[exim] Exim and clustering

Does anyone run Exim in a clustered configuration, whereby everything to do with Exim (config, queue, database, logs) is
stored on a filesystem/LUN which can be made accessible to a standby system? I'm thinking in terms of:

a) Two Linux nodes running heartbeat software and "mirroring" the Exim filesystem(s) using drbd. If NodeA fails, NodeB
mounts the filesystem(s) and starts Exim

and/or

b) Two Solaris nodes running heartbeat with the Exim filesystems stored on a LUN in a SAN. If NodeA fails, NodeB issues
a command to the SAN to re-assign the LUN to it before mounting the filesystem(s) and starting Exim.

"High Availability" seems to be the "in thing" here at the moment - especially since we've just clustered our Exchange
servers (apologies for swearing on the Exim list! ;) ) - and I'm toying with ways of implementing it with regards to
Exim, even though our Solaris+Exim systems have been extremely reliable, you know what PHBs can be like!

What sort of issues might one expect in such a configuration?

--
Paul Walsh

Systems Manager (UNIX),
Information & Communication Technology,
UCE Birmingham, BIRMINGHAM B42 2SU, UK