Re: settings input for upgrade - Mailing list pgsql-performance

From Scott Marlowe
Subject Re: settings input for upgrade
Date
Msg-id CAOR=d=3BQv_MM58kn8aYdrVD8TBoQR=4=gqMVFPH4xUocn5Uzw@mail.gmail.com
Whole thread Raw
In response to Re: settings input for upgrade  (Greg Smith <greg@2ndQuadrant.com>)
List pgsql-performance
On Sun, Aug 21, 2011 at 1:20 PM, Greg Smith <greg@2ndquadrant.com> wrote:
>  deadlock_timeout = 3s
>
> You probably don't want to increase this.  When you reach the point where
> you want to find slow lock issues by turning on log_lock_waits, you're just
> going to put it right back to the default again--or lower it.

All of these random changes brings up the REAL subject, that changes
should be made by measuring performance before and after each change
set and justifying each change.  Just randomly throwing what seem like
good changes at the database is a surefire recipe for disaster.

pgsql-performance by date:

Previous
From: Greg Smith
Date:
Subject: Re: settings input for upgrade
Next
From: Krzysztof Chodak
Date:
Subject: Re: Variable versus constrant size tuples