Migrating an OpenShift etcd Cluster

Summary

Following on from my previous post setting up an OpenShift cluster in Vagrant, this post discusses migrating an etcd cluster within a live OpenShift instance to newer servers.

Moving a standalone etcd cluster is relatively straightforward, but when it’s part of an OpenShift cluster — and especially one that’s live and operational — it is a little more involved.

The ordering of actions is important and there are several aspects to consider when planning such a move:

  • Config management preparation
  • Stopping the cluster
  • Creation and distribution of certificates
  • Data migration
  • Update of OpenShift config
  • Update of config management

Here we are using Ansible to provision and maintain the environment.

You can also use Chef to manage your OpenShift cluster.


Code

The code for this is here:

Video

Here’s a video of the upgrade process:

Steps

VM Setup

This section of the code sets up the VMs using Vagrant.

Cluster Setup

The next section sets up the OpenShift cluster. It:

  • sets up ssh access across all the hosts
  • writes the ansible hosts config file
  • triggers the ansible playbook

Take a Backup

Take a backup of etcd on all three nodes, just in case.

Stop the Cluster

Generate New Certs

For each new node, run the commands to generate the certs for the new nodes, and copy to the codes.

Add etcd Nodes One-By-One

Again for each node:

  • add the new node to the etcd cluster
  • go to the node
  • install etcd
  • extract the certificates
  • update the etcd config
  • restart etcd

NOTE: If you have a lot of data in your cluster, you will want to give the new node ample time to receive the data from the other nodes. In this trivial example, there is little data to transfer. Alternatively, you can copy over the data from one of the original nodes.

Drop the Old Members

Now drop the old members from the cluster and remove etcd from those hosts:

Update the Master Config and Bring the OpenShift Cluster Back Up

The /etc/origin/master/master-config.yaml file needs to updated to reflect the new etcd cluster before bringing back the OpenShift cluster.

Update Config Manager and Re-Run

Learn More

My book Docker in Practice:

Get 39% off with the code: 39miell

Advertisements

One thought on “Migrating an OpenShift etcd Cluster”

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s