From: Thor Tall <tall_thor@yahoo.com>
To: pgsql-admin@postgresql.org
Subject: Slow delete/insert.
Date: Wed, 23 Aug 2006 07:30:17 -0700 (PDT)
Message-ID: <20060823143017.98255.qmail@web50301.mail.yahoo.com> (raw)
I am in the process of converting an old system to a
new system where I have chosen to use postgres in
stead of a home grow system based on b-trees.
The system receives 2650 message a total of 10Mbytes
of data per 15 minutes this information have to be
store in 4 tables in the database. Some of the
messages are updates to existing data and the result
is:
1038 records in table 1 size per record 282 bytes
2074 records in table 2 size per record 36 bytes
21488 records in table 3 size per record 60 bytes
25676 records in table 4 size per record 42 bytes.
Without indexes this should be something like
2,7Mbytes/15min.
The disk speed is as follows taken from hdparm -Tt
/dev/hda:
Timing buffer-cache reads: 128 MB in 0.89 seconds
=143.82 MB/sec
Timing buffered disk reads: 64 MB in 1.45 seconds =
44.14 MB/sec
My problem is that the postmaster process is using
between 70-90% of the CPU time. (seen using top) and
the total wall time with nothing else running.
It takes 6 minutes to ingest the data with an empty
database and 25 minutes (wall time) if all the data is
already in the database.
The processing is done as follows:
1. Start transaction
2. check if message is in table 1 and if so delete
records from table 1(1 row),2(2 rows),3(30 rows),4(50
rows).
3. split the message and insert the date in table 1(1
row),2(2 rows),3(30 rows),4(50 rows)
4. end transaction
I need some suggestions so that I can ingest the data
faster. with the same computer and disks. We are
running redhat 7.2 on the server and I am using the c
interface to write the data to the server.
Thanks for any help,
Thor.
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients from your mail client; the link fills in
the headers below: reply via email
Some clients, webmail especially, drop In-Reply-To and References from
a link, and the reply then starts a new thread: use the mbox below.
To: pgsql-admin@postgresql.org
Cc: tall_thor@yahoo.com
Subject: Re: Slow delete/insert.
In-Reply-To: <20060823143017.98255.qmail@web50301.mail.yahoo.com>
References: <20060823143017.98255.qmail@web50301.mail.yahoo.com>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox