Add rule for docker image pull error

[ajitak]

Add a new rule which catches docker image corruption. It will address kubernetes/kubernetes#47219

---

This change is 

[ajitak]

/assign @Random-Liu
/assign @dchen1107

[Random-Liu]

If this problem is permanent, you may want to add a condition for it.

[ajitak]

Changed it to temporary.

[Random-Liu]

This seems like a kubelet log. Are your should this is docker log?

[ajitak]

Looks like I had published outdated changes. Fixed it.

[dchen1107]

ping ...

Thanks!

[ajitak]

@dchen1107: waiting for @Random-Liu to review.

[Random-Liu]

@ajitak Have you tried the rule for journald? Does it work?\

I guess it will not.

[ajitak]

I have not been able inject messages in journald. I tested regex with docker log string to test the pattern.

[marun]

@Random-Liu @ajitak: What's required to move this forward, or can it be moved out of 1.7?

[ajitak]

@Random-Liu Updated the regex and tested it.

When I run kubectl describe node node-name command, I see following under events:
3m 4s 2 {docker-monitor gke-demo-cluster-default-pool-fdda9125-7gr3} Warning CorruptDockerImage time="2017-06-21T20:25:42.481887487Z" level=error msg="Error trying v2 registry: failed to register layer: rename /var/lib/docker/image/aufs/layerdb/tmp/layer-632022140 /var/lib/docker/image/aufs/layerdb/sha256/011b303988d241a4ae28a6b82b0d8262751ef02910f0ae2265cb637504b72e36: directory not empty"