Ambresh Ranjan wrote: > Problem here is ; i have to synchronize all machine to 1 millisecond > accuracy.
Expect it to take several hours to regain this accuracy, following any significant disruption. > and The path they are connected is unreliable . > So what signal is handled by NTP when service stop/start/restart command is > given . You need to ask Red Hat that; "service" is a Red Hat (and derivatives) specific command. As far as I can tell, for 4.2.4p4, at least, the ntpd distribution supplied scripts are totally obsolete and refer to the, obsolete, xntpd, name. However, I think the basic answer is that: - start runs it without using any signals; - stop uses SIGKILL, which is not caught; - restart is a combination of stop and start (the, obsolete, supplied script doesn't support restart). It is questionable as to whether restart is useful, unless you have changed the configuration file or DNS configuration. You can use the NTP management protocol (as implemented by ntpq) to add and delete servers, but not to do a total restart. _______________________________________________ questions mailing list [email protected] https://lists.ntp.org/mailman/listinfo/questions
