[ntp:questions] Controlling the combine algorithm

unruh unruh at wormhole.physics.ubc.ca
Tue Jul 5 13:29:43 UTC 2011


On 2011-07-04, David Woolley <david at ex.djwhome.demon.invalid> wrote:
> Bottom quoting out of character:
>
> I believe that this refers to the following article, in which the writer 
> wants the "best" (by which he means something like lowest offset) 6 out 
> of 15 servers to contribute to the final time correction.

If he knew they had the lowest offset, why not use that knowledge to set
you clock? How in the world would anyone know they had the lowest
offset? They could all uniformly be off by 10 days, because they all use
some rogue clock as their reference. They would be consistant, but off
nevertheless. How would you know that without refering to some better
clock?

>
> I may have been responsible for wrongly suggesting maxclock, but not 
> having it documented doesn't help in that respect.
>
> > Subject: Selection and Clustering of candidates
>> Mischanko, Edward T wrote:
>>> I choose to poll 15 servers on my network for time synchronization.  NTP
>>> is selecting over 10 servers and clustering them.  Some of the servers
>>> NTP selects have offsets that are higher than the majority; in other
>>> words, they do not have the correct time.  Why are they being selected?
>>> Why are so many being selected when I really want the best 5 or 6.
>>> 
>>> # Miscellaneous stuff
>>> tinker allan 7
>>> tos minclock 4 minsane 3
>>> #
>>> driftfile "C:\Program Files\NTP\etc\ntp.drift"
>>> #
>>> enable stats
>>> statsdir "C:\Program Files\NTP\etc\"     # directory for statistics
>>> files
>>> statistics loopstats
>>> filegen loopstats file loopstats type day link enable
>>> #
>>> #    Need to set up time sources...
>>> #
>>> #
>>> server 000.000.102.1 minpoll 4 maxpoll 10
>>> server 000.000.106.1 minpoll 4 maxpoll 10
>>> server 000.000.109.1 minpoll 4 maxpoll 10
>>> server 000.000.110.1 minpoll 4 maxpoll 10
>>> server 000.000.112.1 minpoll 4 maxpoll 10
>>> server 000.000.120.1 minpoll 4 maxpoll 10
>>> server 000.000.121.1 minpoll 4 maxpoll 10
>>> server 000.000.122.1 minpoll 4 maxpoll 10
>>> server 000.000.123.1 minpoll 4 maxpoll 10
>>> server 000.000.124.1 minpoll 4 maxpoll 10
>>> server 000.000.125.1 minpoll 4 maxpoll 10
>>> server 000.000.126.1 minpoll 4 maxpoll 10
>>> server 000.000.129.1 minpoll 4 maxpoll 10
>>> server 000.000.131.1 minpoll 4 maxpoll 10
>>> server 000.000.134.1 minpoll 4 maxpoll 10
>
>
>
>
> David L. Mills wrote:
>> Steve,
>> 
>> There is something wrong with the nomenclature in your message. By 
>> definition, the number of survivors is the number remaining after the 
>> cluster algorithm has completedd. This number is three by default, but 
>> can be changed using the minclock option. To reduce the numb er of 
>> survivors below three is generally a bad idea, but as I said, the option 
>> is available.
>> 
>> I don't understand what you mean by "it doesn't work". The maxclock 
>> option does not do what you want. It is intended for automatic 
>> configuration, where it specifies the maximum number of servers 
>> remaining after the preempt phase. For configured associations, this 
>> number is irrelevant.
>> 
>> Dave
>> 
>> Steve Kostecke wrote:
>> 
>>> On 2011-07-02, David L. Mills <mills at udel.edu> wrote:
>>>
>>>  
>>>
>>>> The combine algorithm operates on the survivors of the cluster 
>>>> algorithm, as described on the "How NTP Works" page. The number of 
>>>> survivors can be set using the minclock option. I'm not sure why you 
>>>> want to do this, but the option is there.
>>>>   
>>>
>>> Dave,
>>>
>>> I've been having a discussion with someone who wishes to configure a
>>> large number of unicast time servers and restrict the number of
>>> survivors to a number below the default. He was attempting to accomplish
>>> this using maxclock option, but it did not work.
>>>
>>> Thanks,
>>>
>>>  
>>>




More information about the questions mailing list