[ntp:questions] Help needed with excessive drift

mlawdawg at yahoo.com mlawdawg at yahoo.com
Mon May 2 17:33:29 UTC 2005


I'm new to NTP and could use some help. I've ported NTP 4.2.0 from
ntp.org to vxWorks running on pentium(x86) and mips. My port is based
on Casey Crellin's vxWorks port.

I need help in debugging/fixing excessive drift. On my pentium hw, an
intel pc motherboard, the drift was -7.905ppm after 1 hour and
eventually gets pegged at -500ppm. On my proprietary mips platform, the
drift is -371.746ppm after several days of running.

What steps/actions can I take to debug and resolve the drift?

Below is ntpq output, taken after reading "NTP Debugging Techniques" in
the 4.2.0 html documentation. The "rv <assoc for system peer>" shows a
filter offset that is decreasing by 8-9 units per round.

Thanks in advance,
Mark Law
mlawdawg at yahoo.com


ntpq> pe
     remote           refid      st t when poll reach   delay   offset
jitter
==============================================================================
 17.254.0.28     .STEP.          16 u    - 1024    0    0.000    0.000
4000.00
*192.83.249.28   207.126.98.204   2 u  538 1024  377    7.813  110.266
 8.005
+216.52.237.152  128.9.176.30     2 u  660 1024  377    9.975  109.894
87.893
+69.13.233.192   164.67.62.194    2 u  799 1024  357   29.900   89.593
 7.813
+63.247.194.250  129.6.15.29      2 u  801 1024  377   49.986   88.628
 7.813

ntpq> as
ind assID status  conf reach auth condition  last_event cnt
===========================================================
  1   732  8000   yes   yes  none    reject
  2   733  b6f4   yes   yes  none  sys.peer   reachable 15
  3   734  b4f4   yes   yes  none  candidat   reachable 15
  4   735  b4f4   yes   yes  none  candidat   reachable 15
  5   736  b4f4   yes   yes  none  candidat   reachable 15

ntpq> rv 733
status=b6f4 reach, conf, sel_sys.peer, 15 events, event_reach,
srcadr=192.83.249.28, srcport=123, dstadr=192.168.xxx.yyy, dstport=123,
leap=00, stratum=2, precision=-20, rootdelay=2.808,
rootdispersion=43.930, refid=207.126.98.204, reach=377, unreach=0,
hmode=3, pmode=4, hpoll=10, ppoll=10, flash=00 ok, keyid=0, ttl=0,
offset=110.266, delay=7.813, dispersion=24.512, jitter=8.005,
reftime=c620cfc2.be74afd5  Mon, May  2 2005 16:11:46.743,
org=c620d1fd.85ce39b4  Mon, May  2 2005 16:21:17.522,
rec=c620d1fd.69903772  Mon, May  2 2005 16:21:17.412,
xmt=c620d1fd.69903772  Mon, May  2 2005 16:21:17.412,
filtdelay=     7.81    7.81    7.81    7.81    7.81    7.81    7.81
7.81,
filtoffset=  110.27  102.26   93.44   84.80   76.64   67.95   59.11
49.88,
filtdisp=      9.73   25.09   40.44   55.78   71.16   86.55  101.94
117.28

ntpq> rv
status=06f4 leap_none, sync_ntp, 15 events, event_peer/strat_chg,
version="4.2.0", processor="mips", system="i686-pc-cygwin", leap=00,
stratum=3, precision=-7, rootdelay=10.620, rootdispersion=232.494,
peer=733, refid=192.83.249.28,
reftime=c620d1fd.69903772  Mon, May  2 2005 16:21:17.412, poll=10,
clock=c620d43a.c5a1d872  Mon, May  2 2005 16:30:50.772, state=4,
offset=101.811, frequency=-371.746, jitter=38.437, stability=1.295




More information about the questions mailing list