From: Marcus Lasarko (mlasarko@co.ba.md.us)
Date: Sat Dec 09 2006 - 01:23:31 ART
Greetings Jacque!
As long as you have your peer statements between R1 < > R2 you should be acceptable to just add the peer statements, as you do not have multiple peers, no preference concern, etc... It may also be a good idea to define the source interface as a loopback to assure you have an "up" interface between the devices, and of course reachability thereof. (You would not want to define R2's Eth1 as the source.)
Simply establish the server, and should that fail the peers, pending reachability, will go into "symmetric active" mode and work it out. BTW - It helps sometimes if the clocks are close to start, otherwise this peering agreement of what time it is may take a while :0
Of course, If they fail-over/fall-back from the same valid NTP server source then chances are they are already very close in regards to time!
On another note, unless a lab asks for it specifically I would not recommend a stratum 1 configured anywhere, at least not on a device in "Cisco-world". You might want to think about a stratum 3 or 4 and go from there.
Make sure and use the "show ntp associations" detail and "show ntp status" to keep an eye on what is happening during the peering vs. server "client mode".
Best,
~M
>>> "jacque vincent" <jacque_vincent@hotmail.com> 12/08/06 10:40 PM >>>
Hi
I have a network like the following
R1e0-------------e0-R2-e1--------------e0-R3
R1-e0--1.1.1.1/24
R2-e0-- 1.1.1.2/24
R2-e1-- 1.1.2.1/24
R3-e0-- 1.1.2.2/24
R3 is configured as NTP master 1.
Configure R1 and R2 to get time from R3
R1 and R2 should have a NTP peer associations to each other.
How can I get R1 and R2 to synchronise to each other when R3 fails.
This archive was generated by hypermail 2.1.4 : Tue Jan 02 2007 - 07:50:37 ART