* C. Stephen Gunn <[EMAIL PROTECTED]> [010103 06:22] wrote:
> On Tue, Jan 02, 2001 at 07:18:10AM -0800, Alfred Perlstein wrote:
>
> > I notice that if_fxp.c seems to fail silently when unable to
> > allocate network buffers, perhaps the fact that the kernel can
> > be doing about 200 or more messages per-second makes me guess
> > that it's not helping things, especially when the box has a
> > serial console.
> >
> > I'm taking the printf's out of my copy of if_xl.c on a 3.5
> > system, I'll let you know if it still goes nuts for extended
> > amounts of time.
>
> You could change the printf()'s to log()'s, and allow syslogd
> to filter/combine/report them on your console. I'm sure you
> already know this, you could increase NBMCLUSTERS and friends
> in your kernel config, or at boot time. (For posterity in
> the list archives).
I really don't see any point in assisting a cascade-like failure
situation by adding more work that the kernel has to do when faced
with a shortage.
Raising nmbclusters will help, but only to delay the problem.
--
-Alfred Perlstein - [[EMAIL PROTECTED]|[EMAIL PROTECTED]]
"I have the heart of a child; I keep it in a jar on my desk."
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-net" in the body of the message