Amazon Glacier saves my SkyDrive, or why you have to backup your cloud storage in the cloud

Restoring-SkyDrive-via-AWS-Glacier
Click to enlarge

Lots of people think that if they have a cloud-based disk sharing system (I like SkyDrive), they can’t lose data.

Well, I am here to tell you it can — and does — happen.

Over the weekend, I upgraded my MacBook Air to OS X Mavericks. On this Mac, I run Microsoft’s SkyDrive client that syncs all of the files (except music and photos) that one would keep in one’s “documents” folder. I also share that OS X folder with the Windows virtual machines I run on the MacBook. That lets SkyDrive replicate the changes I make on my Mac and/or in Windows VMs to the real Windows machines I also use. Then, I simply set Office 2013’s default directories to the SkyDrive path and as a result, have a unified view of my filesystem no matter which machine I am on.

It all works great. But sumthin’, I don’t know what, caused the SkyDrive client on my MacBook Air to actually delete files more or less randomly from the MacBook SkyDrive store after I upgraded to Mavericks. By the time I discovered it, the deletions had replicated. Ouch, ouch and triple ouch.

Not to worry. Though some of my friends laugh at my retentive backup habits, this is precisely the situation in which a backup of a cloud backup is what’s needed.

I’ve been using CloudBerry for a couple of years to backup to Amazon Glacier. The way I have it set up is that CloudBerry is running a real-time backup and watching the SkyDrive folder (as well as folders on my NAS which store my music and photos). That way, as soon as a file is updated, CloudBerry uploads it to Glacier. I’ve set CloudBerry to keep only the last three versions of files and to delete what’s removed from the filesystem after 90 days. With over 150GB stored on Glacier, I think I pay about $1.50 a month or so for storage — and upload bandwidth is free. Better yet, CloudBerry encrypts files before they are uploaded to Glacier (see the screenshot nearby to see how the files being restored have “garbage” filenames on the way back down). So, the NSA can get the files from SkyDrive but not from Glacier.

With this setup you have 90 days to discover that the oldest file is missing. Why 90 days? Well, if I don’t want something in my SkyDrive, I simply move it to a folder that CloudBerry is also backing up called “Cold Storage.” That’s the stuff I really want to keep forever in Glacier.

Fortunately, I discovered that SkyDrive trashed my files only hours after the calamity-that-isn’t-a-disaster-because-I-have-a-backup.

Getting files back from Glacier is slow, as you can see from the screenshot. But the cost is so low and the convenience so high, I don’t mind waiting four or five hours to restore files.

The moral of the story: anyone who tells you they don’t have to worry about losing stuff because they have it stored in SkyDrive or Dropbox or Google Drive is simply fooling themselves. Any system that replicates changes can also replicate mistakes and probably will. That’s why you need a cloud backup of a cloud backup.


Posted

in

,

by

Comments

One response to “Amazon Glacier saves my SkyDrive, or why you have to backup your cloud storage in the cloud”

  1. […] written before about CloudBerry (also here and here) and was pleased to meet with Director of Marketing Alexander Negrash at the conference for an […]

Leave a Reply

Your email address will not be published. Required fields are marked *