Regular readers of this blog know that I had no recent backups of the sites I hosted when they went offline. I know, I’m an idiot. The fact is that I managed to recover some backups after a while because a secondary server in my former hosting provider had a copy of my files at the time. They lacked the last two weeks of the content, but were far better than the ones I had at home. I also managed to find copies of the MovableType databases used in the sites, containing pretty much the same information but in a much more useful format. As soon as a new hosting provider was found it would be just a matter of upload the database, wouldn’t it? Wrong. As I found yesterday, Murphy decided to intervene in the affairs: the backup containing my two blogs are corrupted. I don’t know how I didn’t discover before. You are allowed to call me an idiot again. Ironically, all others backups are perfect.
The database backup is almost completely garbled. I recovered the category and comments data but pretty much everything else is useless. Luckly, the site backup was not as corrupted, and, as far as I know, has all the files generated by MovableType from the corresponding database. Some files are corrupted, but not in the weblog directories.
Now I will need to create a program to read those files, extract the information in them, correlate it, and rebuild the database. It’s not a complicate thing to do, but it’s boring and tiresome. Also, some information will be lost as not all that was in the database is in the files — like the exact date a entry was posted, for example. Anyway, until I write that program, the previous entries in this blog won’t be available. I apologize for the incovenience. To those who request things I wrote, I will e-mail them directly to you.
Moral of the story: don’t trust backups, especially those generated by automated tools. Never forget to test and replicated them in multiple locations. Some bytes go missing or get garbled in a zipped file and you’re toast. That’s one more lesson learned. I think.