This has been reposted from our lab tumblelog

Keeping all the machines in the lab backed up seems to be a constant issue. I thought I might outline our scheme. 

First… what are the machines?

imageI have a primary machine that has 5+TB of disks, there are around 6 common computers, a small cluster computer, as well as around 3 personal laptops. The strategy for backing up is quite complex and has varied over the years. The tools used currently include

  • Crashplan
  • Backblaze
  • Bitcasa
  • Time Machine
  • Super Duper
  • Dropbox
  • Synology box


I like Crashplan for a few reasons. One is that I can easily keep updated on the status of all the machines and you can back up to local machines (free). This is very handy when you have a suite of machines in the lab. We have a “Family Plan” to back up 10 machines to the cloud.

I have two unlimited plans for my main machine and the main node of the small cluster on Backblaze.


I started using Bitcasa in the beta preview days and it seems like another great option. Currently I am over my free quota.


Several persons in the lab use Time Machine.  I have not found much value as it is just local. If I was interested in local backup I would prefer using that feature on Crashplan.

SuperDuper is a different kind of backup. I use this to clone entire hard drives. Besides using when upgrading, I regularly clone my main machine as well as the nodes on the cluster.

Finally, Dropbox is used as a means to backup – I have a paid account where ~100GB of data stays synced in the cloud

We also maintain a 8TB Synology NAS, however this is primarily used for working with current files. Each lab member has their own folder.


More info @

via the Lab Tumblr:

Backing up computers in the lab