I have had two acs installations running on the same server for a month or so. When I go to add a third, the main installation can't talk to postgres:

[27/Jun/2000:18:44:23 -0400][564.635][acspg] Error: Ns_PgOpenDb
(postgres):  Could not connect to localhost::kg:  connectDBStart() -- 
connect() failed: Connection refused
	Is the postmaster running at 'localhost'
	and accepting connections on Unix socket '5432'?

The main installation preceeds the new one in the rc.local file, but can't make an initial connection to postgres. Also, this new installation does not start successfully at all, theres no AOLserver listening at its address. The above is true whether the second of the three servers is running. The first and the second will run together okay if the new one is not present.

The new install makes 10 connections each for main, subquery and log to its own db, "shebang". It seems to die and bring postmaster down with it. Doing ps -A, I count 13 nsd processes and no postmaster, trying to psql yields:

[bmueller@www bmueller]$ psql shebang
psql: connectDBStart() -- connect() failed: Connection refused
        Is the postmaster running at 'localhost'
        and accepting connections on Unix socket '5432'?

Restart with the new site removed, and psql and the other two acs sites connect with postgres fine. I know more info is needed to solve this, but not what info; I don't see how this process is different from the other two that have been co-existing peacefully.

Try starting the postmaster from a terminal session so you can see what's being logged to stdout/stderr.

One possibility is that you're firing up too many backends.  The default limit's 32, and can be increased by a command line switch to a large number.

If it is running hitting the limit, though, I don't think it's supposed to kill itself.  The error log should give you a better idea as to what's going on.