The bad thing about this is that on a production server, such as the one we had problems with, restoring from backup does not do any good because the backup runs at 2am and the corruption occurs at 8 pm, in between we have had 8 signups, 2 cancellations and perhaps a whole bunch of users making changes, such as adding subdomains, etc.
I guess the only way to be safe on a situation, such as this one is to run a shell script via the cron that copies httpd.conf into multiple locations, so in case this happens again we can just copy the last non-corrupted file.