Deploying Kubernetes on Mesos [Mesosphere HackWeek]

In this first post of the Mesosphere HackWeek series, we take a look at the integration of Apache Mesos and Google’s Kubernetes project.

Kubernetes is an open-source cluster-orchestration system for containers, based on Google’s Borg system, and it’s one of the most requested services on top of Mesos and the Mesosphere Datacenter Operating System (DCOS). It’s designed to simplify how you create distributed applications and run them reliably on any infrastructure. Kubernetes provides declarative primitives—such as pods, labels and services—to group the containers that make up an application into logical units, for easy management and discovery.

During the HackWeek, developers from both the Mesos and Kubernetes communities worked on the integration, marrying Mesos’ resilient and scalable resource allocation with Kubernetes’ pod management. Kubernetes pods can now run alongside other applications and services, including traditional web servers, application servers, Hadoop, Spark and other frameworks — all on a single cluster of machines.

In the weeks following HackWeek, Mesosphere and Google significantly advanced their partnership to deliver enterprise-grade Kubernetes, backed by Mesos and the Mesosphere DCOS. As part of the Mesosphere-Google collaboration, Kubernetes will soon ship as a pre-installed service on the Mesosphere DCOS.

If you want to learn more about the original Mesos-Kubernetes integration, take a look at the “Getting started with Kubernetes on Mesos” document. There is also further information about the underlying architecture available on GitHub. If you want to follow (or contribute to) the upstream integration of the two projects, track it via the GitHub label cluster/platform/mesos.