Questions on production rollout

Hello,

Have heard great things about PostgreSQL as far as performance
and stability and am considering using PostgreSQL in a
production environment. However, I'm a little worried because
our product will be installed on client machines with various
configurations: shared memory params in /etc/system, patch
levels, memory, disk space, etc., etc.

I am wondering if somebody who has had some experience rolling
out PostgreSQL into a production environment might give me some
tips on where the major pitfalls are that I need to be aware of.
Links and/or FAQs would be great.

Just to give an idea about the source of my concerns:

1) A guy here had a compiled version of PostgreSQL for Solaris 8
that he gave me to install on my Solaris 8 system.  He said just
untar it and you're good to go.  I did, then I tried using the
psql command to connect his server that he already had up and
running on his system.  Got an error because I don't have a gcc
library installed.  So here's an example of an install that worked
on one machine but not the other.  Is it possible to build a
version that one can be reasonably sure is going to run on all
systems with a given OS?  I mean, if I know the client machine is
Solaris 8, can I build a self-contained set of binaries that I
can confidently say will run on other Solaris 8 machines?

2) I realize there are other configuration parameters involved.
E.g. already encountered "SHMMAX value to small" message when
trying to start server on my system.  I changed it to something
big, rebooted and it still didn't work.  I'm not too worried that
I can't eventually figure it out but it still has me a little
spooked.  Wouldn't want to put a client thru hell trying to get
the server up and running.  Are there other system configuration
gotchas I should be aware of?

Thanks in advance,

Bill