Aug 26, 2019

Raft in Kubernetes and Swarm Pt. 1: The Raft Algorithm

By Bret Fisher • Issue #10 • View online

If you’re managing a Swarm or Kubernetes cluster, you’re acting as the Ops part of DevOps. As part of this role, fault tolerance and cluster management are vital responsibilities. Here’s part one of a multi-part newsletter on understanding your cluster data store and how to optimize and protect it.

First, Understand the Raft Algorithm

Kubernetes uses the etcd key-value store for its control plane (Master) database, and Docker Swarm Managers use a built-in data store that is similar to etcd. Both use the Raft Algorithm to communicate between these control plane nodes safely and make decisions about state.

Here’s Laura Frank Tacho’s talk on a lot of these topics, with some more details in links below that:

Everything You Thought You Already Knew About Orchestration

Real-Time Container Visibility with Datadog

This weeks sponsor: Enhance visibility into the performance of your entire containerized environment and investigate issues quickly with Datadog. Automatically detect and monitor your containers as quickly as they are created and destroyed with the cluster agent, and easily analyze the health of all your containers in a single screen with the live container map. Try monitoring Docker with Datadog for free and they’ll send you a free t-shirt!

www.datadoghq.com • Share

The Secret Lives of Data

Watch a cluster communicate and elect leaders with moving graphics (very helpful). It’s similar to the one Laura uses above to explain how raft nodes are talking and making decisions, and what happens when one loses communications with the others.

thesecretlivesofdata.com • Share

The Raft Consensus Algorithm

All the details, including tools that implement raft.

raft.github.io • Share

Latest Topics On My YouTube Channel

Remember I’m live every Thursday on a new topic, with guests sometimes. Here’s a quick list of what I’ve done lately:

My two favorite code editors VS Code and SpaceVim

Rancher co-founder on Kubernetes, k3s, k3OS, and more

ENTRYPOINT vs CMD

Doing sysadmin stuff in containers

Thanks For Reading

I’ll be back next week with the next step in understanding your cluster data store: building clusters with fault tolerance.

– Bret

Did you enjoy this issue?

By Bret Fisher

Frequent updates on my projects, videos, and opinions focused on the container ecosystem, including Docker, Kubernetes, Docker Swarm, CI/CD, and container DevOps.

Tweet

Share

If you don't want these updates anymore, please unsubscribe here.

If you were forwarded this newsletter and you like it, you can subscribe here.

Bret Fisher, Virginia Beach, Virginia, USA