Hello folks.  We're running a production mail server under almost
generic 1.2/i386.  The machine is quite stable, if not the best performer
under high memory loads, but last week it crashed after beeing up for 110
days of intense use with the message:

Jun 24 22:03:20 rumpleteazer savecore: reboot after panic: vm_map_entry_create: out of map entries

	After looking in src/sys/vm/vm_map.c, I see the panic message and
understand that it means that the machine ran out of vm map entries in one
of its memory maps, but how can I tell which map it was and how many slots
are in each of the maps which are maintained?  Am I correct to assume that
two vm map structures: the hand crafted one for kernel vertual memory, and
the more dynamic one for user virtual memory?  What are the rules for
constructing these structures and how can I monitor their utilization
during the time the machine is running?
	Since the machine moved over 100,000 messages a day for over 100 days
without a hickup, I suppose this isn't a burning isue, but I am curious to
understand more fully how the vm stuff works under the hood.

	Any help or enlightenment on how to find help on understanding what
happened here would be greatly appreciated.
-thanks
-Brian