On Sun, 2004-12-12 at 06:13, Tom Lane wrote:
> Mark Wong <markw@osdl.org> writes:
> > I never vacuum during the test. Is it possible that all the updates
> > and inserts would affect this?
>
> That's bad; first because it possibly *is* hurting performance, and
> second because if it isn't, your results could legitimately be attacked
> as not representing the long-term-sustainable performance of Postgres.
> VACUUM is real, unavoidable overhead and so we have to account for it
> honestly.
Agreed.
There does seem to be a downward performance trend over the course of
the one-hour tests, fairly consistently across the tests I've seen. This
is from about ~4200 tpm peak to ~4000 tpm peak an hour later.
I suppose that could be the reason for some of the extended transaction
times - though I reported a clear peak in the txn freq/response time
graph (with delays of ~7s). If txn times were lengthening because of
vacuum, I wouldn't expect to see a peak, just a long tail on the
distribution (which we do see...)
--
Best Regards, Simon Riggs