Ok,

when a fork fails, its normally:

1/	out of process table slots
2/	exceeded # procs per user
3/	out of swap
4/	out of pty's
5/	file table full
6/	page translation tables

I'm sure there are others but I think the above are in a reasonable
order of most likely first...

Several times today fork has failed due to some anonymous resource
shortage. But there are only 17 pty's in use, 74 processes running,
and pstat et al suggest no shortages:

# the following output was generated straight after one failure.

root:4025$ sysctl kern
kern.ostype = NetBSD
kern.osrelease = 1.0A
kern.osrevision = 199306
kern.version = NetBSD 1.0A (ZEN) #0: Wed Mar 29 22:56:59 EST 1995
    root@zen.void.oz.au:/d2/current/src/sys/arch/i386/compile/ZEN

kern.maxvnodes = 1354
kern.maxproc = 1044
kern.maxfiles = 3404
kern.argmax = 262144
kern.securelevel = 1
kern.hostname = zen.void.oz.au
kern.hostid = 0
kern.clockrate: tick = 10000, tickadj = 40, hz = 100, profhz = 100, stathz = 100
kern.posix1version = 198808
kern.ngroups = 16
kern.job_control = 1
kern.saved_ids = 0
kern.boottime = Wed Mar 29 23:50:39 1995

kern.domainname = 
kern.maxpartitions = 8

root:4026$ pstat -T
235/3404 files
   1351 vnodes
22M/55M swap space

root:4028$ pstat -sk
Device      1K-blocks     Used    Avail Capacity  Type
/dev/wd0b       32640    17156    15484    53%    Interleaved
/dev/sd0b       24491     6580    17911    27%    Interleaved
/dev/sd1b           0  *** not available for swapping ***
/dev/sd2b       32668  *** not available for swapping ***
Total           57131    23736    33395    42%

What have I missed and which tools might give a clue?

--sjg