Dew Wrobel wrote: > On Oct 5, 2:17 pm, "Richard B. Gilbert" <rgilber...@comcast.net> > wrote: >> Dew Wrobel wrote: >>> I have to setup a couple of servers that will get their time from the >>> internet. >>> I have following the steps listed athttp://www.pool.ntp.org/ >>> The following is the contents of ntp.conf: >>> driftfile /drift/ntp.drift >>> server 0.us.pool.ntp.org >>> server 1.us.pool.ntp.org >>> server 2.us.pool.ntp.org >>> server 3.us.pool.ntp.org >>> When I start NTP, the start up hands with ntpdate trying to get the >>> time from the servers. I have verified that the server names do >>> verify in DNS. >>> Do I need to pick a different set of servers? Any idea/suggestions >>> would be greatly appreciatd. >>> Thanks >> Ntpdate is "deprecated". Perhaps you should eliminate ntpdate and start >> ntpd with the "-g" option. This option tells ntpd to find out what time >> it is by querying the servers and then setting that time. >> >> The results should be similar either way but ntpd -g is the documented >> and supported way to set the time at startup. >> >> If you add "iburst" to each server line in your ntp.conf you should get >> a faster startup. Iburst will cause ntpd to send an initial burst of >> eight requests at two second intervals. The replies fill the "filter >> pipeline" and should get you synchronized a little faster. >> >> Ntpd will need about ten hours to achieve the accuracy it's capable of. >> Initially you should have a reasonable approximation of the correct >> time; e.g. within, say, 100 milliseconds. The longer it runs the >> better the time will get. >> >> If you can possibly avoid rebooting and/or restarting NTPD you will get >> much better time. > > The call to ntpdate is part of the RC script that comes with the OS. > I checked an option under /etc/sysconfig/ntp to not call ntpdate on > start.
Check the version of the NTPD you have. A lot of companies ship NTP Version 3.x while NTP is now at V4.2 (I think!). > > I found a web page about debugging NTP and came across running ntpq > with various parameters. here is the output from that. > I can't help and wondering, based on the as option, it sounds that the > time servers being used are reject. > > Aside from what servers I'm using, I don't have to do anything else, > do I? > > ntpq> pe > remote refid st t when poll reach delay > offset jitter > ============================================================================== > ntp1.truetime.c .INIT. 16 u - 64 0 0.000 > 0.000 0.001 > zinc.ops.tns.it .INIT. 16 u - 64 0 0.000 > 0.000 0.001 > ntp2.usno.navy. .INIT. 16 u - 64 0 0.000 > 0.000 0.001 > ntpq> as The "reach" field says you are not getting replies from any of your chosen servers! The likeliest reasons for that is that are that: 1. Your requests are not reaching the servers 2. The replies to your requests are not reaching you. 3. You didn't wait long enough after starting NTPD before running NTPQ. Check with your network people. NTP uses port 123 which is a privileged port. Your firewall may be blocking it. You need to allow both incoming and outgoing traffic on port 123. In your NTP.CONF file you can add the keyword "IBURST" to each of your server statements. This will cause the first eight requests to be sent at intervals of two seconds. Subsequent packets will be sent a rates ranging from 64 seconds to 1024 seconds; the rates are determined by NTPD and will be adjusted as necessary from time to time! _______________________________________________ questions mailing list questions@lists.ntp.org https://lists.ntp.org/mailman/listinfo/questions
