Issue 14385: job's parallelism defaults to completions.

don't fatal on missing sorting flag

printf->print in hack/verify-flags-underscore.py script

Adding an example deployment yaml

Job update

Reaper and scaler for jobs

Default resources.requests only for Pod resources

allow privileged pods

Fix bug with empty subsets

Disable race detection for all integration tests

This reverts commit eb9cb8fb.

Temporary workaround of problem with not ready endpoints.

E2E tests for horizontal pod autoscaler.

Fixes & tuning in horiontal pod autoscaler and its e2e tests; two of the tests added to "Autoscaling suite".

MESOS: Fix zero MaxOpenFiles in Mesos executor

Properly use framework in load test

Add an option to modify timeout for namespace duration in e2e Framework

Increase some more timeouts to eliminate flakes

Turn off race detection in the etcd tools test as etcd client isn't thread safe.

Move broken service tests to the flaky suite to make GCE suite happy again

Add docs for ubuntu kubernetes cluster upgrading

Fix indendation on devel/coding-conventions.md

Correct comments in mount.go

Fix document link

change etcdIndec to etcdIndex

Bump Container Inspect to loglevel 4

fix the parameter name of k8s version

Use the cluster name instead of the minion tag as the prefix for the firewall rules created in gke e2e tests