![](https://seccdn.libravatar.org/avatar/3035b38ff33cf86f480bb169b8500b80.jpg?s=120&d=mm&r=g)
https://bugzilla.novell.com/show_bug.cgi?id=344356#c8
Carlos Robinson
to comment #2:
why you have following entry? server pool.ntp.org server pool.ntp.org server pool.ntp.org server pool.ntp.org
Why not? The idea is not mine; I read it somewhere. Each time you query for a server in the pool, it gives a different server: cer@nimrodel:~> host pool.ntp.org pool.ntp.org has address 84.88.69.10 pool.ntp.org has address 213.194.159.3 pool.ntp.org has address 195.55.174.243 cer@nimrodel:~> host pool.ntp.org pool.ntp.org has address 195.55.174.243 pool.ntp.org has address 84.88.69.10 pool.ntp.org has address 213.194.159.3 or: cer@nimrodel:~> host es.pool.ntp.org es.pool.ntp.org has address 195.10.6.126 es.pool.ntp.org has address 80.59.234.233 es.pool.ntp.org has address 80.33.117.152 cer@nimrodel:~> host es.pool.ntp.org es.pool.ntp.org has address 80.33.117.152 es.pool.ntp.org has address 195.10.6.126 es.pool.ntp.org has address 80.59.234.233 The idea is to give the program a lot of servers for it to find which are the best and use them. I have seen in the past that some of the servers in the pool are pretty bad, so I list more servers so it can find a bunch of good ones.
found somewhere in comment #7:
[...] 29 Nov 00:31:30 ntpd[19587]: synchronized to LOCAL(0), stratum 10 [...]
and
[...] remote local st poll reach delay offset disp ======================================================================= =192.33.214.57 192.168.1.12 2 1024 377 0.13612 295.28592 0.13649 *LOCAL(0) 127.0.0.1 10 64 377 0.00000 0.000000 0.03065 [...]
looks like your hardware clock is broken. but ntp knows that the time difference between server and local is to huge and sync to the bad local. when you restart ntpdate is called which forces the time correction so ntp works fine until the time diffs to much...
something in the system is wrong but it doesn't looks like ntp is the problem.
Remember that this doesn't happen always, not every day. Once it happened right while I was working; I think the machine stopped for 5 seconds while scanning a photo via USB and the clock started to go slow. Or it can be correct for days and days... in a month it has happened 6 times: cer@nimrodel:~> grep "exceeds sanity limit" /var/log/ntp 4 Nov 15:50:18 ntpd[5076]: time correction of 1190 seconds exceeds sanity limit (1000); set clock manually to the correct UTC time. 12 Nov 09:37:12 ntpd[5061]: 14 Nov 20:42:37 ntpd[5143]: 21 Nov 23:38:27 ntpd[5170]: 27 Nov 16:22:22 ntpd[12905]:
you told this happens if you are at lunch. do you have a program which starts to run if the computer is idle?
The screen saver. I'll be careful to leave it on during lunch to see if it consistently happens then. If that is consistent, I can disable the screen saver - but I can't believe a user's space program can interfere with kernel clock.
And yes, my guess is that ntp is not broken and that there is something wrong with the kernel maintained clock, but I have no idea what, or even what to do to see it. Maybe the kernel chooses the wrong type of clock during boot. It certainly started to go bad the following day to upgrading to 10.3. boot log extract: cer@nimrodel:~> grep -i clock /var/log/boot.msg <6>Time: tsc clocksource has been installed. <6>Real Time Clock Driver v1.12ac <6>Time: acpi_pm clocksource has been installed. <6>intel8x0_measure_ac97_clock: measured 50655 usecs <6>intel8x0: clocking to 48000 doneSetting up the hardware clockdone When you say "looks like your hardware clock is broken" - I don't suppose you mean the CMOS hardware clock? That does not affect after the system is up, as the time is by them correctly set up and the hardware clock is not consulted any more till next boot. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.