[ntp:questions] Frequency Errors +500 PPM...

Richard B. Gilbert rgilbert88 at comcast.net
Fri Feb 4 00:41:29 UTC 2005


Kit Plummer wrote:

>
>
> On Feb 3, 2005, at 3:42 PM, Richard B. Gilbert wrote:
>
>> Kit Plummer wrote:
>>
>>> FWIW, here is a snip of my ntp.log file:
>>>
>>>  3 Feb 10:00:14 ntpd[412]: kernel time sync enabled 0001
>>>  3 Feb 10:04:33 ntpd[412]: synchronized to GPS_BANC(0), stratum=0
>>>  3 Feb 10:36:45 ntpd[412]: time reset -0.873229 s
>>>  3 Feb 10:36:45 ntpd[412]: frequency error -512 PPM exceeds 
>>> tolerance   500 PPM
>>>  3 Feb 10:41:02 ntpd[412]: synchronized to GPS_BANC(0), stratum=0
>>>  3 Feb 11:47:44 ntpd[412]: no servers reachable
>>>  3 Feb 11:48:49 ntpd[412]: synchronized to GPS_BANC(0), stratum=0
>>>  3 Feb 12:03:51 ntpd[412]: time reset -0.871702 s
>>>  3 Feb 12:03:51 ntpd[412]: frequency error -512 PPM exceeds 
>>> tolerance   500 PPM
>>>  3 Feb 12:08:09 ntpd[412]: synchronized to GPS_BANC(0), stratum=0
>>>
>>> What can I do to control the PPM?  This is a SPARC-based VME 
>>> board,   getting its clock from a VME-based SYNCCLOCK32 GPS/IRIG card.
>>>
>>> TIA,
>>> Kit
>>>
>>> On Feb 3, 2005, at 9:46 AM, Kit Plummer wrote:
>>>
>>>> Hey folks,
>>>>
>>>> I am having NTP prolbems.  Obviously.  But, I am not sure it is 
>>>> NTP,   or my reference driver - or hardware for that matter.  The 
>>>> problem   appears to be related to an extremely high PPM.  Am I 
>>>> understanding   correctly that this number is derived from whatever 
>>>> is in the   ntp.drift file?
>>>>
>>>> Well, the real problem appears to be a 1 second step - 
>>>> sporadically.    Though I haven't really been able to prove that 
>>>> either...it is  just  heresay from a system user.
>>>>
>>>> Also,  can someone explain this output from ntpd -d:
>>>>
>>>> peer GPS_BANC(0) event 'event_reach' (0x84) status 'unreach, conf, 
>>>> 1   event, event_reach' (0x8014)
>>>>
>>>> If there is some other information I can provide for 
>>>> troubleshooting   purposes let me know.
>>>>
>>>> TIA,
>>>> Kit
>>>> --------------------------------------------------------------------- 
>>>> -- -
>>>> Kit Plummer
>>>> Operations Research and System Performance Dept.
>>>> Raytheon Missile Systems
>>>>
>>>> _______________________________________________
>>>> questions mailing list
>>>> questions at lists.ntp.isc.org
>>>> https://lists.ntp.isc.org/mailman/listinfo/questions
>>>>
>>>
>> Please don't "top post".
>>
>> Do you encounter this problem if you start ntpd with no drift file 
>> and  with the clock set to within a few seconds of the correct 
>> time?   If  you do, the first thing to do is to get the hardware 
>> fixed.  An error  of 512ppm works out to gaining or losing something 
>> like 48 seconds per  day!   Now few people will argue that the 
>> typical undisciplined  computer clock is not a  "POS"  but one that 
>> is that bad is definitely  broken!
>>
>> I've run ntpd on something like thirty different computers including  
>> Sun/Solaris/Sparc, Compaq/Solaris/X86, DEC Alpha/VMS and  
>> Compaq/RedHat/X86 without encountering this error.
>>
>> Additional helpful information would include the hardware platform,   
>> the operating system and version, and the version of ntpd that you 
>> are  running.
>> _______________________________________________
>> questions mailing list
>> questions at lists.ntp.isc.org
>> https://lists.ntp.isc.org/mailman/listinfo/questions
>>
>
> Is this better?  I suppose you don't like inline comments  
> either...interesting, everyone's got their preferences...
>
> Keep in mind that I've hacked the refclock_bancomm.c to work with my  
> VME-based SYNCCLOCK32 board.  Though I am not thinking that the 
> driver  is the problem.
>
> Another few data points worth mentioning is that the VME chassis is  
> only on for a short amount of time, maybe 4 hours a day max.  I do 
> have  similar problems with a development chassis that is connected 
> all the  time...but, not as severe.
>
> The SBC is a Force CPU-56, running Solaris 8.  The NTP version that  
> I've modified is 1.4.2.
>
> Here are a couple of snips of potentially worthy info (sorry for the  
> length of garbage):
>
> <ntp.log>
> 3 Feb 06:06:35 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 06:10:57 ntpd[637]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 06:10:57 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 06:12:02 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 06:13:08 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 06:14:11 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 06:15:15 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 06:16:18 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 06:17:22 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 06:18:26 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 06:35:30 ntpd[637]: time reset +0.983891 s
>  3 Feb 06:39:48 ntpd[637]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 07:11:59 ntpd[637]: time reset -1.021094 s
>  3 Feb 07:11:59 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:16:14 ntpd[637]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 07:16:14 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:17:20 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:18:23 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:19:29 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:20:33 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:21:39 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:22:45 ntpd[637]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:23:51 ntpd[637]: no servers reachable
>  3 Feb 07:24:56 ntpd[637]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 07:38:58 ntpd[637]: time reset +0.987694 s
>  3 Feb 07:38:58 ntpd[637]: frequency error 512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:43:15 ntpd[637]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 07:43:15 ntpd[637]: frequency error 507 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:44:19 ntpd[637]: frequency error 506 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:45:22 ntpd[637]: frequency error 505 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:46:28 ntpd[637]: frequency error 504 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:47:34 ntpd[637]: frequency error 503 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:48:38 ntpd[637]: frequency error 502 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:49:41 ntpd[637]: frequency error 501 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 07:50:45 ntpd[637]: frequency error 500 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 08:13:52 ntpd[637]: ntpd exiting on signal 15
>  3 Feb 08:14:47 ntpd[1844]: ntpd exiting on signal 15
>  3 Feb 08:19:34 ntpd[191]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 08:19:34 ntpd[191]: time reset -0.395297 s
>  3 Feb 08:19:34 ntpd[191]: kernel time sync disabled 0041
>  3 Feb 08:23:49 ntpd[191]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 08:30:29 ntpd[191]: ntpd exiting on signal 15
>  3 Feb 08:30:31 ntpd[383]: configure: keyword "logconfig=syncall"  
> unknown, line ignored
>  3 Feb 08:32:57 ntpd[383]: ntpd exiting on signal 15
>  3 Feb 08:36:11 ntpd[397]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 08:36:11 ntpd[397]: time reset -0.340365 s
>  3 Feb 08:36:11 ntpd[397]: kernel time sync disabled 0041
>  3 Feb 08:40:26 ntpd[397]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 08:51:17 ntpd[397]: time reset -0.349002 s
>  3 Feb 08:51:17 ntpd[397]: kernel time sync enabled 0001
>  3 Feb 08:55:37 ntpd[397]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 09:02:36 ntpd[397]: ntpd exiting on signal 15
>  3 Feb 09:05:29 ntpd[419]: ntpd exiting on signal 15
>  3 Feb 09:10:20 ntpd[181]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 09:10:20 ntpd[181]: kernel time sync disabled 0041
>  3 Feb 09:11:25 ntpd[181]: kernel time sync enabled 0001
>  3 Feb 09:12:06 ntpd[181]: ntpd exiting on signal 15
>  3 Feb 09:12:07 ntpd[358]: configure: keyword "fileget" unknown, line  
> ignored
>  3 Feb 09:14:36 ntpd[358]: ntpd exiting on signal 15
>  3 Feb 09:17:53 ntpd[369]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 09:22:08 ntpd[369]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 09:33:54 ntpd[369]: kernel time sync enabled 0001
>  3 Feb 09:35:27 ntpd[369]: ntpd exiting on signal 15
>  3 Feb 09:38:39 ntpd[403]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 09:38:40 ntpd[403]: time reset +0.481615 s
>  3 Feb 09:38:40 ntpd[403]: kernel time sync disabled 0041
>  3 Feb 09:41:55 ntpd[403]: ntpd exiting on signal 15
>  3 Feb 09:45:09 ntpd[412]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 09:45:09 ntpd[412]: time reset -0.871878 s
>  3 Feb 09:49:24 ntpd[412]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 10:00:14 ntpd[412]: time reset +0.128403 s
>  3 Feb 10:00:14 ntpd[412]: kernel time sync enabled 0001
>  3 Feb 10:04:33 ntpd[412]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 10:36:45 ntpd[412]: time reset -0.873229 s
>  3 Feb 10:36:45 ntpd[412]: frequency error -512 PPM exceeds tolerance  
> 500 PPM
>  3 Feb 10:41:02 ntpd[412]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 11:47:44 ntpd[412]: no servers reachable
>  3 Feb 11:48:49 ntpd[412]: synchronized to GPS_BANC(0), stratum=0
>  3 Feb 12:03:51 ntpd[412]: time reset -0.871702 s
> </ntp.log>
>
> Around 8:00 AM I deleted the ntp.drift file...thus the improvement.
>
> <server ntpd output>
> clock_update: at 19299 assoc 1
> local_clock: assocID 49204 offset -0.872882728 jitter 0.924624933 
> state  4
> auth_agekeys: at 19320 keys 1 expired 0
> refclock_transmit: at 19362 127.127.16.0
> pp:  34 22:04:38.000934 0
> refclock_receive: at 19362 127.127.16.0

<snip>

The server output doesn't tell me very much!

NTP "1.4.2" is either a typo or a version fifteen or sixteen years 
old!!!!!  Did you mean 4.2.1?

Have you checked the SYNCCLOCK32 for proper operation?  It's not a clock 
I'm familiar with.

If you try to synchronize to a public network server, do you still get 
the error?   If you  can synch to a network server but not to your 
refclock, I suspect a hardware problem with the refclock.



More information about the questions mailing list