Out of our three disks, one had a readable file-system, with a data back to the 20's. So we've lost two weeks of work. Our backup mirroring was just good enough, but we didn't have the redundancy we won't for real safety. Because we're now caching so many web pages, we fulled the primary hard drive so much, that backing up would fail, plus two weeks of the pressure of having a disk being pushed to full up all the time, finally made the primary drive fail. God bless rsync. What was really bad of us, is that we didn't have a copy of the database of the feeds anywhere else except the server. They should have been copies back to our development system made at least once a weak.
Going to improve our backup methodology next time. Onwards with the recovery.