Well, I spoke too soon, I think. Machine 2 stayed within 100 ms only
for a little more than half an hour. It strayed as much as -1.6 seconds
before correcting back to +0.0018 sec. Then it did this
minor-drift-and-correct dance a couple times, and now it's strayed to
over +9 seconds in about 80 minutes. Hey, it just corrected back to
-0.13 sec.
After the initial success with machine 2, I ran the same magic commands
on machine 1, and it's stayed within 1 second for over 5 hours now.
Any more suggestions? Or should I give it more time before passing
judgment?
David
David Rush wrote:
Wow. The box I tried this on has now been within 100 ms (per
stripchart) for over an hour and a half now. It took nearly four
hours for it to "lock in", but things look good now. I'll keep
watching, but this is very promising.
I noticed in the registry that NtpClient\SpecialPollInterval is still
86400.
Will the results of the magic commands be persistent, and survive a
reboot?
David
Ry wrote:
Setting special poll interval to 1 day seems far too long to me. Try
this:
w32tm /config /update /manualpeerlist:us.pool.ntp.org,0x8
/syncfromflags:MANUAL
then run
w32tm /resync /rediscover
The ",0x8" after the servername tells w32tm to use client mode,
automatically determining the polling interval itself (something like
"real" NTP). The resysnc command tells it to throw out all acumulated
error estimates and start from scratch.
_______________________________________________
questions mailing list
[email protected]
https://lists.ntp.isc.org/mailman/listinfo/questions
_______________________________________________
questions mailing list
[email protected]
https://lists.ntp.isc.org/mailman/listinfo/questions
_______________________________________________
questions mailing list
[email protected]
https://lists.ntp.isc.org/mailman/listinfo/questions
