Family Life

All of a sudden a hard drive on my computer got really slow. There were no error messages or anything, it just took a long time to copy anything to that drive. I noticed this a couple of days ago, but being busy I hadn’t really looked into it. Today I used some utility software to look at the internal workings of the drive and found a whole bunch of reallocated sectors.

With modern hard drives, the drive detects errors on the disk in the course of normal use. If it can fix those errors all by itself then it does so, and doesn’t say anything in the hopes no one will notice and the problem will go away. If it gets more errors in the same place, then it’ll copy the contents of that sector over onto a spare it has in reserve. It keeps track of the substitution internally, so the computer never knows a swap has been made. Once again, the drive doesn’t say anything because the problem is fixed, right? No point worrying the user unnecessarily.

My drive had almost 300 of these reallocated sectors. Technically the data on the drive was fine, but that many secret swaps told me there was something about to go seriously bad with the drive. So tomorrow I’ll buy a new hard drive and move all the data over.

The experience highlighted a couple of insidious traits common to humanity. First of all, the drive thought that if only it could mask the internal problem then nobody would ever know and it wouldn’t get into trouble. Unfortunately for me, sooner or later the drive would have failed catastrophically and I would have lost all my data. I would rather have known about it sooner. It’s OK for the drive to try and fix the problem, but it should at least let me know there’s a problem. The moral of the story is, Be sure your (hard drive’s) sins will find you out.

Secondly, I actually had software on the computer to continually monitor the hard drives to detect hidden problems such as this, and report immediately if there was anything suspicious. So why didn’t it let me know? It turns out that I had switched off the reporting function. The software beeped and protested at the slightest hint of anything out of the ordinary: the drive took a millisecond longer to spin up today. Yesterday it took a millisecond less. Today it’s warm, last week it was cold. Pretty much all these messages were irrelevant to the actual health of the drive, and became frequent, annoying interruptions. So I switched them off. Thus, when a real problem came along I wasn’t informed. The software had cried Wolf far too often, and I’d stopped listening.

This entry was posted on Monday, March 10th, 2008 at 10:20 pm. You can subscribe via RSS 2.0 feed to this post's comments. You can comment below, or link to this permanent URL from your own site. Your comments will be moderated if this is your first time, otherwise they will appear immediately. I reserve the right to delete inappropriate comments.