Hi, I'm wondering if there's something wrong with the clock variables of a "PPS clock". Looking aat those, I get: ntpq> cl assID=0 status=0000 clk_okay, last_clk_okay, device="PPS Clock Discipline", timecode=, poll=2455049, noreply=0, badformat=0, baddata=0, fudgetime1=0.000, stratum=16, refid=80.80.83.0, flags=0
(ntpd [email protected], used in a Meinberg LANTIME server) When reading the "peer variables", I get: assID=0 status=21f4 leap_none, sync_atomic/PPS, 15 events, event_peer/strat_chg, version="ntpd [email protected] Tue Oct 7 10:44:09 UTC 2008 (1)", processor="i586", system="Linux/2.6.15.1", leap=00, stratum=1, precision=-19, rootdelay=0.000, rootdispersion=0.386, peer=54170, refid=PPS, reftime=cff11bac.e68208bf Wed, Jul 21 2010 9:09:32.900, poll=4, clock=cff11bb6.b1f500fe Wed, Jul 21 2010 9:09:42.695, state=4, offset=0.000, frequency=75.869, jitter=0.002, noise=0.002, stability=0.002, tai=0, LANTIME=lantime/PZF511/M3x/V5.28g/SN999999999999, access_policy="open access; please notify [email protected] for permanent associations", admin_contact="[email protected]", info_url="http://xxxxxxxx.xxxxxxxxxxxx.de/xxxx.shtml"; (some values obfuscated) Specifically I don't like: 1) "stratum=16", while in the peer it displays: "stratum=1" 2) "timecode=" (without a value, not even double double-quotes (to ease parsing)) My suggestions: The PPS should inherit the stratum of the providing clock. As PPS has no read timecode, we could display "1" when the clock uses the "assert" edge, or "0" when the pulse uses the "clear" edge. Otherwise I'd remove the "timecode" variable from the list. (I was asking Meinberg, whether they use a modified version of ntpd. If it's not the case, I might file a bug report. But send your opinions first!) Regards, Ulrich _______________________________________________ questions mailing list [email protected] http://lists.ntp.org/listinfo/questions
