Re: Partitioning - when is it too many tables?

From: "Tony Wasson" <ajwasson(at)gmail(dot)com>
To: Wes <wespvp(at)syntegra(dot)com>
Cc: "pgsql general" <pgsql-general(at)postgresql(dot)org>
Subject: Re: Partitioning - when is it too many tables?
Date: 2006-03-24 00:05:49
Message-ID: 6d8daee30603231605mba5a061h1276d4ba292f4ebf@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On 3/23/06, Wes <wespvp(at)syntegra(dot)com> wrote:
> This keeps the indexes a more or less reasonable size, and allows quick
> deleting of the old data. Is there any problem with 180 child tables? How
> many would be too many (e.g. if I did one table per 6 hours?)

I am not a guru. Many, many people on the list have more experience
with Table Partitioning (CE), however I will share my experience. I
did some work with CE for our syslog data collection with some nice
results.

* I found that making functions to do the setup work for CE is a must.
* I also found that using 1 rule per table to do query rewriting can
become a bottleneck. This was fine for a small case (20-30 rules), but
it didn't scale nicely when we had 300 rules.

I made 1 table per day and got a nice boost in query speed.

So in short, I think this will work nicely for you. COPY does not
invoke rules, so you should be fine.

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Robert Treat 2006-03-24 00:51:14 Re: [GENERAL] COPY command documentation
Previous Message Just Someone 2006-03-23 23:42:23 Re: Some pgbench results