Re: Performance problems with large telemetric datasets on 7.4.2

From: Heikki Linnakangas <heikki(at)enterprisedb(dot)com>
To: Sven Clement <sven(at)dsign(dot)lu>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Andrew Kroeger <andrew(at)sprocks(dot)gotdns(dot)com>, Mark Lewis <mark(dot)lewis(at)mir3(dot)com>, pgsql-performance(at)postgresql(dot)org
Subject: Re: Performance problems with large telemetric datasets on 7.4.2
Date: 2007-08-06 09:51:09
Message-ID: 46B6EF0D.6030501@enterprisedb.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

Sven Clement wrote:
> Partially I found that one in the PostgreSQL Documentation for the
> 7.x.xversions under the command REINDEX where they claim that you
> should run a
> reindex under certain circumstances and for my comprehension this says that
> with some access pattern (as ours (major writes / one big delete per day))
> the index may be corrupted or otherwise not really useful.

Up to 7.3, periodical REINDEX was needed to trim down bloated indexes.
Since 7.4, empty index pages are recycled so that's no longer necessary.
You can still end up with larger than necessary indexes in recent
versions under unusual access patterns, like if you delete all but a few
index tuples from each index page, but it's rare in practice. And it's
not unbounded growth like in <= 7.3.

In any case, the indexes won't become corrupt.

--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com

In response to

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Sven Clement 2007-08-06 09:54:52 Re: Performance problems with large telemetric datasets on 7.4.2
Previous Message Ragnar 2007-08-06 09:50:25 Re: Performance problems with large telemetric datasets on 7.4.2