Proposal: Store "timestamptz" of database creation on "pg_database" - Mailing list pgsql-hackers

From Fabrízio de Royes Mello
Subject Proposal: Store "timestamptz" of database creation on "pg_database"
Date
Msg-id CAFcNs+qMGbLmeUOnjmbna_K7=UP817BPw9QxhbCTGNScPKVoeA@mail.gmail.com
Whole thread Raw
Responses Re: Proposal: Store "timestamptz" of database creation on "pg_database"  (Josh Berkus <josh@agliodbs.com>)
List pgsql-hackers
Hi all,

This proposal is about add a column "datcreated" on "pg_database" to store the "timestamp" of the database creation.

A couple weeks ago I had a trouble with a PostgreSQL instance, actually our ERP had some strange behaviors with some data loss, but I searched for ERRORs in log files (OS and PG) and I found nothing.

Looking at the files and directories in the cluster noticed something strange, the date / time of the file "base/9999/PG_VERSION" (database of our ERP) was different compared to when we create it. So I used the following SQL to check the date / time of creation of the databases in the cluster:

fabrizio=# SELECT datname, (pg_stat_file('base/'||oid||'/PG_VERSION')).modification AS datcreated
fabrizio-#   FROM pg_database;
  datname  |       datcreated       
-----------+------------------------
 template1 | 2012-12-26 12:11:53-02
 template0 | 2012-12-26 12:11:54-02
 postgres  | 2012-12-26 12:11:54-02
 fabrizio  | 2012-12-26 12:12:02-02
(4 rows)

This isn't an elegant solution to do that, but worked fine. However, why not we have a column to store this information?

Somebody have another idea?

Regards,

--
Fabrízio de Royes Mello
Consultoria/Coaching PostgreSQL
>> Blog sobre TI: http://fabriziomello.blogspot.com
>> Perfil Linkedin: http://br.linkedin.com/in/fabriziomello
>> Twitter: http://twitter.com/fabriziomello

pgsql-hackers by date:

Previous
From: Greg Stark
Date:
Subject: Re: buffer assertion tripping under repeat pgbench load
Next
From: Josh Berkus
Date:
Subject: Re: Proposal: Store "timestamptz" of database creation on "pg_database"