Skip site navigation (1) Skip section navigation (2)

Re: Profiling

From: Cory 'G' Watson <gphat(at)cafes(dot)net>
To: Shridhar Daithankar <shridhar_daithankar(at)persistent(dot)co(dot)in>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: Profiling
Date: 2002-12-17 14:20:21
Message-ID: AD00D410-11CA-11D7-8929-0003939CCA58@cafes.net (view raw or flat)
Thread:
Lists: pgsql-performance
On Tuesday, December 17, 2002, at 07:49 AM, Shridhar Daithankar wrote:

> On Tuesday 17 December 2002 07:13 pm, you wrote:
>> On Tuesday, December 17, 2002, at 07:35 AM, Shridhar Daithankar wrote:
>>> What is the vacuum frequency?
>>
>> Every morning.  This db is almost exclusively INSERT and SELECT.  
>> Well,
>> I take that back, a single table gets UPDATEs rather frequently.
>> Otherwise, INSERT only.
>
> i recommend a vacuum analyze per 1000/2000 records for the table that 
> gets
> updated. It should boost the performance like anything..

By my math, I'll need to vacuum once every hour or so.  Cron, here I 
come.

vacuumdb --table cached_metrics loggerithim

I assume I do not need a --analyze, since that table has no indexes.  
Should I vacuum the entire DB?

Any other settings I should look at?  Note that I'm not necessarily 
having any problems at present, but one can always tune.  This DB is 
used with a web app (mod_perl/DBI) at the moment, but is moving to a 
Java Swing client, which will give me much more data about performance.

Cory 'G' Watson


In response to

Responses

pgsql-performance by date

Next:From: Shridhar DaithankarDate: 2002-12-17 14:29:42
Subject: Re: Profiling
Previous:From: Shridhar DaithankarDate: 2002-12-17 13:49:33
Subject: Re: Profiling

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group