[ntp:questions] need option to ignore 'leap not in sync error'
stenn at ntp.org
Mon Jan 20 12:13:43 UTC 2014
Martin Burnicki writes:
> William Unruh wrote:
> > On 2014-01-16, Maarten Wiltink <maarten at kittensandcats.net> wrote:
> >> Configure the local clock as a source at high stratum on the master.
> >> "server 127.127.1.0 stratum 10"
> >> People will now fall all over you, and me, yelling that this is
> >> anathema, but you have explained to my satisfaction what you want,
> >> and this is how you do it.
> > I would not fall all over you. This seems to be precisely the case for
> > which the "local" directive was designed.
> > It IS inappropiate to use local as a general feature, including on pure
> > clients, as is done in some config files of some distros.
> But if the presence/usage of the "local" refclock on a particular node
> only affects potential clients of that node, why should it hurt to use
> this also on a pure "leaf" node? For the basic operation it doesn't
> matter if a node serves request from some clients, or not.
It seems to me that in general folks do not read the documentation.
When people see a local refclock in a config file, it seems like a *lot*
of them don't understand what that really means.
If a machine has a local refclock configured in it and that is because
of a default configuration instead of a conscious choice, it seems that
the number of problems *grows* in that case.
I'm thinking we want to be pretty careful about when we'd recommend a
local refclock, as it seems much better to recommend orphan mode as that
seems to be a much better solution for a wider number of folks.
What do others think?
> > There are I believe some other options with newer version of ntp (orphan
> > mode?) as well, but I do not know the details.
> Yes. However, as far as I can see there are only advantages if there is
> more than 1 time "master" which shall provide the time to a group of
> clients while no upstream ref time is available.
And there is still the issue of what it the "local master" is
sufficiently off and then a trusted time source is reached?
> On the other hand, if I remember correctly, orphan mode is more
> susceptible to mis-configuration. For example, again IIRC, an orphan
> node never synchronizes to an upstream server if "tos orphan 0" is in
I don't recall offhand, and I should have been asleep over an hour ago.
It would be good to document the circumstances where "tos orphan 0" is
useful, and if there are no such cases it should either be disallowed or
should generate a warning.
More information about the questions