[ntp:questions] Need to verify the offset

Unruh unruh-spam at physics.ubc.ca
Thu Dec 11 06:51:00 UTC 2008


tsnyder at rim.com (Todd Snyder) writes:

>I am currently looking to do something similar.  My requirements are
>longer term, and much more accurate (we're aiming for 5ms across a large
>number of servers).

>I did some poking around, and I found this perl script and have been
>fiddling with it.  I wish I was more of a time nerd (no offense meant)
>and could understand all the output, but I think what you (and I) want
>is here somewhere:

> http://www.kloth.net/software/sntp.php

>In particular, it calculates how far off the polled server is from the
>source you're querying from.  If you were/I were to query from my master
>ntp clock to all servers, on a frequent basis, we'd be able to see
>changes and report on them fairly easily.

Look. ntp is designed to read from the variety of servers, try to figure
out which one(s) are best and adjusting the local clock so as to minimize
the difference between the local clock and the time of those servers ( also
trying to figure out whether some of those servers give bad time, as
measured by the others). You want to query the servers, find the offset
times, figure out whichones are good clocks to compare to -- ie you want to
duplicate what ntp already does. 



>I am trying to find something a little more robust and enterprise
>quality, but this will do until I can do that.  Someone out there must
>have written something enterprise class to monitor time across a
>network, I just need to find them!

Yes. It is called ntp. You can find it at www.ntp.org. Or you can use
chrony (chrony.sunsite.dk) which uses a different algorithm ( much faster
at convergence, and in my tests about a factor of 2-3 better than ntp at
disciplining the clock, primarily due to its faster response to thermal
fluctuations.)



>Cheers,

>Todd. 

>-----Original Message-----
>From: questions-bounces+tsnyder=rim.com at lists.ntp.org
>[mailto:questions-bounces+tsnyder=rim.com at lists.ntp.org] On Behalf Of
>Hushpuppy
>Sent: Tuesday, December 09, 2008 11:30 AM
>To: questions at lists.ntp.org
>Subject: Need to verify the offset

>Hi,
>I have a machine that is syncing to a Stratum-2 server which is syncing
>to a Stratum-1 server, which has a GPS Stratum-0 device attached to it.
>At my work (financial services) the market rules mandate synchronizing
>our clocks to within a maximum 3 second difference of the NIST atomic
>clock.  We also need to log/document that we are within 3 seconds.

>So I have been perusing the output of ntpq but I don't fully understand
>it after some time of study.  I assume I could use the pe command to
>figure out what my current offset is, and then as to determine my assID,
>and finally rv <assID> to look at my peers.  Then I could perhaps add
>the current offset of my current peer, to one of the numbers from the rv
>command (rootdispersion?), to arrive at a maximum offset from my
>Stratum-1 server, and thus from NIST time (+- 100 microseconds  or
>whatever the guarantee is for the Stratum-0 device).  When I say
>"guarantee" I really think we're talking about best effort.  BTW,
>RFC1305 says that rootdispersion is in seconds but I think it's really
>in milliseconds.

>Can anyone help?  Given the below results from ntpq, what numbers would
>I need in order to determine my maximum deviation from NIST time at this
>current point in time?  Can I even determine that from my client machine
>here, or do I need to login to all servers: Client, plus Stratum-2, plus
>Stratum-1, to arrive at a legitimate value?
>Thanks.

>ntpq> pe
>     remote           refid      st t when poll reach   delay
>offset  jitter
>========================================================================
>======
>*XX.XX.XX.21     XX.XX.XX..1       2 u  236 1024  377    0.390
>-0.060   0.017
>+XX.XX.XX..22     XX.XX.XX.1       2 u  980 1024  377    0.333
>1.797   0.081
> 127.127.1.0     .LOCL.          10 l   39   64  377    0.000
>0.000   0.001
>ntpq> as

>ind assID status  conf reach auth condition  last_event cnt
>===========================================================
>  1 11015  9614   yes   yes  none  sys.peer   reachable  1
>  2 11016  9414   yes   yes  none  candidat   reachable  1
>  3 11017  9014   yes   yes  none    reject   reachable  1
>ntpq> rv 11015
>assID=11015 status=9614 reach, conf, sel_sys.peer, 1 event, event_reach,
>srcadr=XX.XX.XX.21, srcport=123, dstadr=YY.YY.YY.16, dstport=123,
>leap=00, stratum=2, precision=-18, rootdelay=106.873,
>rootdispersion=7.431, refid=10.30.25.1, reach=377, unreach=0, hmode=3,
>pmode=4, hpoll=10, ppoll=10, flash=00 ok, keyid=0, ttl=0, offset=-0.060,
>delay=0.390, dispersion=18.684, jitter=0.017, reftime=cce8d703.3ffd812a
>Tue, Dec  9 2008  5:37:39.249, org=cce8fc6c.762843fb  Tue, Dec  9 2008
>8:17:16.461, rec=cce8fc6c.763b45df  Tue, Dec  9 2008  8:17:16.461,
>xmt=cce8fc6c.761d72f9  Tue, Dec  9 2008  8:17:16.461,
>filtdelay=     0.43    0.39    0.50    0.36    0.41    0.46    0.42
>0.45,
>filtoffset=   -0.08   -0.06   -0.10   -0.23   -0.31   -0.41   -0.46
>-0.58,
>filtdisp=      0.00   15.38   30.77   46.13   61.50   76.85   92.21
>107.60

>_______________________________________________
>questions mailing list
>questions at lists.ntp.org
>https://lists.ntp.org/mailman/listinfo/questions

>---------------------------------------------------------------------
>This transmission (including any attachments) may contain confidential information, privileged material (including material protected by the solicitor-client or other applicable privileges), or constitute non-public information. Any use of this information by anyone other than the intended recipient is prohibited. If you have received this transmission in error, please immediately reply to the sender and delete this information from your system. Use, dissemination, distribution, or reproduction of this transmission by unintended recipients is not authorized and may be unlawful.




More information about the questions mailing list