fixes data race in node controller

Etcd listens on localhost for GCE

Remove loopback ip from k8petstore PublicIPs

Redo on building platforms in parallel

Retry replication controller rolling updates on version mismatch.

Using gcr.io/google_containers/etcd:2.0.8

Switch network tester image to gcr.io

Add s/hostDir/hostPath/

AWS: Create instances in the specified AZ

Add latency metrics to REST client

Add ability to set master/minion size for GCE using envvar.

Added rate limiting to pod deleting in NodeController.

Added PullPolicy
Added the change from labels to label-selector

Fixes #6228

Unbreak minion watch on kubelet

Enable log collection from master.

a function (to make it kinder on reviewers for the next piece).

Revert "Revert "Parallelize architectures in both the building and packaging phases of `make release`""

This reverts commit 9f60dde3.

When kubectl does rolling updates of replication controllers, retry updates that
fail due to version mismatches (caused by concurrent updates by other clients).
These failed rolling updates were causing intermittent e2e test failures
(e.g. issue 5821)

First cut at v1beta3 FAQ and versioning rationale.

Make secrets at cluster startup.

Getting started guide for Mesos on Google Cloud Platform

Build etcd image (version: 2.0.8) and push it to gcr.io

Convert SIGTERM into SIGINT in e2e test runner

Add an nginx docker image for use on the master.

Adjust defaults and reporting for serve_hostnames