I'm seeing this message on the syslog, the machine seems to crash
quite soon after this.

Sep  8 22:27:25 news /netbsd: mb_map full
Sep  8 22:47:47 news /netbsd: NetBSD 1.0A (HAUKI) #7: Fri Sep  1

Sep  8 05:04:52 news /netbsd: mb_map full
Sep  8 05:30:28 news /netbsd: NetBSD 1.0A (HAUKI) #7: Fri Sep  1

Sep  7 22:47:18 walrus /netbsd: mb_map full
Sep  7 23:02:47 walrus /netbsd: NetBSD 1.0A  (HAUKI) #5: Fri Aug 11

Sep  7 06:34:03 news /netbsd: mb_map full
Sep  7 06:53:45 news /netbsd: NetBSD 1.0A (HAUKI) #7: Fri Sep  1

Logfiles on news start from Aug 31 17:00 on walrus from Aug 31 02:00:0, 
and these are all the cases of this message in the log.

Seems to me that no fsck happens here, savecore sees no coredump. I have
never seen the machine do this so i dont know if there is some information
printed on the console. 

i386(P5) with ncr-scsi and 2*de pci ethernet. Conf has maxusers at 64, 
GATEWAY also configured.

Other machine is serving interactive users and www. The other is news/dns 
server.

Anybody else seeing something like this ?

Any suggestions, will I hurt myself if i try to grow the NMBCLUSTERS
define ? 

-- 
 Jarkko.Torppa@cute.fi     +358-0-648090          Cute Communications Oy