3

I installed ntp on a centos server.

I modified /etc/ntp.conf replacing the default server entries with one entry to my internal NTP server's IP (1.2.3.4 is the IP of the NTP server): 

server 1.2.3.4 iburst

When I reboot the server and run ntpstat I see this:

unsynchronised
   polling server every 64 s

Then I run sudo ntpdate 1.2.3.4 and see this message:

1 Dec 19:22:35 ntpdate[2339]: the NTP socket is in use, exiting

Then I stop the service (systemctl stop ntpd) and run sudo ntpdate 1.2.3.4 again to see it sync successfully:

1 Dec 19:25:11 ntpdate[2351]: adjust time server 1.2.3.4 offset 0.005316 sec

Now when I start the service back up and run ntpstat I see this message:

synchronised to NTP server (1.2.3.4) at stratum 4
   time correct to within 8065 ms
   polling server every 64 s

If I reboot now and run ntpstat it will say unsynchronized. What am I doing wrong? It should be automatically syncing to the IP of the NTP server I put in the config file right?

red888
  • 4,069
  • 16
  • 58
  • 104
  • how far off is it if you reboot again? as far as i can see, after running `ntpdate` and restarting ntpd, everything is fine - ntp makes small adjustments on its own after being 'roughed in' – pete Dec 02 '15 at 02:02
  • Wrong time zone on the client, out by too much for the version of ntpd running on it? Not saving time to BIOS on reboot? – tink Dec 02 '15 at 02:44

1 Answers1

1

Looks like I was just being impatient. After a reboot if I wait about 6 minutes and ran ntpstat it will show it is synchronized.

The polling interval says 64 s which I assumed meant it polls every 64 s, but it looks like it doesn't work like that exactly.

red888
  • 4,069
  • 16
  • 58
  • 104
  • 2
    Also, while the NTPD is running and you want to manually sync the time, you need to run this command instead so that it uses a different port (because the daemon is already using it): sudo ntpdate -u 1.2.3.4 – TheGrandPackard Dec 02 '15 at 17:42