On 11/13/2010 6:56 PM, Edward T. Mischanko wrote:
What your saying makes perfect sense, but it doesn't seem to work that way
on this Windows XP machine. I would expect it to write to the loopstats
with every poll at a minimum and it doesn't. I am curious to know what
options are frustrating NTPD? Could it be the number of servers I am
polling? I have 10 configured; what is optimal?
"David Woolley"<[email protected]> wrote in message
news:[email protected]...
Edward T. Mischanko wrote:
NTPD fails to update a loopstats file on a regular schedule. This brings
the
question is my clock being updated on a regular schedule?
Depending on your OS, and whether or not you have chosen options that
frustrate the kernel discipline, it will be being updated every clock
tick. It will certainly be being updated every 4 seconds, even if you
have a very old version of ntpd. I think current versions update every
second, in user space mode.
ntpd solves for both rate and offset, and typically only needs to make
measurements at 20 minutes or more intervals in order to be able to
interpolate the time to the best quality possible for the sources and the
machine. chrony also does this, although some believe chrony handles
transients better.
PLEASE DO NOT TOP POST!
Ten servers configured seems a little extreme. Four, five, and seven
servers protect you from the failure of one, two, or three servers
respectively, where failure can mean either not responding, or
responding with the wrong time.
_______________________________________________
questions mailing list
[email protected]
http://lists.ntp.org/listinfo/questions
