Run a Ceph Cluster With Flocker on Your Laptop

At ClusterHQ, our users have asked us for Ceph support. So, here is a first glimpse into the support Flocker provides for Ceph. The great part about releasing a Ceph driver is that we can let anyone who wants to try Flocker, Docker, and Ceph together to do so on their own laptop — no cloud provider account needed!

In this blog post, we will introduce you to our Ceph driver and show you a demo of running a Ceph cluster that works with Flocker and Docker. The post also includes resources on how you can get started yourself.

Introduction

What is Ceph?

Ceph is a distributed object store and file system designed to provide excellent performance, reliability and scalability. Ceph clusters allow you to access object storage through file systems with Ceph FS and block devices via Ceph RBD. To read more about Ceph, visit the documentation.

By using Ceph and Flocker together, you are able to automatically provision and manage Ceph RBD devices to your containers. Application containers accessing Ceph storage can move around your infrastructure freely while Flocker automates the movement of your data from the container's original location to its destination without any need for you to manage Ceph. By doing so, you can decrease application fail-over times, increase the efficiency for distributed database recovery, and keep your data safe and persisted.

Demo

In this demo, we are using Vagrant to automatically deploy Flocker, Ceph, and Docker to a local cluster running on Oracle VirtualBox. This demo can be run easily on your laptop or desktop as long as you have both Vagrant and VirtualBox installed. To read more on how to run the demo yourself, scroll to the links just below the video.

You may also watch the asciinema version of the video so you can follow along and copy and paste commands used in the demo.

If you're interested in trying this out on your own laptop, you can visit this GitHub repository and follow the README to run the demo.

Also, if you would like to use the Ceph-Flocker driver for your Flocker environment but do not want to use Vagrant or Virtualbox, you can navigate the driver's GitHub page.

Using the driver in your Flocker environment means you can deploy Ceph both locally or in any cloud environment that supports running Ceph to get the ultimate flexibility out of your storage and container environment.

You can install the driver and reference the driver in your dataset agent.yml using the following example:

version: 1
  control-service:
    hostname: "user.controlserver.example.com"
  dataset:
    backend: "ceph_flocker_driver"
    cluster_name: "Ceph cluster name, defaults to ''"
    user_id: "Ceph user ID, defaults to 'admin'"
    ceph_conf_path: "path to ceph conf for rados, defaults to '/etc/ceph/ceph.conf'"
    storage_pool: "name of storage pool to use, defaults to 'rbd'"

Conclusion

That's it for now. We would love to know what you think about using Ceph with Flocker to manage your persistent block storage for containers so feel free to reach out to any of our channels for feedback, support, comments and questions.

We'd love to hear your feedback!