Tom Lachecki

Like any half-assed amateur system administrator worth his Sainsbury's Basics table salt, I've been searching for a decent backup/restore solution for a while now.

Background

On my office desk, not counting work-related machines, sits a Windows 7 PC and a whole bunch of external hard drives. Manually copying my 20GB of digital camera photos (plus e-mail archives, invoices, letters and the like) over to one of those external drives every so often is a chore.

In addition, I maintain a remote Virtual Private Server for DNS, Web and Email hosting, not to mention a spot of software development (if ever I get either the time or the motivation, out of work). With a few clients' websites and emails sitting on that server, not to mention a substantial quantity of my own content (plus various system configuration files, which I'd rather never have to recreate), this server also ought to be protected from data loss.

My original VPS was hosted by BitFolk. Although it was unmanaged, the team at BitFolk provided a rsync-based automated backup service if you paid for the provision of a little more data storage. The archived data could be accessed by mounting the remote stores onto the local filesystem. It was a bit of a pain to have to raise a support ticket to add/remove paths for backup, but it worked well. Unfortunately, whilst BitFolk itself and its services are top-notch and cracking value for money, I found that some of the people attached to the project were not so pleasant in person.

Having since moved on from BitFolk to something a little more meaty (hosted by the excellent Clustered), with more disk space to spare I've been looking at setting up my own automated, easily-configurable data resilience "solution" that might do everything I want, for both my office PC and my remote server, with a minimum of fuss.

Attempts

Intermittently dipping into the project every couple of months, my first few attempts almost unwaveringly featured the use of Samba which, among other things, enables the mounting of Linux filesystems onto remote Windows systems as Shared Folders (though I also tinkered with doing the same thing over a maintained SSH connection, with SftpDrive).

Samba is a right pain to configure, but eventually I managed to sort it out. I went through a myriad of tools that each proclaimed to make simple the task of automating backups from one locally-mounted path to another: Bacula, whilst promising, went way over my head; and DirSyncPro had options that did what I wanted, but I could never seem to coax the [slightly buggy] Java interface into actually performing any useful tasks.

Almost as a last-ditch attempt at making life simple, I then tried installing an SSH server on my Windows machine, and scheduling rsync itself to run from inside Cygwin (much like my original VPS solution). At some point through attempting to configure Rsync for Windows I got frustrated and went to the pub instead, swiftly forgetting about the entire enterprise for several months.

The happy ending

Recently finding myself with a couple of hours to spare on a slightly rainy Sunday night, and still keen to avoid "backup to the cloud" tools (such as Amazon S3) that I don't really trust (yet which promise to make life somewhat simpler), I decided that the mission to solve this problem deserved one last chance at a Hail Mary.

After all, I'd just added another client to my collection of websites to host, and I'd lost one of my external hard drives: the one that had, until now, been providing me with the only data resilience that had worked in any meaningful way. Definitely time to fix the issue once and for all.

So, armed with tasty mash and a pretty decent steak, I began skimming (again) through Wikipedia's list of backup software, hoping that I'd previously skipped past an obvious life-saver.

And I had: the brilliant CrashPlan.

CrashPlan is available for a number of operating systems, and functions via internet log-in that can join multiple installations under a single banner, allowing easy transfer of files between any of your own devices. (There are also local back-up features and, for a subscription, the option to use CrashPlan's own online storage.)

My problem was completely solved after doing the following:

• Downloaded, installed and ran on my Windows PC;
• Created a user account from within the application;
• Downloaded, installed and ran on my VPS (thanks in particular to fluxbox, RealVNC server and TightVNC viewer for the graphical interface);
• Logged into my user account from within the application;
• On each machine, specified folders that should be backed-up, how often, and for how long;
• Set my PC to backup to my VPS, and vice versa, simply by choosing one of the devices that had automatically been added to a handy list by the joined login.
• Opened a port for CrashPlan through firewalls both at home, and in front of my VPS.

And that's it. My Windows PC backs up to my VPS, and my VPS backs up to my Windows PC. It just works.

I haven't yet had a reason to try out a major "restore" operation, but if the rest of the product is any indication, my data is safe and resilient as long as either my PC or my VPS exist in this world.

So it's nice that, roughly a month after previously realising that perfection in technology does exist, I have found myself in a similar position again by being able to highly recommend CrashPlan to everyone in the world. Highly.

And if you haven't bothered even considering a backup solution yet… Well, I know everyone says this, but do it. Now that I've proven how easy it is, there is no excuse not to.

It's referendum day in the UK: the day when all vote-registered citizens announce whether they are happy with the existing First-Past-The-Post electoral system, or whether they would prefer the Alternative Vote system.

Strong political divisions have emerged over the possibility of changing the electoral system in this way and, as usual, this has tended to overshadow the actual issue involved. Reforming an electoral system is a big deal and it almost seems like it's turned into a mud-slinging contest before it even left the ground. Still, that's politics.

The "No" camp in particular has released a significant volume of propaganda that seems to sidestep the issue somewhat, , whilst leaning on questionable logic:

The website no2av.org lists, as its reasoning, only scaremongering and sensationalism. Of course, this does not mean in itself that there are no good arguments against AV, but it does mean that the "No" camp hasn't used them, and instead has appealed to the electorate's common instincts with unexpanded rhetoric.

On the other hand, the "Yes" camp has been rather subdued, listing arguments well but doing so quietly. Without that sensationalistic "punch", it's a surprisingly mute campaign from a group fighting for such an extensive change: to reform the very basis of the UK's electoral system.

Yes to AV

Under a two-party system, First-Past-The-Post is great. You vote for your preferred candidate, other people vote for their preferred candidate, and from the results it's clear to see who's the most popular: whoever has the most votes wins.

Under the UK's multi-party system, however, it's not quite so simple. The winning candidate is the one with the most votes individually, but often the number of votes for the other candidates combined is greater than the number of votes for the winner.

The alternative vote claims to "fix" that by, when this scenario is detected at counting, redistributing votes from losing candidates based on voter-specified preference. Your vote will only ever count for someone that you stated you have at least some preference for, and if, say, your first choice is not counted then it doesn't really matter because that candidate was never going to win anyway.

In practice, it's a superior system that promises to be fairer and more representative.

From the "no" camp, the oft-heard assertion that "FPTP has worked well for centuries" is a strange one. The electorate consistently votes centre-left, yet the Conservative party has dominated UK politics for two-thirds of the time since 1900. Only twice in that entire century – 1900 and 1931 – did the Tories actually win a majority of the votes.

Likewise, the oft-heard assertion that "we should keep FPTP because only a handful of countries have adopted AV" is also logically unsound, without some metric with which to determine just how successful the electoral system has been in those that haven't. The US is not relevant as a two-party system is a different beast altogether; and are we really going to assert categorically that all those other FPTP countries have the optimal system?

No to AV

On the other hand, it may be unsavoury in principle that your fifth preference vote may end up bearing as much weight on an election as somebody else's first preference vote.

It's also a concern that (again, in principle) employing a system of elections that — and I mean this with the greatest of respect for my fellow countrymen — many people in the UK will not fully understand might not be entirely appropriate. Do we maintain a slightly broken system that everybody understands, or introduce a fairer system that nevertheless may (perhaps) lead to some people not really knowing what they're voting for?

The anxiety of some predicting electoral candidates watering down their policies to pander to the whims of potential second- and third-preference voters is also palpable. Though on the surface it may seem like a good thing that candidates should have to really impress (rather than simply shoring up their existing followers), we could end up with several candidates whose stated policies aren't really all that dissimilar. And, as any business competition lawyer will tell you, variety in choices is a good thing.

In particular, candidates for major parties may be tempted to pander to followers of minority parties more than usual in order to solicit their Nth-preference votes. In the worst case, this might involve BNP-friendly policies creeping into centre-right candidate policies. In practice, considering the extremity of this specific example, it's unlikely to become a tangible problem; however, it is a fundamental flaw.

Conclusion

Neither way is perfect. Both have their flaws. Democracy and society aren't perfect; we just have to accept that.

First Past The Post is non-representative but relatively simple, and aligns with our instincts of how a race "should" work. Clinging on to FPTP merely under the illusion that it already works perfectly is silly, but it may still be the best of two options.

Alternative Vote is fairer and mathematically superior, but is significantly more abstruse and may embody some moral quibbles as to whether an electorate who doesn't understand the voting system is really any fairer, in balance.

Make up your mind and make your vote. The referendum had to be fought fairly hard for and, whether you want to keep FPTP or you want to give AV a go, make your opinion known.

Bootnote

It's worth giving this article a read. It's like this one, but better-written.