Skip site navigation (1) Skip section navigation (2)

Re: [GENERAL] pg_upgrade problem

From: Alvaro Herrera <alvherre(at)commandprompt(dot)com>
To: depesz <depesz(at)depesz(dot)com>
Cc: Bruce Momjian <bruce(at)momjian(dot)us>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: [GENERAL] pg_upgrade problem
Date: 2011-08-29 18:41:38
Message-ID: 1314643001-sup-6186@alvh.no-ip.org (view raw or flat)
Thread:
Lists: pgsql-generalpgsql-hackers
Excerpts from hubert depesz lubaczewski's message of lun ago 29 14:49:24 -0300 2011:
> On Mon, Aug 29, 2011 at 06:54:41PM +0200, hubert depesz lubaczewski wrote:
> > On Fri, Aug 26, 2011 at 05:28:35PM +0200, hubert depesz lubaczewski wrote:
> > > On Fri, Aug 26, 2011 at 12:18:55AM -0400, Bruce Momjian wrote:
> > > > 
> > > > OK, this was very helpful.  I found out that there is a bug in current
> > > > 9.0.X, 9.1.X, and HEAD that I introduced recently when I excluded temp
> > > > tables.  (The bug is not in any released version of pg_upgrade.)  The
> > > > attached, applied patches should fix it for you.  I assume you are
> > > > running 9.0.X, and not 9.0.4.
> > > 
> > > pg_upgrade worked. Now I'm doing reindex and later on vacuumdb -az.
> > 
> > vacuumdb failed. The fail looks very similar to the one I had on 9.0.4.
> > 
> > After long vacuum I got:
> > INFO:  vacuuming "pg_toast.pg_toast_106668498"
> > vacuumdb: vacuuming of database "etsy_v2" failed: ERROR:  could not access status of transaction 3429738606
> > DETAIL:  Could not open file "pg_clog/0CC6": No such file or directory.

I don't understand the pg_upgrade code here.  It is setting the
datfrozenxid and relfrozenxid values to the latest checkpoint's NextXID,

        /* set pg_class.relfrozenxid */
        PQclear(executeQueryOrDie(conn,
                                  "UPDATE   pg_catalog.pg_class "
                                  "SET  relfrozenxid = '%u' "
        /* only heap and TOAST are vacuumed */
                                  "WHERE    relkind IN ('r', 't')",
                                  old_cluster.controldata.chkpnt_nxtxid));

but I don't see why this is safe.  I mean, surely the previous
vacuum might have been a lot earlier than that.  Are these values reset
to more correct values (i.e. older ones) later somehow?  My question is,
why isn't the new cluster completely screwed?

I wonder if pg_upgrade shouldn't be doing the conservative thing here,
which AFAICT would be to set all frozenxid values as furthest in the
past as possible (without causing a shutdown-due-to-wraparound, and
maybe without causing autovacuum to enter emergency mode either).

-- 
Álvaro Herrera <alvherre(at)commandprompt(dot)com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support

In response to

Responses

pgsql-hackers by date

Next:From: Tom LaneDate: 2011-08-29 18:41:49
Subject: Re: spinlocks on HP-UX
Previous:From: Robert HaasDate: 2011-08-29 18:32:49
Subject: Re: spinlocks on HP-UX

pgsql-general by date

Next:From: Martín MarquésDate: 2011-08-29 18:52:12
Subject: Re: passing cursors from one PL function to another
Previous:From: Pavel StehuleDate: 2011-08-29 18:28:11
Subject: Re: passing cursors from one PL function to another

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group