Received: from gcds4003.gcd.de (mail.gcd.de [194.162.100.194])
	by hub.org (8.9.3/8.9.3) with ESMTP id LAA70759
	for <pgsql-interfaces@postgresql.org>;
	Wed, 12 Jan 2000 11:33:22 -0500 (EST)
	(envelope-from rluettecke@gcd.de)
Received: from gcdw4129 (gcdw4129.gcd.de [192.168.4.129])
	by gcds4003.gcd.de (8.9.3/8.9.3) with SMTP id RAA04326
	for <pgsql-interfaces@postgresql.org>; Wed, 12 Jan 2000 17:32:51 +0100
Message-ID: <009201bf5d1c$60bf9920$8104a8c0@gcdw4129.gcd.de>
From: "=?iso-8859-1?Q?Rolf_L=FCttecke?=" <rluettecke@gcd.de>
To: <pgsql-interfaces@postgresql.org>
Subject: Slow down on Import
Date: Wed, 12 Jan 2000 17:45:05 +0100
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 8bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4

Hi folks,

first some techn. data: SNI-Primergy w/ PII-333, 128Meg and
SCSI-devices, Linux 2.0.36
The following strange behaviour comes up since upgrading to
version 6.5.3 + FreeFile-Patch (hi Tom):
When re-importing a dumped database the job is normally done
in roundabout 8 seconds. When you drop the database and
repeat the step a few times, the speed slows down to at least
1.5 minutes.

Example:

# for i in 1 2 3 4 5
# do
#   destroydb xyz
#   createdb xyz
#   time psql xyz < ./import.db    (small table / 10.000 rows)
#   (output from time: 8.2 seconds up to 1.5 minutes)
# done

There is no additional load on that machine. The main application
(MS-Access w/ ODBC xx.0006) seems sleeping during database-
search - so we tried to figure out the problem and made at least
the above test directly on that machine. Last hint: A complete
copy from "/opt/pgsql" transferred to a another machine shows
the same problem.

What´s wrong? Any ideas?

Regards
 -Rolf