Re: VACUUMs take twice as long across all nodes

From: Vivek Khera <vivek(at)khera(dot)org>
To: "Pgsql-Performance ((((E-mail))))" <pgsql-performance(at)postgresql(dot)org>
Subject: Re: VACUUMs take twice as long across all nodes
Date: 2006-11-02 21:41:35
Message-ID: 1B174499-A20E-4452-9BB1-355B73D374D5@khera.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance


On Oct 27, 2006, at 2:07 PM, Tom Lane wrote:

> 8.2, but in existing releases I can't see much you can do about it
> except REINDEX when things get slow.

This will be so nice for me. I have one huge table with a massive
amount of churn and bulk deletes. I have to reindex it once every
other month. It takes about 60 to 75 minutes per index (times two
indexes) else I'd do it monthly.

It shaves nearly 1/3 of the relpages off of the index size.

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Benjamin Minshall 2006-11-02 23:17:47 Re: Locking vs. Exceptions
Previous Message Alvaro Herrera 2006-11-02 19:15:55 Re: Query plan for "heavy" SELECT with "lite" sub-SELECTs