Postgres Pro
Downloads
Home   >   mailing lists

Re: Redundant database objects. - Mailing list pgsql-general

From Andrew Bartley
Subject Re: Redundant database objects.
Date
Msg-id AANLkTim89HDk1sV6LDnCoZ0FnRZqgzn8iI6qlLd3kNDZ@mail.gmail.com
Whole thread Raw
In response to Re: Redundant database objects.  (Greg Smith <greg@2ndquadrant.com>)
Responses Re: Redundant database objects.  (Greg Smith <greg@2ndquadrant.com>)
List pgsql-general
Tree view
Thanks Greg,

It seems that the underlying stats tables are reset on a periodic basis, can i stop this process? Is it a .conf setting?  I have had a good look around, nothing sticks out. If I can stop it, then i could use pg_stat_reset() then monitor the stat views over an extended period without them being reset by some periodic job.  That, at the moment, is my main concern. 

Thanks again.

Also i need to find similar information regarding functions and views....  Any suggestions?

Thanks 

Andrew Bartley


On 13 July 2010 09:45, Greg Smith <greg@2ndquadrant.com> wrote:
Andrew Bartley wrote:
Unfortunately these views only give me what appears to be a certain time frame.  This does not help all that much.  It will give a list of tables, indexes and sequences that have been used in the time frame, so that is at least a start.

You can use pg_stat_reset() to set those back to 0 again and then see what actually gets used moving forward from the point you do that.  That's a reasonable idea to do anyway to make all those statistics better reflect recent activity rather than historical.  Just be warned that it will screw up many monitoring systems if you have them pointed toward those statistics tables and grabbing snapshots, some will view the reset as the values going negative which doesn't make any real-world sense.

--
Greg Smith  2ndQuadrant US  Baltimore, MD
PostgreSQL Training, Services and Support
greg@2ndQuadrant.com   www.2ndQuadrant.us


pgsql-general by date:

Previous
From: Greg Smith
Date:
Subject: Re: Redundant database objects.
Next
From: Scott Marlowe
Date:
Subject: Re: PG_DUMP very slow because of STDOUT ??