Re: Toast space grows - Mailing list pgsql-performance

From Pavel Rotek
Subject Re: Toast space grows
Date
Msg-id 3556c08c0803070627w5a842140n9a0c9b574909849e@mail.gmail.com
Whole thread Raw
In response to Re: Toast space grows  (Matthew <matthew@flymine.org>)
List pgsql-performance


2008/3/7, Matthew <matthew@flymine.org>:
On Fri, 7 Mar 2008, Pavel Rotek wrote:
> well, this will be the main problem... But... do uncomitted trasactions
> affect toast space?


I think the demonstrated answer to this is yes.


> (begin tx, load previous log, do business action, append new log, flush, do
> business action, append new log, flush, ... do business action, append new
> log, flush, commit tx).


If all you're doing is appending to the end of the log, why don't you make
each "append" a new row in a table. Instead of building massive rows, use
the database for what it was designed for, and have many smaller
independent rows.

Matthew

Because I modify existing application, where logic is already given :-(. If no other way exists, i'll have to do refactoring...

--
"To err is human; to really louse things up requires root
  privileges."                 -- Alexander Pope, slightly paraphrased

pgsql-performance by date:

Previous
From: "Pavel Rotek"
Date:
Subject: Re: Toast space grows
Next
From: Tom Lane
Date:
Subject: Re: Toast space grows