Tag: Fault-Tolerance
20 May 2016 in Demo, Fault-Tolerance

Cheapest Fault-Tolerant Cluster For Deis V1 PaaS

This guide is targeted to the Deis v1.13 LTS branch. With over 6 million downloads, Deis V1 PaaS has never been more popular. These instructions will not work with Deis Workflow, the Kubernetes-native PaaS, which is still currently in beta.

I am going to demo a cheap, quick, and dirty way to bring up a cluster for Deis V1 PaaS on DigitalOcean. This is the cheapest fault-tolerant configuration that I could manage. That is, a cluster on which some nodes can go down—but the cluster and its platform services remain available!

Before we go on, let's distinguish fault-tolerant from highly available.

Fault-Tolerant or Highly Available?

A fault-tolerant Deis cluster continues operating, even in the event of some partial or complete failure of some of its component nodes. This guide will show you how even a cheap Deis cluster can be fault tolerant—and to what extent.

A highly available (HA) Deis cluster is one that is continuously operational, or never failing. HA is the harder of the two to achieve, and this tutorial won't show you how to do that.

High availability is something that isn't provided out-of-the-box in most cases. If HA is one of your goals, your sysadmin or dev team will be better qualified to say what obstacles you will face in maintaining a standard of high availability for your apps.

Read More