Re: insert performance riddle

On Thursday August 11 2005 12:36 pm, Ed L. wrote:
> On Wednesday August 10 2005 6:03 pm, Michael Fuhr wrote:
> > On Wed, Aug 10, 2005 at 05:02:46PM -0600, Ed L. wrote:
> > > I have two identical servers giving abysmal INSERT
> > > performance in pgsql 7.3.4, 7.4.8, and 8.1devel under no
> > > load or I/O contention at all (no dumps, no vacuums, no
> > > apps, etc).  Any suggested investigations appreciated...
> >
> > Have you done any tests that eliminate the client code and
> > client-server communcation?  Something like the following:
> >
>
> Michael, you seem to have nailed it.  The local inserts (via
> Unix domain sockets?) that were running at 6 QPS ran at 6800
> to 41000 QPS in a PL/pgSQL function.

Here's another part of the riddle.  The query durations for the
individual inserts as logged in the server log are ranging
100ms-500ms per insert when connecting via DBI/DBD::Pg from
localhost.  Inside pl/pgsql, they're clearly averaging faster
than 1ms per insert.  I would have thought those times were
independent of the client communication layers...?

Ed