Early in this thread, I was reminded that I should not run ntpdate when ntpd is running. I had a cron job running every hour, which would stop ntpd, run ntpdate, then restart ntpd. That of course was removed. Now, as it should, ntpd is started when the system boots and should maintain the clock, without having to restart it.
I've been working on the time problem in the background, running tests using various combinations of kernel (grub.conf) options and ntp.conf options. I've gotten things pretty close but it's still not dead on. Then I came across something, where ntpd is stopped and restarted every hour which isn't something I've added to the system. Looking at the system log, I see that something is shutting down ntpd every hour then restarting it. Jul 15 05:01:01 uc222 ntpd[1564]: ntpd exiting on signal 15 Jul 15 05:01:01 uc222 ntpd[2414]: ntpd 4.2....@1.1570-o Sat Dec 19 00:58:16 UTC 2009 (1) Jul 15 05:01:01 uc222 ntpd[2415]: precision = 1.000 usec Jul 15 05:01:01 uc222 ntpd[2415]: Listening on interface wildcard, 0.0.0.0#123 Disabled Jul 15 05:01:01 uc222 ntpd[2415]: Listening on interface wildcard, ::#123 Disabled Jul 15 05:01:01 uc222 ntpd[2415]: Listening on interface lo, ::1#123 Enabled Jul 15 05:01:01 uc222 ntpd[2415]: Listening on interface eth0, fe80::211:25ff:fe75:bf3c#123 Enabled Jul 15 05:01:01 uc222 ntpd[2415]: Listening on interface lo, 127.0.0.1#123 Enabled Jul 15 05:01:01 uc222 ntpd[2415]: Listening on interface eth0, 192.168.1.222#123 Enabled Jul 15 05:01:01 uc222 ntpd[2415]: kernel time sync status 0040 Jul 15 05:01:01 uc222 ntpd[2415]: frequency initialized 16.854 PPM from /var/lib/ntp/drift Jul 15 05:01:12 uc222 ntpd[2415]: synchronized to 69.50.231.130, stratum 2 Jul 15 05:01:12 uc222 ntpd[2415]: time reset +0.806621 s Jul 15 05:01:12 uc222 ntpd[2415]: kernel time sync enabled 0001 Jul 15 05:02:26 uc222 ntpd[2415]: synchronized to 129.174.93.11, stratum 3 Jul 15 05:08:40 uc222 ntpd[2415]: synchronized to 69.50.231.130, stratum 2 Jul 15 06:01:01 uc222 ntpd[2415]: ntpd exiting on signal 15 Jul 15 06:01:01 uc222 ntpd[2576]: ntpd 4.2....@1.1570-o Sat Dec 19 00:58:16 UTC 2009 (1) Jul 15 06:01:01 uc222 ntpd[2577]: precision = 1.000 usec Jul 15 06:01:01 uc222 ntpd[2577]: Listening on interface wildcard, 0.0.0.0#123 Disabled Jul 15 06:01:01 uc222 ntpd[2577]: Listening on interface wildcard, ::#123 Disabled Jul 15 06:01:01 uc222 ntpd[2577]: Listening on interface lo, ::1#123 Enabled Jul 15 06:01:01 uc222 ntpd[2577]: Listening on interface eth0, fe80::211:25ff:fe75:bf3c#123 Enabled Jul 15 06:01:01 uc222 ntpd[2577]: Listening on interface lo, 127.0.0.1#123 Enabled Jul 15 06:01:01 uc222 ntpd[2577]: Listening on interface eth0, 192.168.1.222#123 Enabled Jul 15 06:01:01 uc222 ntpd[2577]: kernel time sync status 0040 Jul 15 06:01:01 uc222 ntpd[2577]: frequency initialized 18.147 PPM from /var/lib/ntp/drift Jul 15 06:01:09 uc222 ntpd[2577]: synchronized to 207.171.30.106, stratum 1 Jul 15 06:01:09 uc222 ntpd[2577]: time reset +0.815484 s Jul 15 06:01:09 uc222 ntpd[2577]: kernel time sync enabled 0001 Jul 15 06:02:28 uc222 ntpd[2577]: synchronized to 207.171.30.106, stratum 1 It seems to be something sipx is running, perhaps /usr/bin/sipx-time-manager. I don't see anything in the /var/log/sipxpbx logs or /etc/sipxpbx/ that seem to be related to doing this. The above /usr/bin/sipx-time-manager file seems to read the ntp.conf file and upon recognizing changes, seems to restart ntpd. If this is the case, I have not been changing ntp.conf so not sure why this would occur. Thought I would ask about it because I'm still trying to figure out the clock problems on these blades. Mike _______________________________________________ sipx-users mailing list sipx-users@list.sipfoundry.org List Archive: http://list.sipfoundry.org/archive/sipx-users Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-users sipXecs IP PBX -- http://www.sipfoundry.org/