause   0.8.0               2c40b0526b63        11 weeks ago         241.7 kB
<none>                           <none>              56ba5533a2db        11 weeks ago         241.7 kB
<none>                           <none>              511136ea3c5a        2.009460 years ago   0 B

Does the container engine undergo some issues or I am doing something wrong?

Update1

I gave an example a try:

version: v1 kind: Pod spec: containers: - name: simple-echo image: gcr.io/google_containers/busybox command: ['nc', '-p', '8080', '-l', '-l', '-e', 'echo', 'hello world!'] imagePullPolicy: Always ports: - containerPort: 8080 hostPort: 8080 protocol: TCP restartPolicy: Always dnsPolicy: Default

and it is still error in the log:

evgeny@instance:~$ cat /var/log/docker.log | grep error
time="2015-06-18T16:28:56Z" level=error msg="Handler for GET /containers/{name:.*}/json returned error: no such id: kubelet" 
time="2015-06-18T16:28:56Z" level=error msg="HTTP Error: statusCode=404 no such id: kubelet" 
time="2015-06-18T16:28:56Z" level=error msg="Handler for GET /containers/{name:.*}/json returned error: no such id: docker-daemon" 
time="2015-06-18T16:28:56Z" level=error msg="HTTP Error: statusCode=404 no such id: docker-daemon" 
time="2015-06-18T16:28:56Z" level=error msg="Handler for GET /images/{name:.*}/json returned error: No such image: gcr.io/google_containers/busybox" 
time="2015-06-18T16:28:56Z" level=error msg="HTTP Error: statusCode=404 No such image: gcr.io/google_containers/busybox" 
time="2015-06-18T16:28:56Z" level=error msg="Handler for GET /containers/{name:.*}/json returned error: no such id: docker" 
time="2015-06-18T16:28:56Z" level=error msg="HTTP Error: statusCode=404 no such id: docker" 
time="2015-06-18T16:28:56Z" level=error msg="Handler for GET /images/{name:.*}/json returned error: No such image: gcr.io/google_containers/busybox" 
time="2015-06-18T16:28:56Z" level=error msg="HTTP Error: statusCode=404 No such image: gcr.io/google_containers/busybox" 

I was about to send a feedback via feedback form but Submit fails with: relay__en.js?authuser=1:182 Uncaught TypeError: Cannot set property 'vmFeedbackData' of undefined. Such a story.

Answer 1

The latest container vm image doesn't support the v1beta2 kubernetes API. You will need to update your manifest to use v1beta3 or v1 (with the corresponding yaml changes).

The latest version of the container vm documentation shows a yaml example using the v1 API.

Answer 2

I ran into a similar problem and the solution turned out to be that I was trying to use

securityContext:
  privileged: true

Kubernetes doesn't honor that property when loading a docker image onto a GCE instance without using a Google Container cluster (i.e. using a google-container-manifest metadata item on a compute.v1.instance or a compute.v1.instanceTemplate), so it just never loads the image properly onto the instance.

Removing that property solved my problem.