* Klint Gore <[EMAIL PROTECTED]> wrote:
> Turn on statement logging. I've seen delphi interfaces do extra queries
> on system tables to find some structure information.
I'm already using statement logging of all queries taking longer
than 200ms. It seems that only the INSERT takes such a time.
On Mon, 4 Jul 2005 10:59:03 +0200, Enrico Weigelt <[EMAIL PROTECTED]> wrote:
> * David Mitchell <[EMAIL PROTECTED]> wrote:
>
> Hi,
>
> > Hmm, you said you don't experience this when executing the query
> > manually. What adapter are you using to access postgres from your
> > application? libpq,
On Mon, Jul 04, 2005 at 10:57:29AM +0200, Enrico Weigelt wrote:
> * Alvaro Herrera <[EMAIL PROTECTED]> wrote:
> > On Mon, Jul 04, 2005 at 02:17:47AM +0200, Enrico Weigelt wrote:
> > > * David Mitchell <[EMAIL PROTECTED]> wrote:
> > > > Perhaps if you are doing a lot of inserts and deletes, vacuumin
* Alvaro Herrera <[EMAIL PROTECTED]> wrote:
> On Mon, Jul 04, 2005 at 02:17:47AM +0200, Enrico Weigelt wrote:
> > * David Mitchell <[EMAIL PROTECTED]> wrote:
> > > Perhaps if you are doing a lot of inserts and deletes, vacuuming every 6
> > > minutes would be closer to your mark. Try vacuuming eve
* David Mitchell <[EMAIL PROTECTED]> wrote:
Hi,
> Hmm, you said you don't experience this when executing the query
> manually. What adapter are you using to access postgres from your
> application? libpq, npgsql or something else?
huh, its a delphi application ... (I didnt code it).
> And wh
Hmm, you said you don't experience this when executing the query
manually. What adapter are you using to access postgres from your
application? libpq, npgsql or something else? And what is your method
for running the query 'manually'. Are you running it locally or from a
remote machine or what?
On Mon, Jul 04, 2005 at 02:17:47AM +0200, Enrico Weigelt wrote:
> * David Mitchell <[EMAIL PROTECTED]> wrote:
> > Perhaps if you are doing a lot of inserts and deletes, vacuuming every 6
> > minutes would be closer to your mark. Try vacuuming every 15 minutes for
> > a start and see how that affe
* David Mitchell <[EMAIL PROTECTED]> wrote:
> Perhaps if you are doing a lot of inserts and deletes, vacuuming every 6
> minutes would be closer to your mark. Try vacuuming every 15 minutes for
> a start and see how that affects things (you will have to do a vacuum
> full to get the tables back
Perhaps if you are doing a lot of inserts and deletes, vacuuming every 6
minutes would be closer to your mark. Try vacuuming every 15 minutes for
a start and see how that affects things (you will have to do a vacuum
full to get the tables back into shape after them slowing down as they
have).
* David Mitchell <[EMAIL PROTECTED]> wrote:
> Did you vacuum full?
>
> When you do lots of inserts and deletes, dead tuples get left behind.
> When you vacuum, postgres will reuse those dead tuples, but if you don't
> vacuum for a long time these tuples will build up lots. Even when you
> vacuu
Did you vacuum full?
When you do lots of inserts and deletes, dead tuples get left behind.
When you vacuum, postgres will reuse those dead tuples, but if you don't
vacuum for a long time these tuples will build up lots. Even when you
vacuum in this case, the dead tuples are still there, althou
* Enrico Weigelt <[EMAIL PROTECTED]> wrote:
forgot to mention:
+ linux-2.6.9
+ postgres-7.4.6
+ intel celeron 2ghz
+ intel ultra ata controller
+ 768mb ram
cu
--
-
Enrico Weigelt== metux IT service
On Mon, Jul 04, 2005 at 12:45:37AM +0200, Enrico Weigelt wrote:
> my application reads and writes some table quite often
> (multiple times per second). these tables are quite small
> (not more than 20 tuples), but the operations take quite a
> long time (>300 ms!).
Are you VACUUMing often enough?
* Steinar H. Gunderson <[EMAIL PROTECTED]> wrote:
> On Mon, Jul 04, 2005 at 12:45:37AM +0200, Enrico Weigelt wrote:
> > my application reads and writes some table quite often
> > (multiple times per second). these tables are quite small
> > (not more than 20 tuples), but the operations take quite a
Hi folks,
my application reads and writes some table quite often
(multiple times per second). these tables are quite small
(not more than 20 tuples), but the operations take quite a
long time (>300 ms!).
The query operations are just include text matching (=) and
date comparison (<,>).
I w
15 matches
Mail list logo
