Currently, whenever there is any update, kubelet would force all pod workers to
sync again, causing resource contention and hence performance degradation.

This commit flips kubelet to use incremental updates (as opposed to snapshots).
This allows us to know what pods have changed and send updates to those pod
workers only. The `SyncPods` function has been replaced with individual
handlers, each handling an operation (ADD, REMOVE, UPDATE). Pod workers are
still triggered periodically, and kubelet performs periodic cleanup as well.

This commit also spawns a new goroutine solely responsible for killing pods.
This is necessary because pod killing could hold up the sync loop for
indefinitely long amount of time now user can define the graceful termination
period in the container spec.

Support default types in the resource builder

Changed Resource Consumer for correct parsing POST requests

Cleanup deprecated Forever function

Actually hold NodePorts open in kube-proxy

Copy edits for typos (resubmitted)

Make --validate default on and shows how to turn if off

to add tests for RC selector overlapping and fix few comments

Refactoring of create etcd tests.

Right now there is no method in the resource builder for specifying just the name of a resource. NameParam is useful when a default type is already specified with ResourceTypes.
Signed-off-by: kargakis <mkargaki@redhat.com>

WIP/MESOS: Handle Mesos TASK_ERROR status

Make verify-flags-underscore.py return error on failure

plugin/oidc: add minor documentation details.

Revert "Reenable Cacher"

We were running it in shippable and travis, but since it didn't return
an error it wasn't actually catching things.

Scale subresource and Heapster querying in HorizontalPodAutoscaler

back off restarts of crashlooping containers

Use GCE defaults for per-PR Jenkins

Be more lenient when deriving the node tag from a node name on GCE

Added more API conventions.

Always select non link-local network interface, fixes #11961.

Don't reuse credentials on cluster create

Fix portforward test flake with GOMAXPROCS > 1

- Log error when task state is unknown

- The TASK_ERROR task status was introduced with Mesos 0.21 and is actually used since 0.22.
  It was not handled at all before this patch, leaving errored task in the registry in phase
  "Pending". This will lead to task status updates from the Mesos Master on reconciliation with empty
  slaveId fields, leading to scheduler crashes eventually.
- Handle terminal task with empty slaveId.
  The slave id can be empty for TASK_ERROR.
  The modified code path does not use the slaveId.

Fix sort on sets in hack/verify-flags-underscore.py

Fix undefined variable f in 'finally' clause of verify-flags-underscore