I own a number of computers and, like most people now have a network to connect them all together (except I was running networks back when coax-cable was the in thing to use). The system works quite nicely given it is made from junk parts. Since the estate agents didn’t visit today I used the time to sort out my server and tidy up its storage arrangements.

I think I’ve found a cheap way of backing up the large amounts of data that live on my server.

(more…)

I have a 1gig microSD card that goes in my Nokia 800. It goes in an SD card converter, and because of the design it’s really easy to pull the card out of its converter, leaving that inside the SD slot. Today while removing the card to fill it with music I did just this.

This seems to do really bad things. Either my Nokia was still trying to unmount the card, or by pulling it out in this way the card got corrupted. Either way, after that the card just stopped working. I put it in my PC running Linux and it denied its existence - it even went and turned off the card reader, claiming it was malfunctioning. Putting it in to my Windows PC gave the old “this card is not formatted” message, a “Delayed write failure” message and then half of Windows locking up.

It wasn’t until I pushed the card into my digital camera that I got any sense out of it. My camera bleeped and said the card was unformatted, would I like to format it? Yes, I most definitely would! And it did, no hassle, no errors. It also then quite happily took two pictures.
Then it crashed. Yes, my digital camera crashed. It’s like the card was spreading some sort of card reader virus around. I took the card out the camera and poked it back into my PC. Evidently something good happened because my computer wanted to show me the photos on the memory card.

So you see, sometimes it’s good to have dumb devices that don’t really understand filesystems, FAT tables and other computer concepts.

I wonder how many of these “dead” memory cards and USB drives people own aren’t really dead, they’re just really really corrupt and need a load of zeroes writing across them to sort things out?

I’ve just managed to wipe out half of my latest University assignment. And how did I manage this amazing feat of stupidity you’re probably wondering? Well…

At school I have a place on the network to store files. At home I have a place to store files. I also have a USB pen drive with files on. That’s lots of files and no easy way to copy them around without overwriting them. So I found the rather handy SyncToy from Microsoft and use that to keep everything in sync.

It seems that one day the sync went the “wrong” way and wrote the old version of a file over a newer one. You may pretend there is a long line of expletives just here ->

So backups are great, they stop total data loss. Now what I need is a versioning system so I can roll back things when they go titsup. I wonder how I can do this without there being any manual stage. SVN would work, but I’d need it to automatically update the repository as I won’t remember. I need my home system to always contain all files, in whatever state they’re in.

I’ve just signed up for 10 gig of storage at rsync.net it does exactly what it says on the tin. I have a large amount of space and can simply rsync stuff to it. No farty web frontends, Windows-only uploaders or crap like that. Shovelling the data up there is taking a while though, I hope I never have to do a full restore!

I was going to use a computer at my parents’ house, but it didn’t work properly. They have a dynamic IP address so finding it on the Internet was tricky. Not helped by their crappy router disconnecting at least once a day.

Once I’ve done the initial upload, subsequent backups will be quicker. I’m Going to have to make some sort of redundant backup at home too. I have quite a lot of data that would either take too long to recreate or be lost forever if my server died. It may seem a little excessive having 10 gig of offsite storage for a bunch of digital photos, text documents and source code, but if you’ve ever lost data due to a machine breaking or pilot error it’ll make more sense. I could just burn the lot to DVDs, but that’d require me to be bothered, remember somehow what I’d already backed up and to verify the DVDs still work later on. I have CDRs in my collection that are either scratched, delaminating or just don’t work very well any more.

Off site storage is the way to go. Google holds most of my mail, my RSS subscriptions and my Firefox browser status. I have my bookmarks on del.icio.us and this diary lives on my webhost (which is different to my ISP so I can switch ISPs without too much hassle).

What’d be really good is a BitTorrent style backup system where everybody on the Internet holds some data from other people. Rather than uploading data to a specific machine, the network as a whole could take it from you and give some back. It’d all be encrypted and turned into little meaningless blocks of data. With several billion people on the Internet we’d only need to hold a few gigs of data each.

powered by performancing firefox

I have an external USB HDD for my laptop to backup the HDD in my laptop. Every half hour important data is copied from my laptop to the USB HDD. Some of this important data is all my photos I’ve ever taken - all five gig of them. As my laptop’s HDD is getting full I decided to permanently move the photos onto the external drive.

Today I turned my machine on to show someone some of the photos and… they’d all gone! Never mind they’ll be on my external HDD. Oh no they’re not there either. Err shit.

What I forgot is that the backup script runs this command:

rsync -az –delete /home/james/photos /mnt/brick/photos

Notice the ‘–delete’ part? That means “If the source file gets deleted, the destination file will too”. And guess what, if you move everything from ‘/home/james/photos’ into ‘/mnt/brick/photos’ by hand, they’re no longer there meaning the contents of ‘/mnt/brick/photos’ gets synchronised by having its contents deleted.