Uwe Klein wrote:
semi OT ( either OnTopic or OffTopic ;-): What happens if I kill ntpd during clock slewing ? What happens if I SIGTERM ntpd during clock slewing ?
I don't think you have fully constrained the problem. In some environments the slew will stop, although I'm not sure whether it will stop in about a second or when the whole slew finishes.
Without actually reading the code, I'm not sure what will happen when using the kernel time discipline on a system that supports it.
According to my Linux man pages, kill(2) is Posix, but adjtime(2) is not, so there can exist Posix compliant systems in which an ntpd implementation would continue to slew indefinitely if the kill was sufficiently hard. (adjtimex(2) is Linux only and ntp_adjtime(2) has no man page and no alias in adjtimex.)
Linux implementations, at least, will retain frequency corrections in kernel discipline mode and lose them in user space discipline mode.
_______________________________________________ questions mailing list [email protected] http://lists.ntp.org/listinfo/questions
