On Thu, 11 Jul 1996 19:26:21 -0400 
 jds@smerdon.livonia.mi.us (John D. Smerdon) wrote:

 > Today, after being up for 51 days, it dropped into the debugger with (I
 > think) "mb_map full".  I think this was the second time this happened.

You can alleviate this (a little) by increasing the value of NMBCLUSTERS 
to some number like 1024 (or 2048 or 4096, depending on how busy you 
expect your server to be).  The default is 256 (512 with options 
GATEWAY).  This value is quite often too small for a busy network 
server.

This is fixed in NetBSD 1.2 ... when the mb_map fills up, an attempt is 
made to reclaim map space.

 > If I comment out "options DDB", will it automatically reboot instead of
 > dropping into the debugger?  Are there any problems with doing this?

Yes, but if you remove "DDB", make sure to _completely_ rebuild your 
kernel.  Lots of files have "#ifdef DDB" in them.

 > Is it worth *your* time for me to send-pr a stack trace when panics happen?
 > Or is there something else that I should do?

The trick is, that a stack trace for this is often different from one 
panic to another.  The reason why the panic happened was pretty well 
understood.

 -- save the ancient forests - http://www.bayarea.net/~thorpej/forest/ -- 
Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                               Home: 408.866.1912
NAS: M/S 258-6                                          Work: 415.604.0935
Moffett Field, CA 94035                                Pager: 415.428.6939