running out of swap

[Paul Hendrick]

we have a server running a script every night, and every night our
memory usage jumps after it's run, eventually forcing the machine to
crash when there's no swap space left.
the script calls tar,gzip, postgres db_dumb and vacuumdb, then backs
up the files to a tape.
The memoy leaks are also occuring rarely even without the script being
called, which is the cause of the confusion.
I think it may be postgres leaking memory, but i don't know where or how
to confirm this.
does anybody have experience of this?

the main memory intensive process on the box is python(it's running
zope).
postgres 7.4.5
zope 2.7.5
python 2.3.4
kernel 2.4.20 custom build.

i'd really appreciate any thoughts!