Now that I have a baby, I’m taking a lot more pictures and recording a lot more video than I ever have before. And as you can imagine, I’d hate to lose any of it. So I spent some time over the last few months researching backup options for Mac OS X. I’ve read many articles and reviews, talked with a number of colleagues (including some who write backup software and others who run datacenters), and checked out lots of software and gear (software like Time Machine, SuperDuper!, and Carbon Copy Cloner, external hard drives like those made by Seagate, LaCie, Western Digital, and also “smart” hardware like the ones made by Iomega, ReadyNAS, and Drobo).
I take file backups and redundancy seriously. You might think it’s because, in a past life, I was a systems and network engineer for an aerospace company, in charge of the backups for all of our corporate data, everything from payroll and HR to user-created files. Nobody ever forgets those middle-of-the-night visits to the server room, bleary-eyed, standing at the VT220 hard-wired console, using commands like
restore, working through stack after stack of tapes from the fire-safe, so the VP of Something can get that file he deleted last week (or was it last month?), for that meeting in whatever town tomorrow morning.
But it started further back in time, when I was in my very early teens. I had a Commodore 64, and was using the GEOS operating system) and geoWrite, one of the very first WYSIWYG text editors, to create a paper for school. I’d saved up enough to add a 5¼" floppy drive, and although saving files was much faster than when the cassette tape, it was still slow, and generally I didn’t save very often, except, when I was taking a break or done for the time being.
If you have good file-saving practices, you’re probably shaking your head right now, but remember, I was about 10 or 11 years old at the time and I’d never lost a whole day of work before, especially not the day before it was due. So when the printer jammed and GEOS crashed and I lost most of everything I’d typed up, I was actually a bit surprised. Until that time, I’d always thought of computers and their hardware as being completely reliable.
Ever since then, I’ve been very careful to keep multiple copies of my work. In the old days, that just meant buying a second floppy disk (my entire allowance). Today, it’s not quite so simple. The choices we have both in software and hardware can be overwhelming.
The Short Answer
I’m now relying primarily on a combination of a Drobo, Subversion, automatic .Mac syncing, and occasional drive-cloning. Keep reading for the details.
As with most aspects of my life, my goal is to simplify. The fewer tools, applications, and devices I need, the better. In the past, I used to backup everything, but now, although I do backup my entire system, I’m mainly interested in a backup of what I consider to be critical data. This includes:
- Code and Writing
- Bookmarks, Address Book, Serials, and Passwords
- Other Stuff
What about the applications themselves? What about preferences? What about system-specific settings? Those are all handled by the system level backup/clone. But you know what? After many years of preserving those files, I’ve found that when it comes to actually restoring my machine or setting up a new one, I rarely took those things along with me, preferring to install only what I needed as I needed it. I’ve really been able to keep a minimalist setup in the process, shedding things here and there as I move from system to system. Still, it’s nice to know they’re there when I need them.
I’ve found that a multi-pronged approach works best for me, and I use a few different technologies to make it all work. Reading this, you might think I’m a bit retentive when it comes to data redundancy. And you’re probably correct … but then again, I’ve seen many drives fail in my time, and lots of data get lost. I just don’t want to take that chance with the stuff that matters.
Code and Writing
All of my code lives in a Subversion repository which is offsite and fully backed up, and I always have a relatively local copy ready to go, should something odd happen to the repository. The merits of using a revision control system are many, and well beyond the scope of this article, but if you’re not using something like Subversion, Git, or heck, even CVS, you really should be.
Another nice benefit to keeping code and writing in a repository, aside from the revision history, is that you automatically have an easy path to synchronized data on any machine you use.
Bookmarks, Address Book, Serial Numbers, and Passwords
I use .Mac to sync data between my Mac Pro and my Macbook, and .Mac keeps these things backed up in the process. It’s true that .Mac is a bit expensive, and this data is backed up elsewhere (see below), but I’m hoping that one day, Apple will do something really amazing with .Mac and I’ll be glad I’m signed up for it.
I use Yojimbo, from Bare Bones, to manage all of my serial numbers and passwords. It’s a great utility, and it can even use .Mac to sync (and therefore backup) your data.
Get a nice, reliable host like Mailtrust (formerly Webmail.us), and store your mail there in IMAP folders. They keep good backups, and you can go from machine to machine and everything will be right where you left it. Unfortunately, Mailtrust has recently recently raised their prices, so you might want to check out Google Apps for a free (but possibly less robust) solution. The main thing is IMAP folders and a company that backs things up reliably.
Intermittently, I will copy the contents of the important IMAP folders to my local system within Mail.app, as further redundancy, but that’s probably overkill.
Although I’m doing this much less than I used to (see The Solution below), I sometimes still like to make a complete clone of my system every once in a while using SuperDuper! and a Seagate FreeAgent Pro external hard drive.
Many people use SuperDuper! clones as their primary form of backup, as I did for a long time. I love the idea that at any moment, I have a full clone of my system ready to roll in case of a severe emergency. When it comes to cloning your Mac’s hard drive to a spare disk (external or internal), SuperDuper! is the best. It uses a technique called Smart Update to only update files that have changed, keeping the cloned drive identical to what’s on your Mac’s primary hard drive.
What you get is a bootable clone of your hard drive, as up to date as the last time you’ve run a backup. You can take this drive, plug it in to any Mac, and boot up from it. What you’ll get is your very own system, just the way you left it, ready to roll. From there, you can continue working right where you left off. This is really useful if your primary Mac needs a repair or had a total drive failure, and if you have access to a spare machine.
You can also use SuperDuper! to clone data onto a replacement Mac from a Mac booted into Target Disk Mode or from a cloned external hard drive. For people who like to preserve everything, this is often easier than setting up a new machine from scratch and can be faster than using Apple’s Migration wizard or a Time Machine restore.
What you don’t get is an incremental backup, where changes to files are preserved, or files you’ve deleted are still available. This is where Time Machine comes in.
Apple’s Time Machine is a really nice backup solution that ships as part of Leopard. For people who don’t want to think at all about backups, Time Machine is amazing. It just works, maintaining a copy of your entire system and updating things automatically as you go. It’s even smart enough to pick up where it left off if you have a laptop and disconnect from your backup drive, too. For most people, Time Machine is all you’ll probably need.
I threw an old SATA drive into one of the spare bays of my Mac Pro, and let Time Machine update it as it pleases. I don’t really need to have multiple copies of a single file, saving each revision, except for my code and writing, which are in Subversion, but I had the drive and the bay, so I figured why not? I have used it to restore a few files here and there, and it worked as I expected it to.
I wouldn’t want to use Time Machine to restore my data to a new computer if I was upgrading or replacing current hardware because I’d be forced to do it a certain way, and because it seems like it might take a while. I’d still prefer a clone with SuperDuper! or a manual copy. And while Apple’s previous backup solutions were sometimes problematic, I’ve heard enough good things and positive restore stories to believe that Time Machine might just work as described. My friend Duncan tried a Time Machine restore and it seemed to work just fine.
For my wife and parents, Time Machine is just right, and I can relax knowing that should something happen, they have a good way to get their data back … and I don’t need to babysit the process. I can just plug-in an external drive, let Mac OS X use it for Time Machine when it prompts, and walk away (fingers crossed).
I have a small collection of documents that I like to keep around. There isn’t much, these days, but there are a few things I wouldn’t want to lose. I rarely add to this collection of files, but I need to keep them around. If I were a designer, architect, or something like that, and I wasn’t using subversion, I bet I’d have a lot more of these.
Photos, Video, and Music
Like any modern parent, I take several hundred pictures of my new baby boy every single day. Each picture takes several megabytes of storage, and none of them, not even the poorly framed, out-of-focus ones, will ever be deleted. If you’re a parent, especially a new parent, you know what I mean. And don’t even talk about the video.
What I wanted was a solution that would grow as my library grows, but it had to be simple.
Enter the Drobo. After talking about it at length with a few friends, most notably John Gruber and Andy Ihnatko on Episode 19 of The Talk Show, I decided I’d give it a try. On the company website, the Drobo is described as “fully automated data storage that ensures your data is always protected, your capacity is unlimited and requires no special knowledge or expertise.”
Sounds too good to be true, right?
Well, in fact, it really is just that great. Unlike the RAID systems I’ve looked at (and forget about the dozens of RAID servers I’ve configured and built over the years) which require some degree of configuration and often come with a heavy price tag, the Drobo offers truly redundant storage and is simple to use and easy to setup. And it’s affordable.
Just turn add a few drives, between two and four of them, turn it on, plug it in to your computer, and away you go. The Drobo intelligently spreads your data across all of the available drives. If a drive fails (or if you remove it), the Drobo reconfigures itself to keep your data safe. If you add a drive, your storage instantly increases. In fact, this is how you upgrade your storage capacity. It’s unlike other RAID devices because you don’t need to add drives of the same capacity. Just give the Drobo whatever drives you happen to have handy, say, a 750GB Hitachi, a 500GB Western Digital, and a 250GB Seagate, and it will use them all intelligently, giving you the most redundancy and the most space it can. And as prices on larger drives fall, you can pick one up, remove an old drive, add a new drive, and behold: an instant increase in storage space. And you can do all of this without any interruption. You can continue to work, creating, moving, or deleting files as you please as the Drobo works in the background.
It’s quiet, runs cool, and although it’s USB, I can’t really tell a difference in speed when compared with the same files stored on my primary, internal drive. Maybe high-end media people doing video editing or drive-intensive media work might be able to tell, but for me, it’s just fine. It just sits on top of the Mac Pro and does its thing. I can easily connect the Drobo to any other Mac, or use the DroboShare to make the Drobo a stand-alone NAS device.
So how am I using it?
I’ve moved my iPhoto library, Lightroom catalog, iTunes library, and my main Documents folder to the Drobo. I suppose I could get fancy and mount the Drobo as my home folder if I wanted, but that’s not without its issues. I could probably also create aliases (or symlinks) for the folders in my home folder too. But I’m content to just leave things as they are, keep things in folders right on the Drobo, and access them directly.
Reader Request: If anybody can tell me the best way to easily force iMovie 8 to store everything somewhere other than the default locations, please let me know in the comments.
I’m still cloning things to my external drive once in a while, just to be on the safe side, and letting Time Machine do its thing to that internal drive. Even though I trust the Drobo, having a backup I could easily take offsite is pretty handy. Many people use the Drobo as their Time Machine drive. And it could be used as a SuperDuper! clone too. Mac OS X treats it just like a normal USB drive, and that means it’s bootable.
I’d be curious to hear about how you’re backing things up, so be sure to leave some comments and share your thoughts and experiences.
Truth be told, as much as I really do trust the Drobo and my other drives, there’s still a part of me that wants to create DVD copies of everything for storage in an ETL Verified 1-hour fire protection (up to 1700°F), waterproof safe. I’ll be sure to let you know how that goes.