Oli Warner About Contact Oli on Twitter Subscribe

Murphy’s Law

Saturday, 10 November 2007 raid rant

If something adverse can happen, chances are it will, just to spite you. I’ve just had my first hard drive crash in years and it waited until just after I dismantled my RAID5 array. Just perfect. Strong and colourful language in this post.

Before I get too far into this post, I’ll issue the following warning: the language used in this post will be colourful to the point where you can see a rainbow of verbal filth; therefore if you don’t like bad words, it’s time to look away now.

There are many laws in the word; some of them useful, some of them annoying, but all of them have the habit of rearing their ugly heads when you least want them to prove themselves. Today, I was attacked by Edward A. Murphy, Jr and his cockgobbling Law:

Murphy's Law: If it can go wrong, it probably will

I’ve had a RAID5 array for the past few years. For those not in the know, this is a system of binding hard disks together in a group so that one disk can fail and you don’t lose any sacred data. I’ve had a close miss or two with data-loss in the past but thankfully my RAID has kept me and my data safe.

But I’ve just moved to Linux. I like it so much, that I’m staying here. It’s better in too many ways to ignore. The one large problem with my migration is that my RAID controller isn’t supported. I’m not too shocked because the RAID controller is a piece of third-word shit. Hell, even people in the third-world wouldn’t use this controller if they had a choice. Anway… Linux can’t see through it so I can’t access my RAID array, which means I’m missing a lot of my data. I had to dismantle it.

Taking RAID apart is a very delicate procedure. You can’t just unplug everything and expect everything to be happy. You need to move all the files off it (if you want to keep them), so I set about culling all the crap I didn’t need: I went from 417gigs to 230gigs — small enough that I could fit it on one disk from the array. So I removed on of the disks, put it back in the system as a single drive and copied the surviving files onto the single drive.

So yes. I now have all my most important files sitting on one disk. If that dies, it’s all gone. A single point of failure, that were it to fail, would, in all probability, cause me to have a small breakdown.

Ignoring my system’s frailty, I plundered on, wiping my RAID out and reformatting the disks for use in Linux. I spent some time setting up another RAID array from inside Linux to utilise the two disks. Why? Because data-loss is a fucking cowbitchwhore. While this was happening, one of my disks make a horrific clunking noise. I ignored it and continued to hack around getting the new RAID sorted out.

I got the RAID created and started to copy things off the single drive. I got my documents, my source files, my backup ISO files and my downloaded TV. A good 130gigs. Then things got shit…

I’ve got to hand it to Linux here, when a disk starts misbehaving it doesn’t muck around like Windows, crashing left right and centre. It just rips all access to the drive and puts it to sleep. Very elegant but similarly confusing. I noticed something wasn’t right when a copy command stopped working. I was copying something from the network onto the drive at the time so I assumed it was just a case of a broken network connection.

After a short trip to IRC to ask people that know more than me about how to diagnose these things, it was clear that there was something seriously wrong. I want to let you in on my thought process. It went something a bit like this:

FUCK!

I restarted to see if it would pick it up after a reboot and it didn’t even make it past the BIOS screen. It’s all fucked up. On the plus side, there’s only 100gigs of data that I’ve lost. None of it was particularly rare, but it’s going to be a bloody badger in the eye to replace it all. I’m going to leave it a day or so (to calm down) and I’ll try it again (with some banging/freezing/etc) to see if I can’t whip the 100gigs of schniz off but I’m not optimistic.

The most annoying thing is that I’m down by 300gigs of storage until I can get this drive RMA’d. If only I had disposable cash.