Re: Low Performance for big hospital server .. - Mailing list pgsql-performance

From Dave Cramer
Subject Re: Low Performance for big hospital server ..
Date
Msg-id 41DC3047.2030806@fastcrypt.com
Whole thread Raw
In response to Re: Low Performance for big hospital server ..  (amrit@health2.moph.go.th)
List pgsql-performance
Amrit,

can you post

explain <your slow update query>

so we can see what it does ?

Dave

amrit@health2.moph.go.th wrote:
Today is the first official day of this weeks and the system run
better in serveral points but there are still some points that need to
be corrected. Some queries or some tables are very slow. I think the
queries inside the programe need to be rewrite.
Now I put the sort mem to a little bit bigger:
sort mem = 16384       increase  the sort mem makes no effect on the
slow point eventhough there is little connnection.
shared_buffers = 27853
effective cache = 120000     
 
  If I were you I would upgrade from RH 9 to Fedora Core 2 or 3 after some initial testing.  You'll see a huge improvement of speed on the system as a whole.  I would try turning hyperthreading off also.   

Now I turn hyperthreading off and readjust the conf . I found the bulb query
that was :
update one flag of the table [8 million records which I think not too much]
.When I turned this query off everything went fine.
I don't know whether update the data is much slower than insert [Postgresql
7.3.2] and how could we improve the update method?
Thanks for many helps.
Amrit
Thailand

NB. I would like to give my appreciation to all of the volunteers from many
countries who combat with big disaster [Tsunamies] in my country [Thailand].

---------------------------(end of broadcast)---------------------------
TIP 8: explain analyze is your friend

 

-- 
Dave Cramer
http://www.postgresintl.com
519 939 0336
ICQ#14675561

pgsql-performance by date:

Previous
From: amrit@health2.moph.go.th
Date:
Subject: Re: Low Performance for big hospital server ..
Next
From: Miles Keaton
Date:
Subject: Benchmark two separate SELECTs versus one LEFT JOIN