Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 4 Jun 1999 05:03:28 +0200
From:      Eivind Eklund <eivind@FreeBSD.ORG>
To:        Jordan Hubbard <jkh@zippy.cdrom.com>
Cc:        David Scheidt <dscheidt@enteract.com>, freebsd-ports@FreeBSD.ORG, freebsd-hackers@FreeBSD.ORG
Subject:   Re: a two-level port system? (fwd)
Message-ID:  <19990604050327.I77195@bitbox.follo.net>
In-Reply-To: <10787.928406989@peewee>; from Jordan Hubbard on Thu, Jun 03, 1999 at 03:49:49AM -0700
References:  <19990602173528.B70808@bitbox.follo.net> <10787.928406989@peewee>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Jun 03, 1999 at 03:49:49AM -0700, Jordan Hubbard wrote:
> > of useless.  It's like doing uphill testing of a fat guy on a bicycle
> > against a Lamborghini - you "know" the result beforehand.
> 
> Unfortunately, what you're probably not aware of is that the fat guy
> also has a JATO unit strapped to the back of his bicycle.  Don't make
> assumptions. :-)

I think the bricks some joker put inside the wheels of the Lamborghini
might be more of an issue, actually.  Not that I will take it as an
assumption... :)

> > If extraction of the ports collection (not files in general, just the
> > ports collection) is slower using soft updates than using "async"
> > mode, then it seems some elevator sorting isn't working the way it
> 
> Extraction of ALL the distribution bits is faster with async than it
> is with soft updates.

The IMO most relevant benchmark is minimal install + ports - because
that's where you see the impact of the ports extraction most.

> To put it another, more practical, way - if you timed the
> installation with a stopwatch, with or without ports, the async
> policy would win and Kirk has even pointed that out in other emails.

That's interesting, but not really relevant for my reason for wanting
this - which is the psychological effect on the user.  I'm not really
interested in how long time the installation takes - I'm interested in
the user NOT seeing his 100MBit link drop down to 10KB/s download
speed during extraction of the ports collection.

> Given that, I have to honestly wonder why you've been arguing so
> strongly for soft updates being used in the installation.

Because I've been assuming that dropping the synchronous writes during
the ports collection extract would have a really significant
difference.  My wish is to fix the ports extraction time, and your
previous reply (on IRC) indicated that the reason for not wanting soft
updates on the install floppy was the license, not that it didn't fix
the speed.

Eivind.


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-ports" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990604050327.I77195>