Re: Postgres 9.1 statistics in pg_stat_database - Mailing list pgsql-admin

From Steve Crawford
Subject Re: Postgres 9.1 statistics in pg_stat_database
Date
Msg-id 5113FF56.5020309@pinpointresearch.com
Whole thread Raw
In response to Re: Postgres 9.1 statistics in pg_stat_database  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: Postgres 9.1 statistics in pg_stat_database  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-admin
On 02/07/2013 11:05 AM, Tom Lane wrote:
> =?utf-8?q?Milos_Gajdos?= <gy2kre@centrum.cz> writes:
>> What this shows is that xact_commit increases by 3 after just one row insert. tup_inserted does seem to report
correctvalue. After simple SELECT, xact_commit increases by 2 - kind of strange as SELECT shouldn't really be
increasingxact_commit ? 
> Yes, a SELECT does count as an xact_commit...
>
Additionally (correct me if I'm wrong, Tom), background activity like
autovacuum will generate countable transactions. I can see transactions
increment by a dozen or more on a completely "idle" database if I just
wait a minute or two but the number of those transactions, while visible
on an idle system, is lost in the noise on a busy database.

Other things that generate transactions include connecting to the
database (one, it appears) but \d in psql appears to generate a minimum
of four.

Cheers,
Steve



pgsql-admin by date:

Previous
From: Tom Lane
Date:
Subject: Re: Postgres 9.1 statistics in pg_stat_database
Next
From: Tom Lane
Date:
Subject: Re: Postgres 9.1 statistics in pg_stat_database