[ntp:questions] Order of servers in ntp.conf

unruh unruh at invalid.ca
Thu Aug 15 18:09:38 UTC 2013


On 2013-08-15, Steve Kostecke <kostecke at ntp.org> wrote:
> On 2013-08-15, unruh <unruh at invalid.ca> wrote:
>
>> On 2013-08-14, Steve Kostecke <kostecke at ntp.org> wrote:
>>
>>> The time server specified in each of those lines is the one which is
>>> currently selected as the "sys_peer".
>>
>> But as I understand it, it is simply one of the systems which is
>> not a false ticker, and has not real significance other than that.
>> Ie, its time is not treated any differently than any of the other
>> systems regarded as true chimers by the selection algorithm. Or do I
>> misunderstand something?
>
> The sys_peer is chosen after the select[ion] alrorithm scans the
> associations for selectable candiates _and_ after the cluster
> (combine?) algorithm casts out the outliers.
>
> In the last paragraph at
> http://www.eecis.udel.edu/~mills/ntp/html/warp.html we see:
>
> "The algorithms described on the Mitigation Rules and the prefer Keyword
> page combine the survivor offsets, designate one of them as the system
> peer and produces the final offset used by the algorithm described on
> the Clock Discipline Algorithm page to adjust the system clock time and
> frequency."
>
> Section 6 of the Mitigation Rules page
> (http://www.eecis.udel.edu/~mills/ntp/html/prefer.html) 
> clarifies how the process works:
>
> "As previously noted, the cluster algorithm casts out outliers, leaving
> the survivor list for later processing. The survivor list is then sorted
> by increasing root distance and the first entry temporarily designated
> the system peer. At this point the following contributors to the system
> clock discipline may be available:
>
> * (potential) system peer, if there are survivors;
> * orphan parent, if present;
> * local driver, if present;
> * modem driver, if present;
> * prefer peer, if present;
> * PPS driver, if present.
>
> The mitigation algorithm proceeds in three steps in turn.
>
> 1. If there are no survivors, the modem driver becomes the only survivor
> if there is one. If not, the local driver becomes the only survivor if
> there is one. If not, the orphan parent becomes the only survivor if
> there is one. If the number of survivors at this point is less than the
> minsane option of the tos command, the algorithm is terminated and the
> system variables remain unchanged. Note that minsane is by default 1,
> but can be set at any value including 0.
]
This is ambiguous. If no survivors-> local/orphan. If suvivors<minsane
(which is almost always true if no survivors)-> unchanged. 
If I set minsane to 100 and have only 5 peers, what happens? Is
local/orphan used or is nothing changed.



>

> 2. If the prefer peer is among the survivors, it becomes the system peer

In the above, there are survivors, but fewer than minsane. Does this
apply? or does this apply only if survivors>minsane?

> and its offset and jitter are inherited by the corresponding system
> variables. Otherwise, the combine algorithm computes these variables
> from the survivor population.

"Otherwise the combine algorithm computes these variables from the
survivor population" does not say how it does so. 
Is it the ntp offset of the system peer? or is it the offset
from some (weughted?) average of the survivors? Since some of the
survivor's data could be 3 hours old (poll of 10 with the 1/8 selection) 
and others 1 sec old. Does this enter the weighting?



>
s> 3. If there is a PPS driver and the system clock offset at this point is
> less than 0.4 s, and if there is a prefer peer among the survivors or
> if the PPS peer is designated as a prefer peer, the PPS driver becomes
> the system peer and its offset and jitter are inherited by the system
> variables, thus overriding any variables already computed. Note that a
> PPS driver is present only if PPS signals are actually being received
> and enabled by the associated driver.
>
> If none of the above is the case, the data are disregarded and the
> system variables remain as they are."
>



More information about the questions mailing list