On Sun, Apr 09, 2000 at 09:47:57PM -0500, Matthew Ross Davis wrote:
> Hey folks,
> 
> My Sparc2 running netbsd 1.4 had been puttering along smoothly for a good 500
> days or so, and then all of a sudden with the past few weeks it's panicked and
> rebooted with the following in /var/log/messages:
> 
> Apr  9 08:17:01 silence /netbsd: WARNING: mclpool limit reached; increase NMBCLU
> STERS
> Apr  9 08:18:06 silence /netbsd: WARNING: mclpool limit reached; increase NMBCLU
> STERS
> 
> Reading through mailing list archives, i found a specific reference to this on
> the alpha-port list, with the instruction to use sysctl to chaneg the value of
> kern.mbuf.nmbclusters - however, it doesn't appear as though this is an actual
> variable, at least not on sparc.
> 
> Any ideas as to A) why this is happening, and B) how i can fix it?

a): well, there's something that eats a lot of mbuf, this can be NFS traffic
    for example; in general anything that can produce a lot of small network
    packets.

b): increase NMBCLUSTERS :). Either by recompiling a kernel with
    'options NMBCLUSTERS=1024' (or more, this is just an example); or by
    patching the binary:
    gdb --write /netbsd
    set nmbclusters=1024
    quit

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--