- Changed CRI pod log directory from `/var/log/pods/UID` to `/var/log/pods/NAMESPACE_NAME_UID`. ([#74441](https://github.com/kubernetes/kubernetes/pull/74441), [@Random-Liu](https://github.com/Random-Liu))

* Fix TerminationMessagePath ([#23658](https://github.com/kubernetes/kubernetes/pull/23658), [@Random-Liu](https://github.com/Random-Liu))
* Fix gce.getDiskByNameUnknownZone logic. ([#24452](https://github.com/kubernetes/kubernetes/pull/24452), [@a-robinson](https://github.com/a-robinson))

  * Detach the volume when pod is terminated ([#45286](https://github.com/kubernetes/kubernetes/pull/45286), [@gnufied](https://github.com/gnufied))

  * CRI: Fix StopContainer timeout ([#44970](https://github.com/kubernetes/kubernetes/pull/44970), [@Random-Liu](https://github.com/Random-Liu))

- Migrated the `volumebinding scheduler plugins` to use [contextual logging](https://k8s.io/docs/concepts/cluster-administration/system-logs/#contextual-logging). ([#116803](https://github.com/kubernetes/kubernetes/pull/116803), [@mengjiao-liu](https://github.com/mengjiao-liu))

* Fix `kubectl logs -f` for windows server containers. ([#81747](https://github.com/kubernetes/kubernetes/pull/81747), [@Random-Liu](https://github.com/Random-Liu))
* Fix VMSS LoadBalancer backend pools so that the network won't be broken when instances are upgraded to latest model ([#81411](https://github.com/kubernetes/kubernetes/pull/81411), [@nilo19](https://github.com/nilo19))

    * This keeps the behavior consistent with Docker integration, and fixes an issue that some container Prometheus metrics don't work when there are summary stats for multiple instances of the same pod.

* Fix kubelet request timeout when stopping a container. ([#46267](https://github.com/kubernetes/kubernetes/pull/46267), [@Random-Liu](https://github.com/Random-Liu))
* Add generic NoSchedule toleration to fluentd in gcp config. ([#48182](https://github.com/kubernetes/kubernetes/pull/48182), [@gmarek](https://github.com/gmarek))

* Kubelet: Add image cache. ([#38375](https://github.com/kubernetes/kubernetes/pull/38375), [@Random-Liu](https://github.com/Random-Liu))
* Collect logs for dead kubelets too ([#37671](https://github.com/kubernetes/kubernetes/pull/37671), [@mtaufen](https://github.com/mtaufen))



# v1.4.7

- Kube-scheduler implemented scheduling hints for the `NodeAffinity` plugin. The scheduling hints allowed the scheduler to only retry scheduling a Pod that had been previously rejected by the `NodeAffinity` plugin if a new Node...

- Fixed the `PodAndContainerStatsFromCRI` feature, instead of supplementing with stats from cAdvisor. ([#113291](https://github.com/kubernetes/kubernetes/pull/113291), [@mengjiao-liu](https://github.com/mengjiao-liu))
- Fixed the occasional double-counting of the `job_finished_total` metric. ([#112948](https://github.com/kubernetes/kubernetes/pull/112948), [@mimowo](https://github.com/mimowo))