[ntp:questions] GPS driver failure mode question
Charles Brown
charles.brown at sensis.com
Mon Jan 12 22:57:55 UTC 2009
I linked from the LCL driver to the Orphan Mode documentation, and
this seems to be the answer to my original (poor) question.
Three questions on orphan mode;
First, the orphan mode page says;
"Each orphan child chooses the orphan parent as the root server with the
smallest value."
Which value is that? Stratum?
Second question, using the diagram from the orphan mode section;
Suppose all sources go away, orphan mode kicks in and all the children
and Server B select Server A as the orphan parent. Subsequently, Server
A dies. Will Server B take on the parent role, and all children respond
appropriately?
Third question; Suppose A gets well, still without its external
sources, and comes back. Will Server B defer being parent back to server
A, so Server A becomes parent again, and Server B and all the children
respond appropriately?
--CB
Terje Mathisen wrote:
> Unruh wrote:
> > Steve Kostecke <kostecke at ntp.org> writes:
> >> Never.
> >
> >> The Undisciplined Clock Driver does poll the system clock
> >> (often misleadingly referred to as the "local clock") _but_ the poll
> >> results are discarded. So ntpd never actually serves time _from_ the
> >> Undisciplined Local Clock.
> >
> >
> > Sorry there is some sematic distinction going on here that I do not
> follow.
> > ntp always "serves time" from the system clock. The question is how and
> > whether ntp thinks that system clock itself is being disciplined. If
> it is
> > disciplined by nothing it does not act as a server, and says so. If it is
>
> It can still do so, but definitely not by default!
>
> > "disciplined " by local clock, it is still disciplined by nothing,
> but can
> > act as an ntp server, and tells outside clients that it is being
> > disciplined by the local clock. Am I wrong in my understanding?
>
> Slightly, see the documentation page for the local clock driver:
>
> http://www.eecis.udel.edu/~mills/ntp/html/drivers/driver1.html
>
> Terje
>
> --
> - <Terje.Mathisen at hda.hydro.com>
> "almost all programming can be viewed as an exercise in caching"
>
> _______________________________________________
> questions mailing list
> questions at lists.ntp.org
> https://lists.ntp.org/mailman/listinfo/questions
>
More information about the questions
mailing list