Re: possible vacuum improvement? - Mailing list pgsql-hackers

From Mario Weilguni
Subject Re: possible vacuum improvement?
Date
Msg-id 4D618F6493CE064A844A5D496733D667039014@freedom.icomedias.com
Whole thread Raw
In response to possible vacuum improvement?  (Mario Weilguni <mweilguni@sime.com>)
Responses Re: possible vacuum improvement?  (Tom Lane <tgl@sss.pgh.pa.us>)
Re: possible vacuum improvement?  ("Christopher Kings-Lynne" <chriskl@familyhealth.com.au>)
List pgsql-hackers
>I do not think we need or want a control table for this; certainly I see
>no need for per-table manual control over this process.  There should
>probably be a few knobs in the form of GUC parameters so that the admin
>can control how much overall work the auto-vacuumer does.  For instance
>you'd probably like to turn it off when under peak interactive load.

If (auto)vacuum is clever to check that some tables do not need vacuum
there's really no need for that. That brings me to another point, can't the
statistics collector used for that?

For my database I wrote a statistic display program for web-access, and all
the info autovacuum would need is here.
http://mw.sime.com/pgsql.htm

That brings me to another point, is there interest for this
web-statistics-frontend, maybe for /contrib? I found it extremly useful
because it showed up the weak points in my applications.

Best regards,Mario Weilguni


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: possible vacuum improvement?
Next
From: Michael Meskes
Date:
Subject: Re: BYTEA with ecpg