JAXenter | March 29, 2017

Kubernetes 1.6 and the community portal K8sPort

For open source tools, nothing is more important than an active community. This was also recognized by Kubernetes, which is why the portal K8sPort started last week. Through “Gamification”, users are motivated to be an active part of the community. If so, they receive rewards. The first points were certainly already gained by hard-working developers: Kubernetes 1.6 appeared shortly after the publication of the portal.

K8sPort – for a strong community

Last week, a new portal was launched by the Cloud Native Computing Foundation (CNCF) with K8sPort , whose goal is to expand and reward the activity in the community of Kubernetes. Users gather points based on their commitment to the development of Kubernetes and the community.

There are points for completing so-called “challenges”, which can of course be commits, but also the attendance of conferences, feedback or answering questions such as stack overflow are rewarded. The collected points can then be converted into prizes, such as badges to match the performance or conference ticketing. A conversion into a non-profit donation is also possible.

Kubernetes 1.6 – for better scaling and protection

Some developers have already earned the first points on K8sPort, because yesterday appeared version 1.6 of the orchestration tool Kubernetes. As always, the team does not save on figures and so it should be mentioned here that about 275 developers were involved in Kubernetes 1.6 and about 5,000 commits were made.

Scaling and connecting

The number 5,000 is also important in other respects for the current update, as now Kubernetes Cluster supports up to 5,000 nodes. This increase in the cluster size by 150 percent is possible thanks to the new version of the Key Value Store etcd v3 developed by CoreOS. If that is not enough, federation can combine multiple clusters and address a single API endpoint.

The command line utility kubefed changes from alpha to beta . On board it has improved support for on-premise clusters and it now automatically configures the kube-dns on added clusters. In addition, it can now forward parameters to connected components.

Secure and inspect

In the security area, roll-based access control (RBAC) has now reached the Betastadium. It allows a much more strict and clearer adjustment of the access rights to system components, which makes the work safer. Cluster administrators can selectively regulate the access of users or service accounts on a per-namespace basis according to their needs.

Betaphase has also begun for kubeadm . For this, it was equipped with a new set of command line flags. The new set of basic features also includes an RBAC setup, the use of the bootstrap token system, and an improved Certificates API .

Improved scheduling and other changes

Kubernetes 1.6 was also revised with regard to scheduling. The current release includes a set of powerful and versatile scheduling constructs that give users greater control over the scheduling of pods. It is now possible to limit pods to a particular node in heterogeneous clusters.

Own and individual schedulers can either run alongside the standard scheduler of Kubernetes, or even replace it. It is important that each scheduler is responsible for another set of pods. Like many other features, the multiple schedulers are still in the betaphase.

The implementation of the Docker Command Runtime Interface is activated by default in kubelet . Alpha support is available for other runtimes, such as cri-o, rkt, and frakti. Another alpha feature is, for example, the support for the use of several Nvidia GPUs simultaneously (Docker Runtime only).

Aparna Sinha has summarized further innovations on the Kubernetes blog , there is also detailed information about all the highlights of the update. The release notes for Kubernetes 1.6 on GitHub provide a complete overview of all improvements, new features and bug fixes.