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