4 Comments

CrashPlan for Large, Distributed, Cheap, Off-Site Backup

In the early 90s, my friend’s father took me to EDS where he worked at the time.  I remember him saying, “this is one of the largest data centers in the world.  They have over 3 terabytes of data in there.”  In the homemade box tucked away quietly in my hall closet is a 6x1TB RAID with another 1TB disk for the OS.  Add in the media center and 3 laptops and I’ve got a lot of data just waiting to be lost with a disk failure, theft, or an accidental rm -rf.

What I Need in a Backup Solution

As I thought about my data, I came up with a few criteria before I started scouring the net for a solution.

  1. No constraints on backup size – The data I want to backup exceeds 2TB and is growing.  I’ve used cool apps like DropBox that have arbitrary upper limits like 100GB.  However, the coolest app though won’t do me any good if I can’t backup everything I need. (To be fair, backup is just one tiny element of what DropBox does.  I highly recommend that app for the other things it does, like sync.)
  2. Highly configurable – That 2TB I mentioned lives amongst tons of other stuff that I keep as sort of a cache, but wouldn’t miss it too much if it got deleted.  I need to be able to clearly specify what data I actually want backed up.  Moreover, I need a high degree of control about backup policies, security, etc.  I like solutions that make things simple, but in this case there also needs to be a way to get as complicated as I like.
  3. Distributed backups – Part of the reason I have that 6x1TB RAID array is for super-fast local backup.  Obviously that won’t do me any good if my house burns down, but if a laptop crashes is way easier to grab 500GB from a local machine than it is to pull it across the net.  I want to be able to backup to a service as well as many other computers that I specify both in my house and on the Internet.
  4. Smart, low profile application – Modern OS’s like Mac OS X keep a log of what files have changed.  I don’t want a dumb service that does things like that on its own and consumes my computers’ resources.  I need something that will run in the background and not make any noise.
  5. Accessibility – I need a service that runs on any platform, specifically Mac OS X and Linux.  Moreover, I need to be able to access my backups from the web.
  6. Cheap – I want to pay for storage, not bandwidth.  Less than $10/mo is my general rule of thumb.

There are other minor points, but those are the non-negotiable items.

What I Tried, But Didn’t Like

  1. RackSpace Cloud Files – great for storing files related to a web service, but ultimately too costly and would really only provide a storage point, i.e., no real app dedicated for backup.  (I only considered this because I use RackSpace for web hosting, at which they are quite good.)
  2. BackBlaze – Great service and app for a novice user.  Ultimately it failed big for criteria 2 and 3 above though.  The service is strongly opinionated about what you can’t backup.  Also, you pay per computer.  If you’ve got 1 then $5 a month isn’t much, but obviously $5 per computer can add up quickly ($25 in my case).

CrashPlan

The Solution: CrashPlan

CrashPlan is rad.

I purchased their CrashPlan+ plan.  $6/mo for all my computers and unlimited backup storage space.  The software is pretty great too. Highly configurable, good encryption, nice online support, runs on Win/Mac/Linux, and barely consumes any system resources when active (and can be active as a background/headless process).

As an example, the laptop I’m on now backs up in 3 ways:

  1. Time machine to a Linux share on my local RAID. Why: nice UI, quick access for day-to-day type stuff, good to have for a total failure.  (This option has nothing to do with CrashPlan.)
  2. CrashPlan to a Linux share on my local RAID.  Why: backs up my critical files locally in a manner that allows me to access them easily from any home machine.  The method in option 1 has been problematic in the past, so this makes me feel a little better that the Apple-way doesn’t result in a loss of data. (Read more on sparsebundles if you’re curious.)  In addition, it’s lightning fast.
  3. CrashPlan to a CrashPlan server.  Why: full off-site backup.  In addition to encrypting based on my login/password, they also allow you encrypt with a private key, which means even the folks at CrashPlan can’t see my data.  This is the ultimate backup.  If I was ever unfortunate enough to lose all my computers, I would be able to get my data back from this service.  Moreover, this allows me to access my backed-up data from anywhere in the world.

Final Thoughts

There are many other solutions out there, so please consider your own criteria and pick the best one.  That said, if you ever find a solution that compares to CrashPlan on the criteria I’ve mentioned, I’d love to hear about it.

Also, I had a great conversation with a friend about smart sync when multiple computers are involved.  If that’s a requirement, I recommend checking out DropBox or the like.  While CrashPlan syncs your computer with the backup, it does not attempt to sync all of your computers to have the same versions of files.  Using something like DropBox in conjunction with CrashPlan is a great option.

4 Comments

  1. Excellent comparison. I am in a similar boat but for a client of mine.

    They have a 4 1tb raid5 Synology ds411+ I setup up for them to use as their main share drive. It utilized live by 20 macs at once and does an amazing job of handling them all without any lag.

    I have that NAS rsync to a netgear raid5 drive and also to an external USB 2tb drive for peace of mind. I’ve been hired to backup all this data offsite with a lost cost. Crashplan is what I use at home for my wife’s photography business, it also backs up to a raid5 NAS.

    I’m trying now to install the linux crashplan service to the synology NAS since that’s it’s gui interface. I found someone who has done it with another model of the synology NAS drive but mine fails to start. I’ll eventually figure this out.

    I also have been hired to backup their rackspace hosted website files to an offsite backup. Since rackspace cloud is technically only backing up 1 day at time it’s way to risky. So I found a way to pull those files daily from rackspace to the synology NAS as well but not from the NAS itself. I’m very new to coding in linux but I’m well versed in all windows and mac environments.

    My point is I agree that crashplan is a fantastic, customizable, cheap backup solution for home but is it safe for a business? The crashplan pro is a bit more monthly cost and cannot backup to another non cloud offsite location such as the owners home. Hence my push toward the home service.

    So is it a better solution to remove the netgear NAS from use and create a Linux box similar to what you’ve made, have my synology nas rsync to it, have it pull the rackspace files, run the crashplan client encrypted to my offsite location and the crashplans servers for $6 dollars a month (or whatever it is now)? It seems right and safe and fast. This way they can still utilize the synology NAS as their primary network share for the speed benefits and not have to have a hack job command line crashplan service running on it. Thoughts on my long winded post? :)

  2. Thanks for the outline of your plan. Super helpful.

    I’d love to see how you setup your server. I’m using a Macbook Air a primary machine now and I can’t have all my data there. So I’m looking to build a home server (RAID, of course) to store everything (docs, photos, etc.) that I can access from away from home, and then back that up to crashplan or something.

    1. Aside from the RAID, the server is fairly simple custom build: ASUS board, Intel i7 processor, and 6GB of RAM. I’m running Ubuntu 11.04 (I usually prefer Debian, but Ubuntu handled my RAID card better so I made the switch). I’ve got details about the RAID in a separate blog post.

      I have the server running netatalk and avahi to have it advertise Time Machine backup services. All of the Crashplan backup services are setup using the client software.

      I definitely recommend setting up NFS (or whatever protocol suits your needs) and have OS X auto-mount it. That’s what I do to quickly pass gigs of raw digital video files, music, etc., across the local network Assuming your network hardware is fairly new, it should feel like you’re working locally on your laptop. RAID is especially nice for saving huge files (or huge sets of files) as it allows extremely fast writes.

      Hope that helps.

Comments are closed.