about time

[DE@PHOENIX.SCH.Symbolics.COM (Doug Evans)]

    Date: Thu, 08 Mar 90 12:02:14 EST
    From: bouma@cs.purdue.edu


	I have read with interest the articles about reseting network time
	as we have the same situation Steve Nicoud is experiencing. But none
	has answered the main question of how to figure out which machine is
	supplying the bogus time. #'net-print-host-times just prints the times
	on the lisp machines. What would help (me) a great deal is simply an
	explanation of how the machine figures out what time it is when it is
	booted. Doesn't the internal calendar clock keep running even when in
	the fep? Then why always go to the network to get the (wrong) time when
	booted? Can I at least isolate the problem to one of the machines in our
	namespace? As always, any help is greatly appreciated.

	Bill <bouma@cs.purdue.edu>

A record of who has replied to broadcasts is kept in the variable
CHAOS:*BRD-HISTORY*.  (Note that all addresses are printed in decimal.)
You can find the culprit directly responsible for your wrong time:

  Command: CHAOS:*BRD-HISTORY*
  (("TIME" 9920))
  Command: (zl:hostat 9920)
  Chaosnet host status report.  Type Control-Abort to quit.
  Site   Name/Status              Subnet     #-in    #-out   abort lost     crc ram bitc other
  23300  Phoenix                      46  6413917  1755663       0    0      26   0   20     3 
  NIL