In message <20031005233328.GC16113@audible.transient.net>Jamie Heilman writes
>Ian Spray wrote:
>> The network I/O is poor, and I've spent ages messing with the different PIO
>> and DMA modes in custom kernels and have only managed to slow the system
>> down compared to the conf/DEFAULT settings.  I haven't messed with the
>> nmbclusters option (mentioned in the archive) but I am wondering if my
>> setting the kern.maxvnodes to 40000 has anything to do with the crashes.
>> The vnodes are exhausted within a few hours, but in doing so have I taken
>> the value too high ?  No idea.
>
>Well, I've seen the crashes on my qube2 with both vnodes and
>nmbclusters tweaked and untweaked (I only started tweaking them
>recently), so frankly I doubt it has anything to do with that.
>I suspect the nic driver has issues.

Wow, I know this is a way to totally jinx myself, but my Qube is
doing great now and has been for quite some time.  I have additional
memory (160MB total) plus a bigger hard disk (60GB total).  I don't
use it really hard, but I've certainly transfered multiple gigabytes 
of data with no problems.

OTOH, I'm running -current.  Some time ago, I had problems with
crashes when doing ttcp, and all sorts of other issues, but they
seem to be solved now.  I can get about 2500KB/sec with ttcp,
which is, upon reflection, pretty slow, but at least it works.
You might wanna try updating your system, although I realize this
is a pain.

Cheers,

Kevin
kml@patheticgeek.net