From: | Phoenix Kiula <phoenix(dot)kiula(at)gmail(dot)com> |
---|---|
To: | pgsql-performance(at)postgresql(dot)org |
Subject: | PG performance in high volume environment (many INSERTs and lots of aggregation reporting) |
Date: | 2009-01-28 13:33:50 |
Message-ID: | e373d31e0901280533n670c3337x629849181baafd7@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-performance |
[Ppsted similar note to PG General but I suppose it's more appropriate
in this list. Apologies for cross-posting.]
Hi. Further to my bafflement with the "count(*)" queries as described
in this thread:
http://archives.postgresql.org/pgsql-general/2009-01/msg00804.php
It seems that whenever this question has come up, Postgresql comes up
very short in terms of "count(*)" functions.
The performance is always slow, because of the planner's need to guess
and such. I don't fully understand how the statistics work (and the
explanation on the PG website is way too geeky) but he columns I work
with already have a stat level of 100. Not helping at all.
We are now considering a web based logging functionality for users of
our website. This means the table could be heavily INSERTed into. We
get about 10 million hits a day, and I'm guessing that we will have to
keep this data around for a while.
My question: with that kind of volume and the underlying aggregation
functions (by product id, dates, possibly IP addresses or at least
countries of origin..) will PG ever be a good choice? Or should I be
looking at some other kind of tools? I wonder if OLAP tools would be
overkill for something that needs to look like a barebones version of
google analytics limited to our site..
Appreciate any thoughts. If possible I would prefer to tone down any
requests for MySQL and such!
Thanks!
From | Date | Subject | |
---|---|---|---|
Next Message | Robert Haas | 2009-01-28 14:08:50 | Re: LIKE Query performance |
Previous Message | A. Kretschmer | 2009-01-28 09:43:19 | Re: LIKE Query performance |