Cloud Backup, Episode III
My last post on the topic was more than a year ago though,
so I’ll forgive you if you’ve forgotten. Here’s a recap: originally I was using
a fairly traditional consumer backup service, ADrive.
This worked well because they’re one of the few services that provides access
by Rsync, which made it easy to run scripted backup jobs on my linux
server. Their account structure didn’t really meet my needs, however: you
pay for the storage you have available to you, not what you use. When I hit the
upper limit of my account the next tier up didn’t make financial sense, so I
About 15 months ago I moved my backups over to Google’s Cloud Platform. This gives me an
unlimited amount of storage space, and I just pay for what I use at a rate of
$0.02/GB/Month. This has been working well for me.
In December I
B2. They offer a service very similar to Google’s (or Amazon S3, or
Microsoft Azure, or any of the other players in this space that you may have
heard of), except they cost a quarter of the price at $0.005/GB/Month. There’s
even a free tier, so you don’t pay anything for the first 10GB. When I first
looked at them their toolset for interacting with their storage buckets really
wasn’t where I needed it to be to make them viable, but they’ve been iterating
quickly. I checked again this week, and I’ve already started moving some of my
In time, I plan to switch all my backups over. So far I’ve
moved my documents folder and backups of my webserver, which totals about
2.5GB. That’s nice, because it means I’m still within the free tier. The next
thing to tackle is the backups of all our photos and music, which combine at
around 110GB. That means I have to transfer 110GB of data though, which is
going to be a painful experience. I’m still thinking about the best way to do
it, but probably the direction I’ll go is to spin up a VPS and have it handle
the download of the backup from Google and the upload to Backblaze, then it
doesn’t hog all the bandwidth I have on my home internet connection.
The only other thing to think about with Backblaze is
versioning. Google offers versioning on their storage buckets, but I have it
disabled. With Backblaze there is no option (at least not that I’ve found) to
disable this feature – meaning previous versions of files are retained and
count toward my storage bill.
I’m torn on this. The easy thing to do would be to disable
it, assuming one of Backblaze’s future iterations of their service offering is
in fact the ability to turn this off. I’m thinking though that the smarter
thing to do is make use of it.
For me and my consumer needs, that will most likely mean I
put together a PHP script or two to more intelligently manage it, however. Some
past versions are nice, but some of the files in my backup are changed pretty
frequently, and I definitely don’t need an unlimited history.
Still, I’m very much pleased with the price of B2, and
watching the feature set rapidly improve over the past couple of months gives
me confidence that I can move my backups over and keep them there for the
long-term, because the transition from one service to another is not something
I want to put myself through too often.