DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma;
	h=message-id:date:from:to:subject:cc:in-reply-to:mime-version
	:content-type:content-transfer-encoding:content-disposition
	:references;
	b=U6qLfcBJCcAjLzqMOXy9Rbvh/lWa8B9IlNxSWMraU9WOjXZ+2L9fTju+K4+mpB3W+z
	hdTP6yGzX6Pu7/zCSgH5N7Rn0BV899Q9knfxV9CdmA6DOJWDyl0e5bbfsF+W0zuFQKGp
	vozmBBU+yukBBU/1kShPDtbve7kied/aV+rdI=
Message-ID: <dcc563d10901131450m72c01387y11db04359cbdcfdf@mail.gmail.com>
Date: Tue, 13 Jan 2009 15:50:39 -0700
From: "Scott Marlowe" <scott.marlowe@gmail.com>
To: "Scott Whitney" <swhitney@journyx.com>
Subject: Re: Strange deadlock error last night
Cc: pgsql-admin@postgresql.org
In-Reply-To: <20090113223407.BDF247E46BA@mail.int.journyx.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
References: <dcc563d10901131416w73527ec4i29bbd2920631dbc1@mail.gmail.com>
	<20090113223407.BDF247E46BA@mail.int.journyx.com>

On Tue, Jan 13, 2009 at 3:35 PM, Scott Whitney <swhitney@journyx.com> wrote:
> Thanks for all the information, guys. I think Tom was right. Our application
> was doing a couple of full vacs at the same time. It's weird that we didn't
> run into this in the past.
>
> You're all absolutely right about the upgrading, but in our environment,
> it's not 2-3 minutes. It's 2-3 weeks. I've got to fully vet the app on the
> platform internally with full test plans, etc, even for the most minor
> upgrades; corp policy.

That's preparation time versus execution time.  While it may take 2-3
weeks to get the testing done, once done, the actual down time is
minimal.  I would suggest that your current testing methodology is
horribly flawed.  I know it was foisted on you by outside forces, so
it's not really "your" idea.  But you know what I mean.  There's the
cost of possible data loss / down time due to unforseen problems
caused by an update, and there's the cost of possible data loss / down
time due to known flaws in your current code.  If your testing
methodology was strict enough, it would have found such flaws that
were fixed in subsequent releases, but it hasn't.  Meantime, you spend
weeks testing for possible new bugs, while running known broken code
with real data loss / down time inducing bugs.

There's a reasonable cost / benefits analysis, and right now,
someone's set the cost far too high for far too few benefits.  Been
where you are, and I know how much it sucks.