igor rabota
|
NTPD fails to sync drifting clock.
|
igor rabota
04/24/2013 10:35 AM
post100850
|
NTPD fails to sync drifting clock.
Hello,
I got something really strange with QNX 6.4.1 ntpd. The clocck itself have a big drift of approx 3 ms per second. So,
after launching ntpd and waiting a bit, instead of synching to server, it syncs with localhost and the clock continues
to drift. Server is reliable, all other clients (win, linux) are sync-ed to it with no problems.
ntp.conf:
server 10.9.2.80 iburst prefer
server 127.127.1.0
fudge 127.127.1.0 stratum 6
launching ntpd cmdline:
ntpd -d -l /tmp/ntp.log >/tmp/debugntp.log&
after couple seconds the clock jumps to 0 offset, and then steadily drifts. Here is output of windows w32tm stripchart
(first line error because daemon was not started yet):
17:18:54 error: 0x80072746
17:18:56 d:+00.0017367s o:-00.7920553s [ * | ]
17:18:58 d:+00.0039651s o:-00.7985764s [ * | ]
17:19:00 d:+00.0040767s o:-00.8047629s [ * | ]
17:19:02 d:+00.0070356s o:-00.8139787s [ * | ]
17:19:04 d:+00.0060316s o:-00.0005456s [ * ]
17:19:06 d:+00.0015506s o:-00.0087431s [ * ]
17:19:08 d:+00.0026439s o:-00.0146072s [ * ]
17:19:10 d:+00.0037182s o:-00.0216746s [ * ]
17:19:12 d:+00.0019183s o:-00.0291349s [ * ]
17:19:14 d:+00.0028667s o:-00.0359352s [ * ]
17:19:16 d:+00.0020591s o:-00.0419477s [ * ]
ntpq -p output (after a while, 5-10 min):
/ # ntpq -p
remote refid st t when poll reach delay offset jitter
==========================================
10.9.2.80 10.10.2.100 4 u 5 64 377 1.828 1948.17 977.776
*LOCAL(0) .LOCL. 6 l 8 64 377 0.000 0.000 0.977
Can you please advise what is wrong or how to make this configuration behave?
Attached are logfiles as defined in commandline above.
|
|
|