03/07/2018 |

Configure OpenIO SDS as a Backup Target

After explaining the many ways that you can install OpenIO SDS, showing you how to add a load balancer, and how to play with the software to understand its basic commands, it is time to show how to install third-party software to take advantage of it.

In this tutorial I'll show you how to install Duplicati and use this software for your home backups. This article will highlight how easy it is to configure OpenIO SDS as a backup target for other solutions as well. Duplicati is the easiest to use at home, and I wanted a new private-cloud-ish backup solution for my Mac.

I have limited resources at home, but if you want me to try other software, or you want to try other software yourself and have the chance to win a Raspberry Pi kit, join our Slack community and let me know.

What is Duplicati?How to size an OpenIO cluster on x86 servers

Duplicati is backup software for Windows, Mac, and Linux. It has a nice feature set including encryption, a cloud-ready design, a simple UI, and is open source.

Even though the Mac has Time Machine, this only works for local backups; Duplicati lets me back up my laptop from anywhere. The final goal of setting up this software is to share the IP address of my SDS object store on the internet and perform backups even when I travel.

Let's do it!

To install Duplicati on macOS, you first need to install the Mono framework.  Here is the link to download the package.

Duplicati will need a bucket to save its backups; you can create it with any client. To perform this operation, I used awscli which was already installed on one of my Linux hosts.

You can then download and install Duplicati on your Mac.

When you double-click the app, an new icon appears in the menu bar.  Click it and choose Open.

This is what you should see in the web browser.

Screenshot 2018-06-23 14.25.35


In a Terminal window, I create the bucket I'll use for my backups:

# aws --endpoint-url http://oio01:6007 --no-verify-ssl s3api create-bucket --bucket duplicati-backups

After doing that I just need to configure my backup job by clicking Add Backup:

Screenshot 2018-06-23 14.32.37


After clicking Next, I add the name of my backup and the encryption passphrase (you can also leave it empty).

Screenshot 2018-06-23 14.34.37


In the next step, I add the relevant information for the backup destination; your screen should look like this:

Screenshot 2018-06-23 14.38.04

 A few notes here:

  1. If Duplicati asks for a URL do not put "http://" in front of the endpoint address.
  2. Be sure to specify the bucket you created previously.
  3. Look at the advanced options. The default is Signature V4 (but you need Keystone installed for this); if you want to use TempAuth (Signature V2) then you have to look in the Advanced Options menu and choose s3-ext-signatureversion and enter 2 in the related field.


Screenshot 2018-06-23 14.43.16


If everything is OK, click Next. On the next screen, choose what you want to back up.


Screenshot 2018-06-23 14.44.49


Schedule your backups now.


Screenshot 2018-06-23 14.47.06




And, last but not least, choose the retention.


Screenshot 2018-06-23 14.48.48


Click Save, and if you want to run your backup immediately click Run. You'll see a dialog showing information about the backup’s progress.

That's it! It’s easy and straightforward.

Final note

This was just an example but, for most backup solutions, the process is similar. Now that practically all backup software can use an S3 endpoint as a target, object storage is becoming a popular solution because it is much more flexible, scalable, and cost effective than traditional backup appliances.

Even more so, backup is one of the essential workloads that can lead to the object store improving overall infrastructure efficiency and consolidation of secondary data in your organization.

OpenIO SDS has an additional advantage when it comes to inexpensive but safe storage. It can be installed on any type of hardware, even a combination of different types of hardware, and is available with a friendly subscription model. In practice, it is possible to recycle decommissioned hardware and build large-scale clusters with heterogeneous nodes. This gives a second life to old hardware without sacrificing reliability, and it reduces storage infrastructure costs and moves them from CAPEX to OPEX.How to size an OpenIO cluster on x86 servers