Re: Wrong docs on wal_buffers?

From: Greg Smith <greg(at)2ndquadrant(dot)com>
To: Samuel Gendler <sgendler(at)ideasculptor(dot)com>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Josh Berkus <josh(at)agliodbs(dot)com>, Jeff Janes <jeff(dot)janes(at)gmail(dot)com>, pgsql-performance(at)postgresql(dot)org
Subject: Re: Wrong docs on wal_buffers?
Date: 2011-01-07 18:46:12
Message-ID: 4D275F74.3040405@2ndquadrant.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

Samuel Gendler wrote:
> I was trying to make the point that, in an attempt to make things very
> easy for novice users, we are actually making them quite a bit more
> complex for novice users who want to do anything besides start the server.

People who can't start the server often abandon PostgreSQL, never come
back again. And in the worst case, they popularize their frustration
via blogs etc. That contributes to the perception the program is hard
to use far more than people who run into problems only with
performance--people expect database tuning to be non-trivial, but they
have minimal tolerance for trouble when first trying to use a program.
From an advocacy perspective, there is no higher priority than making
sure things work as smoothly as feasible for people who have never run
PostgreSQL before. Changing the software so it doesn't work out of the
box on a system with minimal shared memory defaults, as seen on Linux
and other operating systems, would be one of the worst possible changes
to the database anyone could make.

About once a month now I came across someone who used my pgtune
program: https://github.com/gregs1104/pgtune to get a reasonable
starting configuration. That program is still rough and has
limitations, but the general idea turns out to work just fine. I'm
bored with discussions about making this easier for users unless they
involve people volunteering to help with the coding needed to turn that,
or something better than it, into a release quality tool. The path to
sort this out is mapped out in excruciating detail from my perspective.
The only thing missing are code contributors with time to walk down it.
So far it's me, occasional code refinement from Matt Harrison, some
packager help, and periodic review from people like Josh Berkus. And
that's just not good enough to make progress on this particular front
quickly enough.

--
Greg Smith 2ndQuadrant US greg(at)2ndQuadrant(dot)com Baltimore, MD
PostgreSQL Training, Services and Support www.2ndQuadrant.us
"PostgreSQL 9.0 High Performance": http://www.2ndQuadrant.com/books

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Greg Smith 2011-01-08 01:59:01 pgbench to the MAXINT
Previous Message Samuel Gendler 2011-01-07 17:25:15 Re: Wrong docs on wal_buffers?