RES: Lock Problem - Mailing list pgsql-sql

From André José Guergolet
Subject RES: Lock Problem
Date
Msg-id 3FFA1818C1550246933DEA311C39B07612304A@CGEXC03.compugraf.com.br
Whole thread Raw
Responses Re: RES: Lock Problem  (Scott Marlowe <smarlowe@g2switchworks.com>)
Re: RES: Lock Problem  (Andrew Sullivan <ajs@crankycanuck.ca>)
List pgsql-sql
Sorry, I have a table with 360 rows, in this table I control the state of machines on network:


Ip             State        StateDate
172.20.0.39        Running    2006-08-23 00:00:00
172.20.0.59        Running    2006-08-23 00:00:00
172.20.0.72        Running    2006-08-23 00:00:00
172.20.0.84        Running    2006-08-23 00:00:00
172.20.0.35        Running    2006-08-23 00:00:00
172.20.0.17        Running    2006-08-23 00:00:00
172.20.0.28        Running    2006-08-23 00:00:00
172.20.0.39        Running    2006-08-23 00:00:00
172.20.0.14        Running    2006-08-23 00:00:00
172.20.0.33        Running    2006-08-23 00:00:00
172.20.0.19        Running    2006-08-23 00:00:00

My system, checks if my script is running in each machine at this table, this table has 360 rows and has 50-100 updates
perminute in columns STATE and STATEDATE. 

I list this states with a webpage. This webpage updates the list every 10 seconds. My page executes only "select * from
machinestates".

If I stop the updates, I never get my page stopped at the select command.

I read about "DIRTY Transaction", is it the way?
I have another solution?

Thanks.
Att.
André Guergolet




-----Mensagem original-----
De: pgsql-sql-owner@postgresql.org
[mailto:pgsql-sql-owner@postgresql.org]Em nome de Andrew Sullivan
Enviada em: quinta-feira, 24 de agosto de 2006 17:53
Para: pgsql-sql@postgresql.org
Assunto: Re: [SQL] Lock Problem


On Thu, Aug 24, 2006 at 05:27:58PM -0300, André José Guergolet wrote:
> Hello all,  I'm using the Postgres 8.0 and my product creates 7
> connections at the server.
>
> One of my connections do an update at a simple table with 360 rows
> and I've got many table locks. How I can take more scability?

It's pretty hard to tell, given what you're telling us.  The UPDATE
will take a write-blocking lock on each of the rows for the duration
of the transaction.  You shouldn't have any table locks, unless
you're not telling us something.  You should post more detail.  Why
do you think you have table locks, to begin with?

A

--
Andrew Sullivan  | ajs@crankycanuck.ca
The plural of anecdote is not data.    --Roger Brinner

---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster


pgsql-sql by date:

Previous
From: Andrew Sullivan
Date:
Subject: Re: Lock Problem
Next
From: Scott Marlowe
Date:
Subject: Re: RES: Lock Problem