[ntp:questions] Orphan Mode Configuration Questions (Was: Re: Synchronize clocks without a reference clock)

Per Hedeland per at hedeland.org
Fri Mar 23 21:23:02 UTC 2007


In article <slrnf06dv2.20v.kostecke at stasis.kostecke.net> Steve Kostecke
<kostecke at ntp.isc.org> writes:
>On 2007-03-22, Per Hedeland <per at hedeland.org> wrote:
>
>> Steve Kostecke wrote:
>>
>>>Here's the configuration you need:
>>>
>>>> [1.] participants are identically configured both as broadcast
>>>> client and as broadcast server.
>>
>> So the part that says "operates in subnet configurations in all modes"
>>is wrong? Do you know this for a fact and/or have tried it?
>
>There have been quite a few comments, assertions, and outright
>speculation, about orphan mode in this newsgroup from individuals who
>readily admit to _never_ having used it.
>
>I _have_ tested orphan mode with in both broadcast and multicast modes.
>See the "Orphan Mode Configuration" thread for a minimal working
>ntp.conf file.
>
>The ntpd at my posting address (check the headers) is a multicast orphan
>node. Go ahead and query w/ ntpq/ntpdc if you wish.

Hm, I'm afraid my post came across the wrong way. Let me say that I
really enjoy your posts which I find to be consistently informed,
informative, and to-the-point wrt to the questions asked / problems
reported - hence what I really wanted to say above was: If you (by
whatever means) know for a fact that orphan mode requires broadcast,
I'll take your word for it - but if not, and since you appear to have a
test setup, maybe you could try... (without wanting to "demand" that:-).

>> I would imagine at least that if all the participating servers have
>> each other configured as "peer", there shouldn't "technically" be a
>> need for a broadcast setup.
>
>It is entirely possible that orphan mode is compatible with "meshed
>peers". Why don't _YOU_ test that mode for us?

Well that question seems to be moot per your other post - thanks a lot
for that! And I expect that it will be useful for the OP too, since his
setup wouldn't allow broadcast to work.

The problem then is really that the "offical documentation" (hm, seems
it always takes me 5 minutes of random clicking to find what I'm looking
for there - an ntp.conf(5) page would be soo much simpler - but anyway
it's http://www.eecis.udel.edu/~mills/ntp/html/manyopt.html) appears to
be self-contradictory.

If someone reading (I know that you don't maintain those pages) would
fix the wording under "orphan _stratum_" to make it clear that the
broadcast config described is just an example, I think it would be a
great improvement.

Oh, and I hope you put your orphan findings on the wiki somewhere -
thanks again!

--Per Hedeland
per at hedeland.org




More information about the questions mailing list