} >> You're only making the problem worse by running the time service too, as soon as } >> I have time I'll be making my server return 1999-12-31 23:59:59 to time queries.
I don't perceive people using the time (port 37) service to be a problem, as long as they don't poll too often. It may not be the optimal way to keep the time on their computer, but it's their computer, not mine, so it's not my problem. NIST offers ntp (137/udp), time (37/tcp and udp), and daytime (13/tcp) on all it's time servers: http://tf.nist.gov/service/its.htm So NIST has set a precedent. I see no reason to fight it. There are a number of windows programs using time and/or daytime. "atomtime" is one that defaults to daytime and is apparently popular. A while back I also turned on daytime, and sure enough a handful of people are using it now. The time service gets about ten times the traffic as daytime, but both are insignificant. daytime is defined by RFC 867, which reads "There is no specific syntax for the daytime". The format for daytime in most Linux boxes is stupid (no timezone), so I changed it. This is my /etc/inetd.conf entry: daytime stream tcp nowait nobody /bin/date date -uR sample output: Thu, 20 Apr 2006 00:32:09 +0000 } service time } { } socket_type = stream } protocol = tcp } user = nobody } wait = no } port = 37 } server = /bin/echo } server_args = -ne \274\027\301\377 } log_on_success = } log_on_failure = } instances = 5 } } What is \274\027\301\377 ? I had to try this out of curiosity and got: Thu 06 Jan 1949 09:13:24 AM PST -- E Frank Ball [EMAIL PROTECTED] _______________________________________________ timekeepers mailing list [email protected] https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
