We are using a pair of adsm servers with hacmp. In normal mode, each
of them has one adsm running. In case of failure, all filesystems are
mounted to the remaining server and he has 2 dsmserv's running, each
with it's own db. (We had to use different tcpports for admin and
http). It's a little tricky, and you need to double adaptors and you
need ssa disks, but it's worth the money.

If we (or the hacmp) recognizes the server crash, we are up with both
servers a on one machine a few minutes later. Our customers won't
notice it.

Roger