Joerg Sonnenberger wrote:
On Fri, Jan 09, 2009 at 07:20:32PM +0100, Frank Kardel wrote:We restart it because the NTP server(s) to use may have changed. I'm pretty sure that NTP does not detect this on the fly.You are right. ntpd does not detect a change in conf files.Can you make ntpd reread the config file after SIGHUP? Joerg
There is currently no such code in there. One could try what happens if config reading is re-run at some arbitrary time, but I am not sure whether the right things happen when e. g. the crypto pwd is changed. changing peer configuration would certainly accumulate. A correct re-reading solution would probably mean some thingking and coding.In the ntp develepment version is is possible to send new config lines via ntpq. In the stable version (the one NetBSD ships) it is possible to configure or unconfigure peers via ntpdc.
That would be an alternative to re-start - but it is script trouble, Frank