Sounds like you were bitten by the technical debt that was created by placing an important database behind a single point of failure - a single physical server.

Sounds like you were bitten by the technical debt that was created by placing an important database behind a single point of failure - a single physical server. Of course something would break sooner or later, PA-RISC and Itanium servers were great, but still not magic. Important things go onto clusters, or at least have a (hot or cold) standby server.