Re: integrate pg_upgrade analyze_new_cluster.sh into vacuumdb

From: Kevin Grittner <kgrittn(at)ymail(dot)com>
To: Jeff Janes <jeff(dot)janes(at)gmail(dot)com>, Oskari Saarenmaa <os(at)ohmu(dot)fi>
Cc: Peter Eisentraut <peter_e(at)gmx(dot)net>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: integrate pg_upgrade analyze_new_cluster.sh into vacuumdb
Date: 2014-03-03 19:13:47
Message-ID: 1393874027.72504.YahooMailNeo@web122304.mail.ne1.yahoo.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Jeff Janes <jeff(dot)janes(at)gmail(dot)com> wrote:

> But I do wonder what experience people have with the 3 stage
> process, how useful is it empirically?  If you can't open the
> database for general use until the 3rd phase is done, then you
> would just jump to doing that stage, rather than working through
> all 3 of them.  If you can open the database and muddle through
> without statistics for a while, why not muddle through for the
> little bit longer that it would take to collect the full set
> right off the bat, rather than making intermediate passes?

It's not always a "little bit" of time.  For a description of my
experience with a home-grown 3 stage process before one was built
into pg_upgrade, see this post:

http://www.postgresql.org/message-id/1373465348.51692.YahooMailNeo@web162906.mail.bf1.yahoo.com

Basically, we cut our down time from hours to minutes without
serious impairment of performance.

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Simon Riggs 2014-03-03 19:19:45 Re: ALTER TABLE lock strength reduction patch is unsafe
Previous Message Tom Lane 2014-03-03 19:05:10 Re: UNION ALL on partitioned tables won't use indices.