Peter Stickney wrote:
OK, Ive got kernel 2.6.9, slackware 10.0. ntpd 4.2.0. This one box is the NTP server for a small ( 60 ) Win2000 network with 5 more linux servers also getting their time from this one box. The issue looks like the box doesnt want to sync to pool.ntp.org servers with the daemon. BUT. If I run 'ntpdate pool.ntp.org' it syncs fine.
No, it doesn't really synchronize. It merely gets the time from one of the pool systems (who knows which one) and sets the local clock to that. That's the biggest difference between ntpd and ntpdate.
As a result, none of the clients will sync with this one because they know it's off. Windows says "The NTP server isn't sync'd, time not set."
Correct, they are acting properly. Your NTP server is not synchronized because you didn't set it up properly so that clients are right to refuse to use this time.
The server looks like this: ntp.conf server us.pool.ntp.org server 127.127.1.1 restrict 192.168.0.1 mask 255.255.240.0 nomodify nopeer notrap restrict 127.0.0.1 nomodify fudge 127.127.1.1 stratum 10 driftfile /etc/ntp.drift logfile /var/log/ntp.log
Rule #1 of NTP is that you need at least 3 servers. You only specified one. It's not going to be able to work this way.
snip ntp.log 12 Jul 19:22:39 ntpd[25341]: synchronized to LOCAL(1), stratum=10 12 Jul 19:28:02 ntpd[25341]: synchronized to 69.44.57.60, stratum=3 12 Jul 19:36:36 ntpd[25341]: synchronized to LOCAL(1), stratum=10 12 Jul 19:40:51 ntpd[25341]: synchronized to 69.44.57.60, stratum=3 12 Jul 19:42:14 ntpd[25341]: synchronized to LOCAL(1), stratum=10 12 Jul 19:45:09 ntpd[25341]: synchronized to 69.44.57.60, stratum=3 12 Jul 19:58:02 ntpd[25341]: synchronized to LOCAL(1), stratum=10 12 Jul 20:00:13 ntpd[25341]: synchronized to 69.44.57.60, stratum=3 12 Jul 20:15:47 ntpd[25341]: time reset +3.526547 s 12 Jul 20:20:07 ntpd[25341]: synchronized to LOCAL(1), stratum=10 12 Jul 20:25:26 ntpd[25341]: synchronized to 69.44.57.60, stratum=3 12 Jul 20:42:47 ntpd[25341]: time reset +1.678287 s 12 Jul 20:47:05 ntpd[25341]: synchronized to LOCAL(1), stratum=10 12 Jul 21:01:01 ntpd[25341]: synchronized to 69.44.57.60, stratum=3 12 Jul 21:18:05 ntpd[25341]: synchronized to LOCAL(1), stratum=10 12 Jul 21:19:11 ntpd[25341]: synchronized to 69.44.57.60, stratum=3 12 Jul 21:36:28 ntpd[25341]: time reset +2.766987 s 12 Jul 21:40:47 ntpd[25341]: synchronized to LOCAL(1), stratum=10
Notice the way this is synchronizing, switching back and forth between the remote and local clock, a bad sign.
I set the sntp of the windows boxes at logon with net time /setsntp:lovejoy [ lovejoy being the server in question ] the other linux servers have ntp.conf #lovejoy server 192.168.1.14 restrict default noquery notrust nomodify restrict 127.0.0.1 restrict 192.168.1.0 mask 255.255.255.0 fudge 127.127.1.0 stratum 10 server 127.127.1.0 driftfile /etc/ntp.drift logfile /var/log/ntp.log
Again just one server? Local doesn't count.
It seems that lovejoy is synching, but in fact it isnt. The other linux servers have stuff like... 15 Jul 16:39:25 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM 15 Jul 16:43:42 ntpd[15930]: synchronized to LOCAL(0), stratum=5 15 Jul 16:43:42 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM 15 Jul 16:44:46 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM 15 Jul 16:45:52 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM 15 Jul 16:46:56 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM 15 Jul 16:48:00 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM 15 Jul 16:49:04 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM 15 Jul 16:49:04 ntpd[15930]: synchronized to 192.168.1.14, stratum=11 15 Jul 17:07:34 ntpd[15930]: time reset -0.868923 s 15 Jul 17:07:34 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM 15 Jul 17:11:53 ntpd[15930]: synchronized to LOCAL(0), stratum=5 15 Jul 17:11:53 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM 15 Jul 17:12:58 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM 15 Jul 17:14:03 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM 15 Jul 17:15:08 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM 15 Jul 17:16:14 ntpd[15930]: frequency error -512 PPM exceeds tolerance 500 PPM
You need to start ntpd with the -g option to allow it to tolerate the large error for the first time only.
15 Jul 17:17:11 ntpd[15930]: synchronized to 192.168.1.14, stratum=11 15 Jul 17:32:22 ntpd[15930]: time reset -0.834325 s So it looks like it's synching to lovejoy [ 192.168.1.14 ] but at a very low strat. And those PPM errors go away afer a nice ntpdate on lovejoy.
Of course. The clock is too far off and you didn't start it with -g.
The windows boxes, forget it, they dont want to synch at all. Event Log: The NTP server %1 isn't sync'd, time not set
Correct, lovejoy is not synchronized so it shouldn't
BUT. I can go to lovejoy, killall ntpd, ntpdate us.pool.ntp.org, restart ntpd and that fixes it. Obviously. Even the windows boxes will come back:
That just avoids the problem because now the clock is close enough. You need to start with -g.
Event Log: Time service corrected the clock error by 266 seconds So it looks like it should be working, but it just isnt. I would assume because ntpdate works and synchs everything, that its no networking issue, ports opened or whatever. I just ntpdated a few mins ago, so let's see what it looks like on monday. Who's in synch with who or whatever. Any help would be greatly appreciated.
If you follow the basic rules in setting things up and include at least servers you won't have this problem.
Danny _______________________________________________ questions mailing list questions@lists.ntp.isc.org https://lists.ntp.isc.org/mailman/listinfo/questions
