Re: autovacuum on a -mostly- r/o table - Mailing list pgsql-performance

From Edoardo Ceccarelli
Subject Re: autovacuum on a -mostly- r/o table
Date
Msg-id 451AAB93.6010602@axa.it
Whole thread Raw
In response to autovacuum on a -mostly- r/o table  (Edoardo Ceccarelli <eddy@axa.it>)
Responses Re: autovacuum on a -mostly- r/o table  (Tobias Brox <tobias@nordicbet.com>)
List pgsql-performance
Rod Taylor wrote:
On Wed, 2006-09-27 at 18:08 +0200, Edoardo Ceccarelli wrote: 
I have read that autovacuum cannot check to see pg load before
launching 
vacuum but is there any patch about it? that would sort out the
problem 
in a good and simple way.    
In some cases the solution to high load is to vacuum the tables being
hit the heaviest -- meaning that simply checking machine load isn't
enough to make that decision.

In fact, that high load problem is exactly why autovacuum was created in
the first place. 
True,
but autovacuum could check load -before- and -during- it's execution and it could adjust himself automatically to perform more or less aggressively depending on the difference between those two values.
Maybe with a parameter like: maximum-autovacuum-load=0.2
that would mean: "never load the machine more than 20% for the autovacuum"

...another thing is, how could autovacuum check for machine load, this is something I cannot imagine right now...

pgsql-performance by date:

Previous
From: Edoardo Ceccarelli
Date:
Subject: Re: autovacuum on a -mostly- r/o table
Next
From: Tobias Brox
Date:
Subject: Re: autovacuum on a -mostly- r/o table