[ntp:questions] Re: Massive Jitter & Drift Preventing Syncronization?

Richard B. Gilbert rgilbert88 at comcast.net
Wed Dec 21 14:35:13 UTC 2005


Dan Check wrote:

> Hi,
>
> I'm running into terrible system clock problems with a Sun Fire X2100  
> running a Debian etch kernel (2.6.12-1-amd64-k8-smp #1 SMP Wed Sep 28  
> 02:57:49 CEST 2005 x86_64 GNU/Linux).  The system clock falls out of  
> sync really aggressively -- at a rate of a couple of minutes every  
> hour.  Any help on this would be hugely appreciated.
>
> NTP doesn't work at all unless I add the burst and iburst options to  
> the servers, and then it seems to sync periodically,  but it doesn't  
> do the gentle time correction that I've come to know and love from  
> ntp.  Is this because my system clock is naturally too far out to  
> lunch?  Is there a different way that I can get it to a baseline of  
> acceptable performance, and would that help NTP to do its job?
>
> I've currently got ntp requests going out over both eth1 and eth0,  
> onto the internet and onto a local subnet, so I don't think network  
> problems are a root cause here.
>
> Here's some ntpq output (notice the massive jitter):
>
> cairo:~# ntpq -p
>      remote           refid      st t when poll reach   delay    
> offset  jitter
> ======================================================================== 
> ======
> eddie.psaux.com 69.25.96.13      2 u   41   64  377  126.531  -178278  
> 1469.35
> snm.sd.dreamhos 164.67.62.194    2 u    1   64  377   76.850  -174947  
> 5826.25
> dns.ziandra.net 192.43.244.18    2 u    1   64  377  146.722  -187680  
> 9415.62
> 69.38.165.70    .STEP.          16 u    - 1024    0    0.000    0.000  
> 4000.00
> 69.38.165.70    17.254.0.27      3 u    2   64  377   32.126  -174999  
> 8972.49
> cudns.cit.corne 192.43.244.18    2 u    2   64  377   68.414  -188343  
> 9871.54
> hickory.cc.colu 18.145.0.30      2 u    2   64  377   15.159  -176542  
> 7009.97
> 192.168.1.202   216.165.129.244  2 u    1   64  377    0.089  -176454  
> 7924.23
>
> And here are the logs... notice the frequent large syncs...  And it  
> gives me a "kernel time sync disabled" but never says "kernel time  
> sync enabled" after that, which is the behavior that I'm used to...
>
> Dec 19 14:28:41 cairo ntpd[24714]: ntpd 4.2.0a at 1:4.2.0a+stable-2-r  
> Fri Feb  4 19:20:53 CET 2005 (1)
> Dec 19 14:28:41 cairo ntpd[24714]: precision = 1.000 usec
> Dec 19 14:28:41 cairo ntpd[24714]: Listening on interface wildcard,  
> 0.0.0.0#123
> Dec 19 14:28:41 cairo ntpd[24714]: Listening on interface  wildcard, 
> ::#123
> Dec 19 14:28:41 cairo ntpd[24714]: Listening on interface lo, XXXXXX#123
> Dec 19 14:28:41 cairo ntpd[24714]: Listening on interface eth0,  
> XXXXXXXX#123
> Dec 19 14:28:41 cairo ntpd[24714]: Listening on interface eth1,  
> XXXXXXXX#123
> Dec 19 14:28:41 cairo ntpd[24714]: kernel time sync status 0040
> Dec 19 14:28:41 cairo ntpd[24714]: frequency initialized 0.000 PPM  
> from /var/lib/ntp/ntp.drift
> Dec 19 14:29:19 cairo ntpd[24714]: synchronized to 66.33.216.11,  
> stratum 2
> Dec 19 14:29:19 cairo ntpd[24714]: time reset -53.294103 s
> Dec 19 14:29:19 cairo ntpd[24714]: kernel time sync disabled 0041
> Dec 19 14:32:46 cairo ntpd[24714]: synchronized to 69.38.165.70,  
> stratum 3
> Dec 19 14:33:37 cairo ntpd[24714]: synchronized to 128.59.59.177,  
> stratum 2
> Dec 19 14:33:39 cairo ntpd[24714]: no servers reachable
> Dec 19 14:33:50 cairo ntpd[24714]: synchronized to 69.38.165.70,  
> stratum 3
> Dec 19 14:33:52 cairo ntpd[24714]: synchronized to 128.59.59.177,  
> stratum 2
> Dec 19 14:34:42 cairo ntpd[24714]: synchronized to 69.38.165.70,  
> stratum 3
> Dec 19 14:34:42 cairo ntpd[24714]: synchronized to 66.92.250.40,  
> stratum 2
> Dec 19 14:34:44 cairo ntpd[24714]: no servers reachable
> Dec 19 14:34:57 cairo ntpd[24714]: synchronized to 66.92.250.40,  
> stratum 2
> Dec 19 14:35:47 cairo ntpd[24714]: synchronized to 66.33.216.11,  
> stratum 2
> Dec 19 14:35:52 cairo ntpd[24714]: no servers reachable
> Dec 19 14:37:17 cairo ntpd[24714]: synchronized to 66.33.216.11,  
> stratum 2
> Dec 19 14:38:07 cairo ntpd[24714]: no servers reachable
> Dec 19 14:40:27 cairo ntpd[24714]: synchronized to 132.236.56.250,  
> stratum 2
> Dec 19 14:41:16 cairo ntpd[24714]: synchronized to 66.33.216.11,  
> stratum 2
> Dec 19 14:41:21 cairo ntpd[24714]: no servers reachable
> Dec 19 14:41:30 cairo ntpd[24714]: synchronized to 132.236.56.250,  
> stratum 2
> Dec 19 14:42:24 cairo ntpd[24714]: no servers reachable
> Dec 19 14:42:43 cairo ntpd[24714]: synchronized to 128.59.59.177,  
> stratum 2
> Dec 19 14:42:43 cairo ntpd[24714]: time reset -190.225502 s
> Dec 19 14:43:58 cairo ntpd[24714]: synchronized to 132.236.56.250,  
> stratum 2
> Dec 19 14:44:54 cairo ntpd[24714]: no servers reachable
> Dec 19 14:52:49 cairo ntpd[24714]: synchronized to 66.33.216.11,  
> stratum 2
> Dec 19 14:52:53 cairo ntpd[24714]: no servers reachable
> Dec 19 14:54:44 cairo ntpd[24714]: synchronized to 128.59.59.177,  
> stratum 2
> Dec 19 14:54:45 cairo ntpd[24714]: no servers reachable
> Dec 19 14:54:57 cairo ntpd[24714]: synchronized to 66.92.250.40,  
> stratum 2
> Dec 19 14:55:45 cairo ntpd[24714]: no servers reachable
> Dec 19 14:55:17 cairo ntpd[24714]: synchronized to 132.236.56.250,  
> stratum 2
> Dec 19 14:55:17 cairo ntpd[24714]: time reset -175.662019 s
> Dec 19 14:56:32 cairo ntpd[24714]: synchronized to 132.236.56.250,  
> stratum 2
> Dec 19 14:57:25 cairo ntpd[24714]: synchronized to 66.33.216.11,  
> stratum 2
> Dec 19 14:57:25 cairo ntpd[24714]: synchronized to 206.111.81.7,  
> stratum 2
> Dec 19 14:57:27 cairo ntpd[24714]: no servers reachable
> Dec 19 15:03:03 cairo ntpd[24714]: synchronized to 69.38.165.70,  
> stratum 3
> Dec 19 15:03:05 cairo ntpd[24714]: no servers reachable
> Dec 19 15:03:08 cairo ntpd[24714]: synchronized to 128.59.59.177,  
> stratum 2
> Dec 19 15:03:59 cairo ntpd[24714]: synchronized to 132.236.56.250,  
> stratum 2
> Dec 19 15:04:03 cairo ntpd[24714]: no servers reachable
> Dec 19 15:06:28 cairo ntpd[24714]: synchronized to 132.236.56.250,  
> stratum 2
> Dec 19 15:07:19 cairo ntpd[24714]: no servers reachable
>
>
> Thanks,
> Dan
> _______________________________________________
> questions mailing list
> questions at lists.ntp.isc.org
> https://lists.ntp.isc.org/mailman/listinfo/questions
>
My first guess would be that your local clock has a frequency error in 
excess of 512 parts per million.  500 PPM translates to about 43 seconds 
per day.  NTPcan not and will not work until you get that fixed.  I also 
find both the "clock hopping" and the "no servers reachable" messages  a 
little disturbing.
The no servers reachable suggests that you have network problems somewhere.

My second guess would be that you are listening to some bad servers 
which might account for the clock hopping.

How well, or how badly, does the system keep time when ntpd is not running?




More information about the questions mailing list