Hi, I wanted to let my ntp client use the servers it receives from the
dhcp server (3.0 from packages). I made it working like this:
I added ntp-servers to /etc/dhclient.conf, and the following function to
the /sbin/dhclient-script script:
add_new_ntp() {
if [ -n "$new_ntp_servers" ]; then
sed "/^server/d" /etc/ntpd.conf > /etc/ntpd.conf.new
for address in $new_ntp_servers; do
echo "server" $address >> /etc/ntpd.conf.new
done
mv /etc/ntpd.conf.new /etc/ntpd.conf && pkill -KILL ntpd && ntpd
fi
}
I use add_new_ntp after add_new_routes in BOUND|RENEW|REBIND|REBOOT)
Is this the right way to do let ntp use dhclient?
And, is what I did in add_new_ntp() the best way to do it?
It seems that the script works. After executing "/bin/sh /etc/netstart
bce0" I see the following in /var/log/daemon:
May 2 22:35:51 kaat dhclient[30663]: DHCPREQUEST on bce0 to
255.255.255.255 port 67
May 2 22:35:51 kaat dhclient[30663]: DHCPACK from 127.0.0.1
May 2 22:35:51 kaat ntpd[2737]: Lost child: child terminated; signal 9
(Killed)
May 2 22:35:51 kaat ntpd[27146]: ntp engine ready
May 2 22:35:51 kaat dhclient[30663]: bound to 192.168.1.80 -- renewal
in 7200 seconds.
May 2 22:36:11 kaat ntpd[27146]: peer 192.168.1.7 now valid
May 2 22:36:14 kaat ntpd[27146]: peer 192.168.1.6 now valid
May 2 22:38:13 kaat ntpd[27146]: clock is now synced
May 2 22:52:39 kaat ntpd[12672]: adjusting clock frequency by
-12.976140 to -12.976140ppm
But I have two additional questions about this:
1) my dhcpd runs at 192.168.1.5 (remote), so why DHCPACK from 127.0.0.1?
2) "adjusting clock frequency by -12.976140 to -12.976140ppm." These
values are the same, what just happend?
Note: I don't run a DB server on the machine, so I don't think playing
with the time hurts something. But anyway, I just want to be sure I get
it right.
Regards,
Tom
