Re: Re: Slowdown problem when writing 1.7million records

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: "Gordan Bobic" <gordan(at)freeuk(dot)com>
Cc: pgsql-general(at)postgresql(dot)org, "Stephen Livesey" <ste(at)exact3ex(dot)co(dot)uk>
Subject: Re: Re: Slowdown problem when writing 1.7million records
Date: 2001-02-27 19:34:50
Message-ID: 7787.983302490@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

"Gordan Bobic" <gordan(at)freeuk(dot)com> writes:
> Are you using autocommit? Are you using fsync? You should probably
> periodically commit the data every few thousand inserts if you don't
> have autocommit enabled.
> [ various other speedup suggestions from other people ]

The thing that interests me here is the apparent slowdown as more data
is inserted. I don't see a good reason for that, and none of the
suggested remedies seem likely to eliminate it if we don't know what's
causing it.

INSERT per se should be an essentially constant-time operation, since
it's just slapping another record at the end of the table. Unless
you're using advanced features like triggers/rules/foreign keys,
it seems like any progressive slowdown would have to be blamed on
updates of indexes. But that's not normally a performance problem
except in pathological cases (zillions of equal keys for example ---
but Stephen's only index is a PRIMARY KEY, therefore UNIQUE, therefore
no equal keys). So I'm confused, and think this deserves more careful
examination as to *why* the slowdown, rather than just looking for
a quick&dirty workaround.

regards, tom lane

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Norman J. Clarke 2001-02-27 19:52:11 Re: Does PostGresSQL have this feature?
Previous Message David Lynn 2001-02-27 19:26:05 Odd behavior with views and numeric columns