Vacuum template databases, Urgent: Production problem

From: Pallav Kalva <pkalva(at)livedatagroup(dot)com>
To: pgsql-performance(at)postgresql(dot)org
Subject: Vacuum template databases, Urgent: Production problem
Date: 2006-03-14 16:44:12
Message-ID: 4416F2DC.4010400@livedatagroup.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

Hi,

Do we have to vacuum template0 database regularly ? We got this warning this morning while vacuuming databases. As a part of my daily vacuum job I do vacuum of quartz, helix_fdc and affiliate databases which are the
one's which are heavily updated and used. But today I realized that usps, template1 and template0 is also being used in a transaction somehow based on this (SELECT datname, age(datfrozenxid) FROM pg_database;) query.
Actually we dont do any updates on usps , template1 and templat0 databases but some how still the age(datfrozenxid keeps incrementing.

My question now is do I have to vacuum daily template1 and template0 databse, is there any harm on vacuuming these databases daily ?, since these are postgres system tables I am kind of worried.
I was told that template0 is freezed but not sure why the age(datfrozenxid keeps incrementing.
I am going to vacuum usps from now anyway. We are using Postgres version 8.0.2

If some one can please help me on this it would be really great, this is a production database and we cant afford to loose anything.

Thanks!
Pallav.

Message from the log
---------------------
WARNING: some databases have not been vacuumed in 1618393379 transactions
HINT: Better vacuum them within 529090268 transactions, or you may have a wraparound failure.

SELECT datname, age(datfrozenxid) FROM pg_database;
datname | age
-----------+------------
quartz | 1076729648
helix_fdc | 1078452246
usps | 1621381218
affiliate | 1078561327
template1 | 1621381218
template0 | 1621381218
(6 rows)


SELECT datname, age(datfrozenxid) FROM pg_database;
datname | age
-----------+------------
quartz | 1076770467
helix_fdc | 1078493065
usps | 1621422037
affiliate | 1078602146
template1 | 1621422037
template0 | 1621422037
(6 rows)


I ran this just 2 minutes apart and you can see the age value changes for
template0 and template1

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Tom Lane 2006-03-14 17:27:50 Re: Vacuum template databases, Urgent: Production problem
Previous Message Merlin Moncure 2006-03-14 14:06:36 Re: firebird X postgresql 8.1.2 windows, performance comparison