On 7/8/2011 10:38 AM, Florian Heigl wrote:
Hi,I'm doing some testing of the ntp config I made and don't really understand what just happened. Primarly I'm looking at two ntp servers and one clock out of a larger setup. NTP is quite critical so I've already done some head-scratching to come to this setup. Setup on "ntp01": server meinberg1 server meinberg2 prefer server meinberg3 server meinberg4 peer ntp02 iburst # this line reads ntp01 on the other server # local fallback clock server 127.127.1.0 fudge 127.127.1.0 stratum 10 driftfile /var/lib/ntp/drift Note about the local clock: Lik Right now only Meinberg3 is online, and, to make things more interesting, I unplugged it's GPS cable. It kept at stratum 1/2 for a long duration and now went dpwn to stratum 3 local source. I would have hoped both "ntp" servers would still use meinberg3, instead they decided to pick their local clock. remote refid st t when poll reach delay offset jitter ============================================================================== meinberg1 .INIT. 16 u - 64 0 0.000 0.000 0.000 meinberg2 .INIT. 16 u - 64 0 0.000 0.000 0.000 meinberg3 LOCAL(0) 3 u 34 64 7 0.192 -1531.5 5.156 meinberg4 .INIT. 16 u - 64 0 0.000 0.000 0.000 ntp02 LOCAL(0) 11 u 24 64 17 0.216 -1529.3 7.325 *LOCAL(0) .LOCL. 10 l 30 64 17 0.000 0.000 0.001 I really do love working with NTP, just right now I don't understand why it does what it's doing (even if it may just be what I configured into it, right?
I'd suggest a little more diversity in servers. Four Meinberg servers, presumably located at the same site, could leave you with NO servers if the Meinberg site went down or if you lost connectivity to the site.
_______________________________________________ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
