Excerpts from William Herrin on Sun, Nov 30, 2008 11:05:51AM -0500: > We went through this before nearly a year ago. The short version is > that the jitter at the cache duration boundary will impact that ntp > session regardless of whether you keep or drop the packet. > > Solutions include: > > 1. TRRP style: place the NTP daemons that will communicate out over > the internet on bare (non-upgraded) addresses and sync the map-encap > machines to those nearby ntp servers. > > 2. Any pull-cache: modify the NTP software so that it ignores > out-of-bounds answers when it hasn't sent a packet in a while and > tries again a few seconds later.
The issue will rarely come up. In order for it to arise, a site border router must come up cold with no knowledge of prefixes where it wants to find its NTP servers. Once it maps these prefixes, which will very probably be few and well-used, it will have mappings for them until it comes up cold again. So as yet another alternatives, avoid completely cold restarts, or give it what it needs via configuration or synchronization. There are probably other possibilities. Scott _______________________________________________ rrg mailing list [email protected] https://www.irtf.org/mailman/listinfo/rrg
