Well, yesterday was a busy day. I converted a client's old server over to a new box running CFMX in the morning and headed over to another client's location to help them try to recover some corrupted data. (They didn't have any recent backups that were good—their raid controller card had begun writing junk data to the drive, so things were in bad shape.) I was able to get them back up in running to a degree, but there's 2 weeks of data I wasn't able to recover. They're sending their box out to a data recovery firm and hopefully we'll be able to get that missing data back. I was a little disappointed I couldn't help them out better than I did, but that's why I've stressed to them in the past the importance of backing data up. This isn't the first time I've had to help the client get the box up and running after a hardware failure.
Anyway, the day was pretty hectic. When I was at the one clients location, a couple of oddities arose on the box I put live, so I had to come back to the office to fix those issues. Fortunately, nothing was major and the problems didn't affect the core audience. I still was hoping for zero down time and thought I had planned accordingly.
Something to keep in mind when transferring a server: check Google for links to your site and make sure that you haven't forgotten about any aliases or non-standard links. There was a function of the site left in there for legacy purposes (to map old links from a static version of the site built in the mid-90s) and a few of the variations weren't working. Since those aren't referenced internally, when running on my stress tests and bot scans, it never came across the problem. Anyway, a search on Google may have found some of the variations before the client's client did. :)
Comments for this entry have been disabled.