recurring problems requiring a vacuum

Our web/db application is to the state where we are using it to load in a lot
of content.  The system seems to run fine for a while, and then bogs down with
postgres processes chewing up lots of CPU time and nothing happening.

When this happens I vacuum the database, and it seems to fix the problem.  The
trouble is, it keeps recurring.  This last time I vacuumed, I got the following
message (among others):

    NOTICE:  Rel pg_proc: TID 8/12: OID IS INVALID. TUPGONE 0.

This message keeps recurring, so I assume it is just continuing to warn me that
some tuple is gone.  (From some internal table?)

My question is, why does the db keep grinding to a halt and requiring vacuums
in this way?  The db structure is faily straightforward; each table is of the
form:

    create table foo (foo_id serial primary key, ...);

and I have not created any indices other than these.  There is not a lot of
content in the database; ie, things are still fairly small.

I am running Postgres 6.5.2 on FreeBSD 3.3-STABLE.
The web application uses PHP 3.0.12 on Apache 1.3.9.

Any hints or pointers appreciated.  Thank you!

Charles

__________________________________________________
Do You Yahoo!?
Thousands of Stores.  Millions of Products.  All in one place.
Yahoo! Shopping: http://shopping.yahoo.com