Interesting new features in K8s v1.20

 Kubernetes 1.20(The Raddest release) was released in December 2020 and it had lot's of new features(42 enhacements).
You can find the complete changelog here: K8s 1.20 CHANGELOG.

But there were few interesting features and changes which caught my eye and I thought of listing them them out here.

K8s 1.20(The Raddest Release)

i) Support for Docker Shim going away.

Starting with version 1.20, Kubernetes will deprecate Docker as a container runtime for CRIO(), also it's likely that from version 1.22, support for Docker as a container runtime will be removed.

Docker runtime came with a maintenance overhead (e.g: Docker daemon could build images too.),  hence it was necessary to make k8s independent of runtimes.

ii) IPv4 / IPv6 dual stack support.

IPv4/IPv6 dual-stack networking enables the allocation of both IPv4 and IPv6 addresses to Pods and Services. IPv4/IPv6 dual-stack networking allows the simultaneous assignment of both IPv4 and IPv6 addresses.

iii) Kubectl debug command goes GA.

Now with kubectl you can use the debug argument from version 1.20,
You can use the kubectl debug command to add ephemeral containers to a running Pod.

Example usage of kubectl debug:-

kubectl debug -it podName --image=busybox --target=podName

Note: Earlier we could use the same feature via 

kubectl alpha debug <<...args>>

iv) Volume snapshot operations graduated to stable state.

Finally after a long time volume snapshot's operations are GA in v1.20. This will help all external storage providers (E.g NetApp) to build stable volume snapshot workflows and maintain volume workflows in line with the standard kubernetes volume lifecycle.

If you already have NetApp storage then you should checkout their cool CSI Trident(https://github.com/NetApp/trident)

v) Pod Hostnames now can be FQDN's.

Earlier we only had short names for pod hostnames, starting version 1.20 K8s will also allow FQDN's to be pods hostnames.