Did you know that hard drives are rather like light bulbs? They have a finite life and are simply not expected to last forever. At some point, the hard drive inside your computer will fail.
If you are reading this blog, you probably have pictures from your walks and climbs in beautiful mountains. If they are only stored on your computer’s hard drive, you would lose the lot.
By not backing-up you’re simply gambling that the hard drive will fail after you’ve moved on to a newer machine, whether it’s MS Windows, a Mac or Linux.
Losing my pictures would be awful but there’s plenty of other data I’d rather not lose; emails, letters, source code, old academic work. Horror stories of losses abound on the web and I personally know people that it has happened to.
If you backup already, to DVDs or an external hard disk perhaps, but keep that backup in the same building as the computer, you have no protection against the worst case scenarios of fire/theft. I’ve heard it said that people who lose their house in a fire or other disaster come to terms with the loss of everything, except for their pictures.
Backing up on to DVDs or even better, on to external hard drives is a great idea. Nothing is faster for restoring a backup from than a disk in your hand. But you still risk finding that the backup disk has simply died. They all do, eventually.
As long as you have broadband, online backup is a no-brainer. It’s easy to set up, cheap, and once it’s done, it’s done. No need to remember to start the backup, or swap the disk. Nothing more to do – ever.
It can take quite some time to back up initially; days or even weeks if you have a lot of data. But so what? As long as it doesn’t need looking after, it doesn’t matter. And once it’s been done once, only the changes are uploaded.
Its not expensive – especially once you consider how much you’d pay to get your photos back if you were to lose them.
There are many online services and I’ve tried several of those that get better reviews.
Overall I’d recommend Jungle Disk (see below) but I must admit to having been impressed by both CrashPlan Central and BackBlaze.
Edit: 23 May 2010: It struck me that whatever online backup solution you use, the passwords MUST be kept off-site somewhere so that you can access the backup from scratch on a new machine (say, in case your place burns down!). The best way to do that is probably to print them off and keep them somewhere with a friend or relative you trust.
I’ve been using Jungle Disk backed by Amazon S3 (not Rackspace Cloud Files) for several years and can’t find anything that ticks all the same boxes.
But my criteria are perhaps more demanding than some. Apart from checking for data-corruption, encryption and the high bandwidth that any good backup provider will have, I also want data stored in more than one geographic location looked after by a company that’s large enough to be unlikely to go bankrupt.
Amazon S3 writes every file to multiple different geographic locations (and then immediately checks each one to ensure the write was correct).
Most people would probably be happy with BackBlaze or CrashPlan (not Mozy – see below, or Carbonite – which I won’t waste anyone’s time even mentioning further).
The only thing that puts them in front of Jungle Disk is that they require slightly less set-up and they offer a fixed price per month option. Whereas Jungle Disk varies in cost depending on how much data you store. Though for storing less than 10gb of data, Jungle Disk is comparable per month – they just don’t do yearly billing unfortunately. Meaning a forex charge on your card every month if you’re outside the USA.
So, assuming you’re not quite as paranoid as I am, and can live with the small risk of the datacenter being wiped out in a far-fetched disaster (and so aren’t going for Jungle Disk+S3); which one would I recommend?
Probably BackBlaze. Although I prefer CrashPlan’s interface, I prefer BackBlaze’s use of a large, trusted datacenter.
Jungle Disk: jungledisk.com
For me this is the only service that does the job – it’s the most full featured and flexible set-up I’ve found. I use the “Desktop Edition” at $3 a month (though the new “Simply Backup” version at $2 looks good too).
It backs up any kind of file without restriction, it can back up any number of machines, it uploads as fast as your connection can go (but can be “throttled” at times you decide, so your browser doesn’t crawl) and runs on MS Windows, Mac and Linux. There’s a 5gb file size-limit which might be inconvenient for some but is actually larger than most services allow.
I’ve used it for two years on three machines. I have over 70GB of data backed up and I have used the restore “for real” on several occasions. It just worked.
Why do I use S3 and not Rackspace Cloud Files? Because S3 use multiple, geographically distributed data-centres to hold each file. With Rackspace, each file is held in three separate physical locations in the same datacenter (geographically co-located). Albeit, each with a separate power supply and Internet connection – but that’s no defence against a plane/earthquake/localised alien invasion.
However, using Rackspace has a different advantage: You don’t have to set up a second account (with Amazon S3) since Rackspace own Jungle Disk. That also means one less forex charge on your card of course.
You pay for the Jungle Disk application and the S3 (or Rackspace) storage separately. And since it’s only about 15cents per gig per month you’re unlikely to have to pay very much.
As a side bar – I’ve made lists of a few folders and files on Windows XP that you might want to avoid backing up, that can be copied straight into Jungle Disk. For example, your Firefox and Internet Explorer caches are certainly not worth backing up.
(There’s another good overview at onlinebackupsreview.com)
Note: I’ve just discovered Cloudberry Online Backup which has a one-off license fee cost for Cloudberry and then uses S3 as the storage provider – that might be worth investigating since there’s no monthly payment (and forex hit!) but I’ve not evaluated Cloudberry so can’t comment on it directly.
Although overall I’d recommend Jungle Disk, I was impressed by BackBlaze.
It’s simple, it works, it seems to backup and restore just fine. It’s a flat $5 monthly or $50 yearly (which is great if you are outside the USA as there will be less forex charges on your card). It’s also very good value if you have a couple of terabytes to backup.
Although the “backup the lot” approach of BackBlaze is one of their main selling points, I don’t like the way it decides to backup almost everything by default. The interface to selectively exclude certain folders needed some work when I evaluated it in late 2009 (if you can’t easily tell it to exclude things, it’ll use up bandwidth backing up junk – and make any restore slower than it needs to be). There’s a 4gb file size limit but as I mentioned above, that’s quite common.
BackBlaze reviews occasionally mention that they are new, but they’ve been running since early 2008 as far as I can ascertain, so it’s not all that young any more.
However, although they use a very large, dedicated datacenter (which is also used by Sun Microsystems and Cnet) it still means your data is only held in one geographic location. No plane/earthquake/alien insurance there. But you may be less paranoid about such things than I.
I tried backing up a couple of different data sets consisting of a couple of gigabytes of data and then restored them a couple of times. All ran very smoothly.
(Discount code at onlinebackupsreview.com)
CrashPlan Central: crashplan.com
Again, I’d recommend Jungle Disk but, CrashPlan comes at a better price than Jungle Disk and has a nicer (more configurable) interface than BackBlaze.
Unfortunately they run their own, single, data centre. Impressively specified though it is.
Single geo-location for a backup is not unusual, quite the reverse, but at least BackBlaze use the services of a very large and dedicated datacenter provider rather than doing it themselves.
CrashPlan say that it “Supports files larger than you’ll ever need”. Which may be true, but you might want to check what that actually means if you have very large files.
However, the ability to configure exactly what is backed up via a very nice interface is very appealing.
All the backup and restore testing I did ran absolutely fine.
I’m sorry this section is so long when it only describes what I wouldn’t use – but since Mozy is so popular I thought I’d better explain exactly why I won’t use it.
I really want to like this application for its nice easy interface and general simplicity – but it’s just not reliable. Even the rather partisan review on onlinebackupsreview.com mentions “Some users have reported problems when trying to restore data”.
A “problem restoring” is an absolutely massive flaw in a backup application.
I installed it and backed up 2.5gb of data a couple of years ago. It did have a habit of locking up during a backup but a restart of the machine would un-stick it (everything has bugs, bugs in uploading I can forgive). But then an upload just failed on a particular file and would get no further. It transpired the backup was corrupt on their server (a bug they have since fixed – apparently) and after a lot of to-and-fro with Mozy Support I was advised to re-upload. Not great if you have 60gb to upload and utterly fatal if you happen to have just lost your HD! So I stuck with Jungle Disk.
I installed Mozy again a year or so later (like I said, I want to like it) but after it installed it simply refused to start-up.
Now, I know what I’m doing, but pretending I didn’t, I asked Mozy Support for help (I need to be able to recommend this service after all).
I got the usual “Silly user! Here’s how to install the product…” email reply that most first-line tech support will fob you off with, after they’ve completely failed to read your email.
They took a little convincing that I had installed it but it simply wasn’t working. They eventually suggested removing it completely (and provided a thorough description of how to do that) followed by re-installing – which was the answer I was looking for.
I did, it worked, I backed up. I restored a set of files… fine.
I restored the same set of files again for good measure – error! What error? Well who knows!?
There was an “error2” reported at the end of the restore – and the log showed many of the files hadn’t been restored… except that when I checked with a binary file compare tool, they had been.
I compared both the first successful restore and the one reporting errors – they were identical. The files that were logged as having errors had actually been restored just fine.
So although it did work, it reported that it hadn’t and that’s no good if you really are restoring after a data loss and therefore can’t compare with the originals to see if it’s a real error or not. You simply should not have to be deciding anything at all regarding “errors”.
So, I’m sorry Mozy but an unreliable backup isn’t a great deal better than no backup at all as far as I’m concerned.
The above options are far more reliable in my experience.