Saturday, March 15, 2014

Backups, backups, backups

I have over a decade of work plus a huge amount of raw data (including video) that takes up a lot of space. I also have financial records and photographs of every family event since 2002 in digital form. At the start of 2014 my home network attached storage (NAS) died prematurely (thanks for that Seagate). Luckily I had my old USB drives that I'd used prior to buying the NAS, so most of the data were recovered. But it is a rather sweaty-palmed moment when you dust off an eight year old hard drive and hope it will still work. They all did - but I was lucky.

My data fall into three broad categories:

  1. Replaceable, but annoying to lose: e.g. Copies of CDs and my digital copies of 12" vinyl
  2. Irreplaceable, and sad to lose: e.g. photographs of holidays, early drafts of my PhD thesis
  3. Irreplaceable, and a disaster to lose: e.g. financial records that I have a legal obligation to keep, projects that I'm currently working on (some of the latter are on Dropbox, but it isn't an appropriate place to put sensitive data and means exporting data from the EU, which has legal implications).
I investigated which NAS units got consistently rated highly. For obvious reasons I was particularly keen on reliability, but also something that would just work. I'm a psychologist, not a system administrator. I took the plunge with a Synology DS213j for £160. Let's breakdown that model number: DS for DiskStation, 2 refers to the number of hard drives it holds, 13 refers to 2013, and j means it is a cut down version - this is for home or small office use. The Synology NAS is easy to use. You log in via your web browser and it gives you a desktop environment that feels very familiar. Options do what you expect them to. I didn't need the manual much, but there is an easy to navigate help system. It easily passed the just works test. There are also loads of extra modules you can load in through an app store to do additional things such as displaying photos or hosting your own website on the device.

Like many NASs of this sort, the Synology comes without hard drives. I found myself steering clear of Seagate and went with two Western Digital Red 3TB drives. The Red range is designed for NAS and a bit higher end than your bog standard hard drive in a desktop. Having two drives meant they could be set up as a mirrored "RAID" array. So if one drive dies the other carries on quite happily. Just put a new one in and the data are copied back across - job done. Hard drives are cheap relative to the value of the data they hold so I thought 2 x £95 for the drives was ok.

That sounds like pretty good backup doesn't it? But what if I'm burgled and the device is stolen? That's all my data gone. Oh dear, back to square one.

The solution is to have more than one copy in more than one place. Having been very pleased with the Synology I bought another one! This time I went for the DS214se for £120. The main difference is a slower CPU inside and less RAM. That's absolutely fine for a single user, but you wouldn't want this serving a whole office or to load it up with too many of the extra apps. I got another WD Red 3TB hard drive. I've now installed this at my parents' house in Surrey (i.e. over a hundred miles from me in Birmingham) and it lives discreetly next to their Virgin Media superhub. So now I'm backing up my files to this remote NAS over the internet. (It would have been more sensible to do this on the same network to start with because uploading 3TB of data over my ADSL is taking a while!) I started with the most important data - all the copied CDs can follow in time.

I've also protected the NAS in my apartment by adding an Uninterruptible Power Supply (UPS). I went with an APC CS350 for £65. We've had a few power cuts in the last year and this will keep it powered up (along with the ADSL router) for a few minutes whilst things save and shutdown properly. The UPS can communicate and tell the NAS when it is running on battery (translate as "go into safe mode now"). 

So is that sufficient? Well I'm willing to bet it is more robust than most people's backup solutions. But remember that 3rd category of data I identified? That's the stuff I absolutely cannot lose. I wanted an extra level of protection. So I used an option on the Synology NAS to upload some of the most precious data to Amazon Glacier. This is what is referred to as "cold data storage" and very cheap to store (about $0.01 for 1GB per month). The reason it is cheap is that data are stored offline. If you want your data back it will take several hours before it is even accessible. This is good for archiving files that don't change very often (if ever). Putting a few GB in there costs me a few cents a month - literally pennies.

My next step will be to get a big USB drive and backup the NAS to that. I can keep the drive in my office at work and carry it home for a monthly backup. This is relatively cheap and low tech. The weakness is the human element. I get busy so I put things off. Plus monthly backups could mean the loss of everything that I've created in that time. I could backup more regularly, but I'm not sure I trust myself to do it!

Am I paranoid? Maybe, but my near disaster at the start of the year was a shock. Hard drives are mechanical devices, so the question is when rather than if they will fail. Yes, you can use Dropbox for some stuff but it is mostly good for a few GBs. The largest paid for option from Dropbox I could find was 500GB (so 1/6th of what I have) and that costs $499 per year. I've spent about £600 in total which sounds like a lot, but I hope it will last a few years and the data have a very high value to me. If one device fails totally I can immediately use the other instead.

A small scale version of what I've got might be a DS214se with a WD Red 2TB drive - which would be less than £200. Then get an external drive for backup and keep that at work.

If your computer hard drive failed right now, what would you lose? Scary thought, isn't it?