On Sat, 21 Oct 1995 16:35:56 -0400 
 "Michael K. Sanders" <msanders@confusion.net> wrote:

[ Sorry for the additional cross-post, but it seems to me that follow-ups
  to this should take place on tech-kern. --thorpej ]

 > 16MB RAM, 32MB swap... there wasn't anything particularly intensive going on 
 > when I got the following panic:
 > 
 > mb_map full
 > panic: kmem_malloc: mb_map too small

This can happen if your system is dealing with alot of network traffic.  
Add an option like:

options		NMBCLUSTERS=1024

to you kernel config file, and it will probably go away.  I definitely 
had to do this on my hp380 which is the diskless boot server for my hp319 
and MVME-147, plus my source tree server, etc.  In fact, I had to bump up 
to 2048; the larger number provided by `options GATEWAY' (512 on the 
hp300) just wasn't sufficient...

As computer networks get larger, and applications grow increasingly 
network-oriented, I would tend to favor the default NMBCLUSTERS being 
raised from 256/512 to 1024/2048 (without/with GATEWAY).  Just like we 
recently raised CHILD_MAX frm 40 to 80, the lower value just doesn't seem 
to make sense for modern systems anymore...

Comments?

--------------------------------------------------------------------------
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