[ntp:questions] NTP 4.2.6p3 multicast server not visible in orphan peer to peer mode?

Wu, Bailey Bailey.Wu at thalesgroup.com.au
Thu Apr 7 02:59:41 UTC 2011

Thanks for the explanation Steve.

To answer some of your questions:

I didn't use manycastserver as I didn't think it would work with the 3.5f HPUX clients. 
Once I get access to those clients I'll be able to verify.

As to the 'true' option; this is to mark the server as a truechimer. 
Normally I would not use this, but there have been tests with our setup where
the clients marked both servers as falsetickers and refuse to synchronise with any.
I thought it would be a good option to use to stop clients from ever marking the servers
as falsetickers.

The maxpoll is just to stop ntp from increasing the poll interval. Just so I could see
the poll happening without waiting for too long.

I find that with the 'prefer' option server2 will sync with server1 faster. Without it
server2 would just stay with it's local reference clock and mark server1 as a candidate.
Eventually it would probably sync with server1.

I find that with orphan mode server2 operates without being sync'd to anything (server1 being down),
but the clients cannot ntpdate to a server that is not synchronised to anything.
clients may well start with times wildly different to that of the servers, hence when started
a script will ntpdate to any reachable servers.  

I initially used the reference setup in the following tutorial for server1 and 2

but it does not work when both servers are connected, since the refid will be the same
and server2 will reject server1 as a loop. So I changed from in server2
to, which works nicely.

Also, the orphan line does not seem to make a difference whether its there or not.

I am using lower stratum numbers because the network is isolated, and will never operate
with any outside networks or with any other time source (for the life of the network).
When I said isolated I really mean it :P

-----Original Message-----
From: questions-bounces+bailey.wu=thalesgroup.com.au at lists.ntp.org [mailto:questions-bounces+bailey.wu=thalesgroup.com.au at lists.ntp.org] On Behalf Of Steve Kostecke
Sent: Thursday, 7 April 2011 11:11 AM
To: questions at lists.ntp.org
Subject: Re: [ntp:questions] NTP 4.2.6p3 multicast server not visible in orphan peer to peer mode?

On 2011-04-06, Wu, Bailey <Bailey.Wu at thalesgroup.com.au> wrote:

> Steve, I like your second configuration. I applied it yesterday with
> the following modifications:
> # /etc/ntp.conf for server 1
> driftfile /var/ntp/ntp.drift
> broadcastserver
> server
> fudge stratum 4
> # /etc/ntp.conf for server 2
> driftfile /var/ntp/ntp.drift
> broadcastserver
> server
> fudge stratum 6
> server server1 true maxpoll 6 prefer

You've re-implemented your original broken configuration and thrown away
everything I gave you.

If you want to use manycast associations you need to use
'manycastclient' on the manycast clients and 'manycastserver' on the
manycast servers.

BTW What is 'true' on your 'server server1' line ? Why do you think you
need 'maxpoll 6'? Why prefer?

> It seems to do what I want. But fail over to server2 is very slow,
> takes about 10min for server2 to sync with it's own clock

This is normal. A time source must become unreachable before ntpd will
deselect it. This can take 8 * maxpoll.

> and all clients to then sync to server2. I had to use for
> server2 otherwise it will refuse to sync with server1 because it fails
> the peer loop test.

The configuration I gave you set up the following:

1. Server 1 is the LAN master (using the Undisciplined Local Clock at
stratum 10)

2. Server 1 is a manycastserver for both the clients _and_ server 2.

3. Server 2 can degrade to orphan mode at stratum 12 if server 1 becomes
unreachable (i.e. goes down)

4. Server 2 is a manycastclient (this could be changed to a regular
unicast client) and is also a manycast server for 

> I'm curious about the "tos orphan 12" line you have for server2,
> what's the effect of this?

It is to allow server2 to become the LAN time server (at stratum 12) in
the event that server 1 goes down.

The reason for operating server 1 at stratum 10 and server 2 at stratum
12 is to prevent problems if your network does gain an external
connection and real time sources become reachable.

> would this make server2 realise server1 is down and then switch to its
> local clock faster?


> Does this mean all clients that are connected need to have this orphan
> mode line as well?

No. Only the machines which will be orphan mode servers need the tos
orphan line.

Steve Kostecke <kostecke at ntp.org>
NTP Public Services Project - http://support.ntp.org/

questions mailing list
questions at lists.ntp.org

This e-mail transmission and any documents, files and previous e-mail messages
attached to it are private and confidential. They may contain proprietary or copyright
material or information that is subject to legal professional privilege. They are for
the use of the intended recipient only.  Any unauthorised viewing, use, disclosure,
copying, alteration, storage or distribution of, or reliance on, this message is
strictly prohibited. No part may be reproduced, adapted or transmitted without the
written permission of the owner. If you have received this transmission in error, or
are not an authorised recipient, please immediately notify the sender by return email,
delete this message and all copies from your e-mail system, and destroy any printed
copies. Receipt by anyone other than the intended recipient should not be deemed a
waiver of any privilege or protection. Thales Australia does not warrant or represent
that this e-mail or any documents, files and previous e-mail messages attached are
error or virus free.

More information about the questions mailing list