Friday, December 25, 2015

Kubernetes: Scheduling the Future at Cloud Scale

Kubernetes: Scheduling the Future at Cloud Scale
As remarkable as Docker containers are, they still need to be heavily scheduled and orchestrated to run efficiently, and seamlessly replaced and re-balanced when they fail. It’s a big job, well beyond the abilities of Chef, Puppet, and similar tools. This O'Reilly report introduces you to Kubernetes, a cluster management system from Google that’s capable of scheduling and launching roughly 7,000 containers a second.

Dave Rensin, who directs Global Cloud Support and Services at Google, explains piece-by-piece how this system works. If you’ve created at least one simple container with Docker, you’ll understand how to get started with Kubernetes by using one of several options.

  • Examine pods, a collection of containers bundled and scheduled together
  • Run through volumes, the filesystems your container can see and use
  • Create and organize pods with labels and annotations
  • Use replication controllers to manage replicas: multiple copies of a pod
  • Set up services, long-lived endpoints that identify a set of pods in your cluster
  • Get started using your own servers, VMs from the public cloud, or a managed offering from a major cloud provider
  • Find examples on GitHub of Kubernetes in use, including WordPress and Guestbook

0 commentaires:

Post a Comment

Note: Only a member of this blog may post a comment.