Today I was doing some database work for a client and I had to delete some recent entries from a database. Like a moron, I forgot to add a “where” close and emptied the entire thing! No big deal, I’m the one who wrote the backup script so I went to grab the backup for yesterday. Yup, you know it. There wasn’t one. The backup stopped working back on July 3rd! Ouch. That burns. Luckily the host is also a good friend of mine and I shot him an email asking if he had a recent backup of the raw MySQL files. I knew that he did regular backups of the entire server so the raw files should be there. I don’t count on that, you should never count on anyone elses backups, but in this case it saved my bacon. Gary had the backup and gave me the quick instructions on restoring them, which I’ll share with you:

• Get a copy of your raw mysql file. In my case, I didn’t want the whole database, I wanted one table from one database. Luckily MySQL stores each table in a 3 files called:

• database_table.MYI

• database_table.MYD

• database_table.frm

So, once I located them – Gary sent them to me – I just needed to restore them. It’s pretty easy.

• Stop mysql
• cp database_table.* /var/lib/mysql/ That’s the typical path, but of course it may vary from host to host.
• Restart mysql

And if everything went well (it usually does) you’re back in business.

Caveat: This won’t work if your database type is innodb. Ugh. The disturbing thing to me is that my automated backup stopped running for some reason. It looks like the process was killed at some point and never restarted. You should add “check that backups are working” to your weekly, or at a minimum, monthly checklist. Never ever count on your host or someone else to do it. Murphy’s laws are still very much alive and in effect.