Date: Fri, 30 Nov 2001 20:39:05 -0500 From: Leo Bicknell <bicknell@ufp.org> To: Luigi Rizzo <rizzo@aciri.org> Cc: Mike Silbersack <silby@silby.com>, Alfred Perlstein <bright@mu.org>, freebsd-hackers@FreeBSD.ORG Subject: Re: TCP Performance Graphs Message-ID: <20011130203905.A2944@ussenterprise.ufp.org> In-Reply-To: <20011130173033.G33041@iguana.aciri.org>; from rizzo@aciri.org on Fri, Nov 30, 2001 at 05:30:33PM -0800 References: <20011130171418.B96592@ussenterprise.ufp.org> <Pine.BSF.4.30.0111301717290.10049-100000@niwun.pair.com> <20011130173033.G33041@iguana.aciri.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Nov 30, 2001 at 05:30:33PM -0800, Luigi Rizzo wrote:
> I just committed to current (and soon to stable) some code to log
> _failures_ in mbuf allocations, but that is only meant as an aid
> to remove worse code in the drivers.
Note that if we implement a 'fair share' buffering scheme we would
never get a failure, which would be a good thing. Unfortuantely
fair share is relatively complicated.
> (Plus, just setting a threshold is not good, you also want
> some histeresys, because you can easily conceive a system that
> runs at XX % mbuf occupation, whatever XX you pick.)
With fair share or some other type of setup I would agree. Given
our current 'things fail badly if you run out' I think a warning
at 90% or 200 left, or something would be a real good idea. With
the current allocation scheme this is on par with "/foo file system
is full" messages, we should have a "networking stack is full,
build a kernel with more mbuf's message".
--
Leo Bicknell - bicknell@ufp.org - CCIE 3440
PGP keys at http://www.ufp.org/~bicknell/
Read TMBG List - tmbg-list-request@tmbg.org, www.tmbg.org
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20011130203905.A2944>