Forum OpenACS Q&A: Response to Postgres, distributed databases, two phase commits, or other replication strategies....

I've been doing some work using DRBD

http://www.complang.tuwien.ac.at/reisner/drbd/

Which basically gives you a Network Raid 1 (one server primmary r/w and one secondary r/o).  Using heartbeat you should be able to create a h/a type setup on the database.

I currently using on a NFS server pair for my mail systems so and it handles its daily beatings fairly well.

Of course, since you need to fail over, it can take about 30-60 for the fail over to work so you could be done during those times.

Also, these servers are on the same network.