I asked: > I have a problem with the system clock. I had to change the time for a > DST test some developers had to do, but I cannot get the time back close > enough for ntpd to resume its duties. > > First I manually set the time to get it within a few seconds of the correct > time. Then I ran ntpdate <host> to get the time exact, with the intention > of running ntp once the time was very close. The problem is that ntpdate > tells me what the offset would be, but does not change the time. Here is > what happens when I run ntpdate three times with a "date" command in rapid > succession: > > 11 orwebtest1!/etc/init.d >> ntpdate wintermute.tsg.cbot.com > 15 Feb 09:09:30 ntpdate[1165]: adjust time server 164.74.19.238 offset -86195.033568 sec > 12 orwebtest1!/etc/init.d >> date > Thu Feb 15 09:09:33 CST 2007 > 13 orwebtest1!/etc/init.d >> ntpdate wintermute.tsg.cbot.com > 15 Feb 09:09:40 ntpdate[1169]: adjust time server 164.74.19.238 offset -86194.336740 sec > 14 orwebtest1!/etc/init.d >> date > Thu Feb 15 09:09:43 CST 2007 > 15 orwebtest1!/etc/init.d >> ntpdate wintermute.tsg.cbot.com > 15 Feb 09:09:52 ntpdate[1175]: adjust time server 164.74.19.238 offset -86193.553653 sec > > So as you can see it is detecting a big gap in time there (offset in the > 80 thousand second range) but it is not correcting it. If I try to run > ntp, which should do an ntpdate first, As soon as it wants to sync it fails: > > ntpq> pe > remote refid st t when poll reach delay offset disp > ============================================================================== > ptstcon1.prices c10-srvcon3.cbo 3 u 61 64 17 1.53 -862165 1875.02 > wintermute.tsg. c10-srvcon3.cbo 3 u 60 64 17 3.37 -862164 1875.02 > admdev1.admin.c c10-srvcon3.cbo 3 u 60 64 17 3.27 -862164 1875.02 > oetest1.oe.cbot 0.0.0.0 16 - - 64 0 0.00 0.000 16000.0 > acodstst1.cbot. ptstcon1.prices 4 u 60 64 17 0.78 -862164 1875.02 > > (look at those offsets) > but when the disp is low enough to sync, ntp aborts and on the console is > > Feb 15 09:05:11 orwebtest1 xntpd[484]: time error -86212.266372 is way too large (set clock manually) > > A reboot did NOT help, btw. The exact same thing is happening after a full > reboot. Does anyone have a suggestion on how to get the time back to normal > on this server? The answer: Although I have not received the question myself, enough others have to see that I am having difficulty reading a calendar today. Today is the 14th, not the 15th. The slew of 862164 is awfully close to 60*60*24=86000, which is 24 hours. When I set the date to within a few seconds, the ntp had no problem at all. Its interesting that ntpdate couldn't correct me, but thats not what it is supposed to do... +-----------------------------------------------------------------------+ | Christopher L. Barnard O When I was a boy I was told that | | cbarnard@tsg.cbot.com / \ anybody could become president. | | (312) 347-4901 O---O Now I'm beginning to believe it. | | http://www.cs.uchicago.edu/~cbarnard --Clarence Darrow | +----------PGP public key available via finger or PGP keyserver---------+ _______________________________________________ sunmanagers mailing list sunmanagers@sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagersReceived on Wed Feb 14 15:31:23 2007
This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:44:04 EST