* Rescheduler relies on taints, which were changed in a backward incompatible way. Rescheduler 0.3 shipped together with Kubernetes 1.6 understands the new taints and will clean up the old annotations, but Rescheduler 0.2 shipped together with Kubernetes 1.5 doesn't understand the new taints. In order to avoid eviction loop during 1.5->1.6 upgrade you need to...

- When the cluster size is small and the scheduler doesn't get unscheduled Pods frequently, the scheduler doesn't try to reschedule Pods in some cases. ([#119784](https://github.com/kubernetes/kubernetes/pull/119784), [@sanposhiho](https://github.com/sanposhiho)) [SIG Scheduling and Testing]

## Dependencies

### Added
_Nothing has changed._

          * 'node.kubernetes.io/network-unavailable=:NoSchedule' if NetworkUnavailable is true
          * 'node.kubernetes.io/disk-pressure=:NoSchedule' if DiskPressure is true
          * 'node.kubernetes.io/memory-pressure=:NoSchedule' if MemoryPressure is true
          * 'node.kubernetes.io/out-of-disk=:NoSchedule' if OutOfDisk is true

([#63955](https://github.com/kubernetes/kubernetes/pull/63955), [@k82cn](https://github.com/k82cn))
- Remove rescheduler since scheduling DS pods by default scheduler is moving to beta. ([#67687](https://github.com/kubernetes/kubernetes/pull/67687), [@Lion-Wei](https://github.com/Lion-Wei))

          * 'node.kubernetes.io/network-unavailable=:NoSchedule' if NetworkUnavailable is true
          * 'node.kubernetes.io/disk-pressure=:NoSchedule' if DiskPressure is true
          * 'node.kubernetes.io/memory-pressure=:NoSchedule' if MemoryPressure is true
          * 'node.kubernetes.io/out-of-disk=:NoSchedule' if OutOfDisk is true

## External Dependencies

* Default etcd server version is v3.2.18 compared with v3.1.12 in v1.10 ([#61198](https://github.com/kubernetes/kubernetes/pull/61198))
* Rescheduler is v0.4.0, compared with v0.3.1 in v1.10 ([#65454](https://github.com/kubernetes/kubernetes/pull/65454))
* The validated docker versions are the same as for v1.10: 1.11.2 to 1.13.1 and 17.03.x (ref)

- Fixed a scheduler panic when using PodAffinity. ([#82841](https://github.com/kubernetes/kubernetes/pull/82841), [@Huang-Wei](https://github.com/Huang-Wei))
- Take the context as the first argument of Schedule. ([#82119](https://github.com/kubernetes/kubernetes/pull/82119), [@wgliang](https://github.com/wgliang))

- Ensured that the `node-role.kubernetes.io/master` taint is applied to the master with NoSchedule on GCE. ([#78183](https://github.com/kubernetes/kubernetes/pull/78183), [@cheftako](https://github.com/cheftako))

* Fix an issue where filesystems are not unmounted when a backend is not reachable and returns EIO. ([#67097](https://github.com/kubernetes/kubernetes/pull/67097), [@chakri-nelluri](https://github.com/chakri-nelluri))
* Add NoSchedule/NoExecute tolerations to ip-masq-agent, ensuring it to be scheduled in all nodes except master. ([#66260](https://github.com/kubernetes/kubernetes/pull/66260), [@tanshanshan](https://github.com/tanshanshan))

- Remove deprecated `--generator, --replicas, --service-generator, --service-overrides, --schedule` from `kubectl run`
  Deprecate `--serviceaccount, --hostport, --requests, --limits` in `kubectl run` ([#99732](https://github.com/kubernetes/kubernetes/pull/99732), [@soltysh](https://github.com/soltysh))