[ntp:questions] NTP on Windows Server 2008 R2

Terje Mathisen "terje.mathisen at tmsw.no" at ntp.org
Wed Feb 8 17:45:54 UTC 2012


Geir Guldstein wrote:
> I am currently working with NTP 4.2.6p5 on Windows Server 2008 R2 (64
> bit). NTP works reasonably well before I install a certain device
> driver, which of course must coexist with NTP in the final
> configuration.
>
> I observe the following on a fresh OS installation before the
> mentioned device driver has been installed:
> After ntpd has been running for a few hours ntpq reports offsets<
> 2ms.
> When ntpd starts it reports the following to the OS event log:
> ntpd 4.2.6p5 at 1.2349-o Jan 19 9:35:18 (UTC+01:00) 2012  (3)
> Raised to realtime priority class
> Performance counter frequency 3.021 MHz
> Clock interrupt period 15.600 msec (startup slew -0.1 usec/period)
> Windows clock precision 15.600 msec, min. slew 6.410 ppm/s
> HZ 64.102 using 43 msec timer 23.256 Hz 64 deep
> proto: precision = 0.300 usec
> Listen and drop on 0 v4wildcard 0.0.0.0 UDP 123
> Listen and drop on 1 v6wildcard :: UDP 123
> Listen normally on 2 WAN fe80::795c:5b82:ec33:dd9 UDP 123
> Listen normally on 3 WAN 192.168.1.52 UDP 123
> Listen normally on 4 Loopback Pseudo-Interface 1 ::1 UDP 123
> Listen normally on 5 Loopback Pseudo-Interface 1 127.0.0.1 UDP 123
> peers refreshed
>
> I stop NTP, delete the ntp.drift file, install the device driver,
> restart the system and get the following:
> Offsets reported by ntpq large are variable, they can be as large as
> 100-200ms and ntpd steps the clock. The system behaves this way even
> when it has been running for several hours.
> When ntpd starts it reports the following to the OS event log:
> ntpd 4.2.6p5 at 1.2349-o Jan 19 9:35:18 (UTC+01:00) 2012  (3)
> Raised to realtime priority class
> Performance counter frequency 3.021 MHz
> Clock interrupt period 15.600 msec (startup slew -0.2 usec/period)
> Windows clock precision 0.500 msec, min. slew 6.410 ppm/s
> using Windows clock directly
> proto: precision = 500.000 usec
> Listen and drop on 0 v4wildcard 0.0.0.0 UDP 123
> Listen and drop on 1 v6wildcard :: UDP 123
> Listen normally on 2 WAN fe80::795c:5b82:ec33:dd9 UDP 123
> Listen normally on 3 WAN 192.168.1.52 UDP 123
> Listen normally on 4 Loopback Pseudo-Interface 1 ::1 UDP 123
> Listen normally on 5 Loopback Pseudo-Interface 1 127.0.0.1 UDP 123
> peers refreshed
>
> If I uninstall the device driver and restart the system NTP again
> works reasonably well.
>
> I will of course contact the device/driver vendor for support, but
> wonder whether anyone here has advice to share.

That sounds a lot like a disk or network driver which messes up 
severely, running far too long inside either a real-time thread or (even 
worse) in the interrupt handler.

What kind of driver is this?

SW RAID?

Terje
-- 
- <Terje.Mathisen at tmsw.no>
"almost all programming can be viewed as an exercise in caching"



More information about the questions mailing list