Date: Thu, 10 Mar 2011 09:09:00 -0600
From: Kenneth Marshall <ktm@rice.edu>
To: Scott Whitney <swhitney@journyx.com>
Cc: pgsql-admin@postgresql.org
Subject: Re: pg_clogs hanging around
Message-ID: <20110310150900.GA8169@aart.is.rice.edu>
References: <20110310140405.GZ8169@aart.is.rice.edu>
	<5606497.14104.1299766714935.JavaMail.root@zimbra.int.journyx.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: 
 <5606497.14104.1299766714935.JavaMail.root@zimbra.int.journyx.com>
User-Agent: Mutt/1.5.16 (2007-06-09)

On Thu, Mar 10, 2011 at 08:18:34AM -0600, Scott Whitney wrote:
> > On Thu, Mar 10, 2011 at 07:56:26AM -0600, Scott Whitney wrote:
> > > >
> > > > If you have hardware problems like that you have way more
> > > > problems. You could have corruption (silent) occurring in any of
> > > > the other
> > > > database files. Good luck.
> > >
> > > I am, in fact, aware of that, but every single machine ever
> > > manufactured will have hardware problems such at this at some point.
> > > It stems quite simply from Ohm's Law, one gross over-simplification
> > > of which is as simple as "if it's got a resistor in it, it's going
> > > to fail at some point", as I'm sure you know. It's merely a matter
> > > of whether proactive replacement, backups, standby systems, etc
> > > ameliorate that risk. When we had our failure a couple of years ago,
> > > it did not.
> > >
> > > Regardless, my question still stands, and I do, in fact, care about
> > > ANY database blocking cleanup of clogs (or anything else). There's
> > > this concept of "if this then what else," and if template0 (or
> > > anyone else) is blocking that ability to properly clean those up,
> > > what else is possibly screwed up in a similar fashion.
> > >
> > > So, what can I do to resolve this issue?
> > >
> > 
> > True, entropy rules. I think that you can use "VACUUM FREEZE"
> > to allow the clogs to be cleaned up.
> > 
> I thought I was. I've got -F in there. vacuumdb -a -v -F
> 

Sorry about that, you are. I have never done it from the command-line
just from psql. I was thinking F for FULL... :)

Cheers,
Ken