At 01:52 AM 2/18/2011, Chris Albertson wrote...
I stopped ntpd then ran ppstest. It produces the following:
source 0 - assert 1298011155.999464502, sequence: 64785 - clear
1297840243.197985597, sequence: 36
I don't know what all of this means but I see the time increment by
about 1 second every second
It's reporting Unix timestamps for each PPS.
locke:/# echo 1297840243.197985597 | awk '{print strftime("%c",$1)}';
Wed 16 Feb 2011 02:10:43 AM EST
Looks like PPS is coming into the box just fine, but it looks like your
clock is off by a couple of days?
You can try adding the Atom driver for PPS, and see how that behaves.
It can coexist with the Oncore driver...
# PPS -
# time1 - offset calibration
# flag2 - PPS edge 0=Assert, 1=Clear
# flag3 - kernel discipline, 0=disable (default), 1=enable
# flag4 - record timestamps
# stratum
# refid
server 127.127.22.0 minpoll 4 maxpoll 4
fudge 127.127.22.0 flag2 0
Also, you could try switching to PPS clear by adding the following line
to /etc/ntp.oncore.0
CLEAR
I think ASSERT is the default, and correct for you, but it's worth a
try. This has got to do with the polarity of the PPS, and which edge is
used. I seem to recall that back when I had to patch in PPS support,
both assert and clear got incremented as shown by ppstest. But now with
it in the 2.26.34+ kernel, only assert happens, and clear won't even
work? Perhaps someone with more kernel knowledge will comment.
_______________________________________________
questions mailing list
[email protected]
http://lists.ntp.org/listinfo/questions
