Re: [BUGS] BUG #1552 followup

On 8/2/05, Dan Armbrust <daniel.armbrust.list@gmail.com> wrote:
> Jaime Casanova wrote:
> On 8/2/05, Dan Armbrust <daniel.armbrust.list@gmail.com>
> wrote:

> I've been trying to track down some performance problems that I am
having
> doing large inserts on tables with foreign keys.

I'm 99% sure that the
> issue I am having is BUG 1552
> -
http://archives.postgresql.org/pgsql-bugs/2005-03/msg00183.php

When
> I read the whole thread, I got the impression that a fix (or maybe
a hack)
> had been put into the 8.0 branch. I'm running 8.0.3, and I'm
seeing the
> issue. I can go into more details if desired...
So my questions are:
1)
> Would that fix have made it into the 8.0.3 release - alternatively,
how can
> I figure that out myself?
2) Is there a bug tracker somewhere, or do you
> just track bugs on the
mailing list? I haven't been able to find anything on
> your site.

Thanks,

Dan


> What part of the discussion gives you that impression... I understood
that
> Tom will (in the future) see a fix when Neil finish the framework
to replan
> prepared statements (or is rebuild cached plans where the
table size has
> grow significantly?)

PD: This is not the list for this discussion... i am
> forwarding you to
another(randomly choosen) list.


> This posting seems to indicate that there was a change made:
> http://archives.postgresql.org/pgsql-bugs/2005-03/msg00183.php
>

After re-read all thread i found that Tom states this as a real
solution to the problem:
http://archives.postgresql.org/pgsql-bugs/2005-03/msg00208.php

and try this as a band-aid... the fact that he was looking for
_better_ suggestions let me think he was not sure this will fix it:
http://archives.postgresql.org/pgsql-bugs/2005-03/msg00259.php

It's there since 8.0.2 release:
http://www.postgresql.org/docs/8.0/static/release-8-0-2.html
--> Planner adjustments to improve behavior on freshly-created tables (Tom)

>
> What is the proper way to discuss bugs?  The only thing I have found on the
> website is the link for reporting new bugs, (which just sends an e-mail to
> the bugs mailing list?)
>
> Since I didn't see any bug tracker software, I figured (wrongly) that you
> must just discuss bugs on the bugs list...
>
>

you are right bugs discussion occurs in bugs lists. Maybe a misread
your message but because the thread is stating that the real solution
is other (i wasn't aware of the change) i thing other list maybe
performance like Josh suggest to the original bug posting. Now i think
you were right in bugs, because there was an attempt to baind-aid it.

in IMHO, bugs makes the server crash or gives you wrong data... this
seems to me at least (not an expert i have to admit) like a
performance problem that can be found workarounds in pgsql-performance
until a real solution can be implemented.

--
regards,
Jaime Casanova
(DBA: DataBase Aniquilator ;)