[ On Thursday, March 21, 2002 at 23:17:45 (+0100), Wojciech Puchar wrote: ]
> Subject: no buffer space - again :(
>
> it's not too low NMBCLUSTERS!!!!
> it MUST be error somewhat.
> it's 8MB machine with NMBCLUSTERS as high as 2048!!!!
> after "No buffer..." the only fix is rebooting.

.... or adding more NMBCLUSTERS....  :-)

It's really unfortunate that NetBSD doesn't keep (or is it "make easily
available"?) historical stats about network buffer usage like FreeBSD
now does:

$ netstat -m
1715/5824/131072 mbufs in use (current/peak/max):
        1694 mbufs allocated to data
        21 mbufs allocated to packet headers
1543/5482/32768 mbuf clusters in use (current/peak/max)
12420 Kbytes allocated to network (12% of mb_map in use)
0 requests for memory denied
0 requests for memory delayed
0 calls to protocol drain routines


I've wanted to dig the code that does this out of FreeBSD and port it to
NetBSD for some time now, but I haven't gotten around to it yet (read:  I
haven't had a production machine suffer like yours is where I need to
very carefully tune NMBCLUSTERS without wasting memory unnecessarily in
the way I'm able to do on that FreeBSD machine I show above).

-- 
								Greg A. Woods

+1 416 218-0098;  <gwoods@acm.org>;  <g.a.woods@ieee.org>;  <woods@robohack.ca>
Planix, Inc. <woods@planix.com>; VE3TCP; Secrets of the Weird <woods@weird.com>