I wrote a little while ago that I was running Linux (Ubuntu in this case) inside a VirtualBox virtual machine, and it was all good.  Before that I’ve played with lots of methods of getting my favourite unix utilities (like rsync) working under Windows.  I’ve used Cygwin, and pre-compiled Windows versions and stripped-down Cygwin versions, and second machines running Linux and VM’s.

One of the main drivers for getting those things working is to back up my websites, held on my hosting account.  I can ssh into my hosting account, and that means if I can get rsync going locally, I can use it with ssh to copy all changes to my local machine.  It’s efficient (rsync only copies changes) and it’s easy.  The pain is always finding a decent compliant version of rsync.

Anyway, I already said that when I started using the Linux VM I ported my script across to that, and along with the VirtualBox shared folders, I could backup my websites and they were visible under XP.  It wasn’t pretty but it worked, and it meant I had to start up the VM.  At the start that wasn’t a problem because I was using it quite a bit but as the days went on and I stopped launching it, backups were less frequent.

And then today – random disaster.  I crashed the VirtualBox VM image, and after a couple of restarts it eventually stopped booting.  This wasn’t a great problem as I had snapshots of working images, so I just rolled back to one of those with two clicks.  Two clicks which took less time than the following thought took to get from one end of my brain to the other ‘I made the snapshots weeks ago, and since then I’ve written a lot of scripts and downloaded a lot of files and you just erased them all you idiot’.

So, I set about repatching Ubuntu and setting up various settings that I’d lost and made a few more snapshots.  But I needed a more permanent, reliable website backup solution.

Which means I’ve installed Cygwin again.  I know there are Windows binaries for rsync, and I know there are other apps which claim to do the same thing, but you can’t (in my view) beat the simplicity of Cygwin and the unix binaries.   Now I have a working cron daemon, ssh configured, rsync installed, and my little script which does all the work.  The rsync command is pretty simple,

rsync –recursive –links –safe-links –rsh=ssh –stats –human-readable me@mywebhost:/myhomedir/ /path/to/local/copy/

Then I just tar up the resulting files, compress them, make sure the filename has a date in it, and I can be confident I’ve got copies of everything I need.  Since most of my sites rely on mysql for their data, I also run some jobs on my webhost to mysqldump all the data into files three times a week, and I then back those files up locally.  I could mysqldump the content remotely, but it’s a hell of a lot quicker to do it on their system, compress them, and then rsync the compressed files.

Installing ssmtp lets me send mail from the Cygwin command line, so the script can send me a mail when it’s finished, and I’ll schedule it in cron to run once a week or something.  Much better.

Plus, I get all the fun of vi, grep and awk 🙂