1024Hz tick on Intel - clock drift? (NTP complaints)

Barrie Spence (baz@sqf.hp.com)
Thu, 02 Dec 1999 19:16:37 +0000

We're running 2.2.13 kernels on PII/PIII machines with HZ changed to
1024
(to allow sufficient resolution for measurement purposes), but it would
appear that the clock appears to drift too much for the NTP daemons to
correct (I presume using adjtime()).

This is a RedHat 6.1 base (xntp3-5.93-13 RPM).

>From syslog:

Nov 29 15:23:25 hpqs0580 ntpdate[522]: step time server 15.144.184.32
offset 0.833641 sec
Nov 29 15:23:25 hpqs0580 xntpd: xntpd startup succeeded
Nov 29 15:23:25 hpqs0580 xntpd[532]: xntpd 3-5.93e Mon Aug 16 20:55:26
EDT 1999 (1)
Nov 29 15:23:25 hpqs0580 xntpd[532]: tickadj = 5, tick = 977,
tvu_maxslew = 5110, est. hz = 1023
Nov 29 15:23:25 hpqs0580 xntpd[532]: precision = 7 usec
Nov 29 15:23:25 hpqs0580 xntpd[532]: read drift of -333.622 from
/etc/ntp/drift
Nov 29 15:29:16 hpqs0580 xntpd[532]: synchronized to 15.144.184.33,
stratum=3
Nov 29 15:29:17 hpqs0580 xntpd[532]: time reset (step) 0.679307 s
Nov 29 15:29:17 hpqs0580 xntpd[532]: synchronisation lost
Nov 29 15:35:13 hpqs0580 xntpd[532]: synchronized to 15.144.185.1,
stratum=3
Nov 29 15:35:14 hpqs0580 xntpd[532]: time reset (step) 0.747457 s
Nov 29 15:35:14 hpqs0580 xntpd[532]: synchronisation lost
Nov 29 15:41:30 hpqs0580 xntpd[532]: synchronized to 15.148.227.5,
stratum=4
Nov 29 15:41:31 hpqs0580 xntpd[532]: time reset (step) 0.762709 s
Nov 29 15:41:31 hpqs0580 xntpd[532]: synchronisation lost
Nov 29 15:47:24 hpqs0580 xntpd[532]: synchronized to 15.144.184.33,
stratum=3
Nov 29 15:47:25 hpqs0580 xntpd[532]: time reset (step) 0.689713 s
Nov 29 15:47:25 hpqs0580 xntpd[532]: synchronisation lost
Nov 29 15:53:21 hpqs0580 xntpd[532]: synchronized to 15.144.185.1,
stratum=3
Nov 29 15:53:22 hpqs0580 xntpd[532]: time reset (step) 0.745505 s
Nov 29 15:53:22 hpqs0580 xntpd[532]: synchronisation lost
Nov 29 15:59:38 hpqs0580 xntpd[532]: synchronized to 15.148.227.5,
stratum=4

I'm going to try the latest ntp-4.x release to see if it helps any...

Any guidance/insight would be appreciated.

Thanks,
Barrie

-- 
Barrie Spence (313-2465)                Agilent Technologies UK Ltd
E-Mail:   baz@sqf.hp.com                South Queensferry, UK
Play:     barrie@calvin.demon.co.uk	#include <std/disclaimer.h>


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/