Something is buggy with iburst...

[Achim Gratz]

Hal Murray via devel writes:
> To my eye, the 38m under when is suspicious.

It'll actually go up to 51 minutes (3072 seconds).

> I've seen similar problems.  In my case, the whole when column was crazy.

Well, I've had several servers go similarly out of whack, but not yet
all of them at the same time.

In any case, I am really asking about what I might look for to get at
the root cause.  I could just restart otherwise.

> Here is what I think is going on...
>
> The when column is showing the difference between the last time a response 
> arrived and "now".  I think it is using ntpq's now rather than the server's 
> now.

The times in question are synchronised to way better than 1ms, typically
less than 50µs between all of the NTP chimers on my local network (as
viewed from a client-only machine) and less than about 20µs for the two
with GPS/PPS.  At the moment all four servers are within 7µs as viewed
from my PC.

> Things may get more complicated if/when the server does a time step.  But I 
> think a step resets all that stuff.

None of that happened as far as I can tell.  I know it's not the server
since all other clients see that server OK.  It's just this one
connection that's affected.  Something's changed from yesterday
interestingly enough: yesterday the reachability was only every second
bit (reach 525 or 252), but today it's a solid 377.  Ntpmon shows the
this:

assoc=52322: conf, reach, sel_candidate, 4 events, sys_peer
srcadr=192.168.178.29, srcport=123, dstadr=192.168.178.28, dstport=123
leap=no-leap    reftime=2017-05-09T16:17:10.49Z rootdelay=0μs
stratum= 1      rec=2017-05-09T16:17:20.68Z     rootdisp=1.144ms
precision=-20   xmt=2017-05-09T16:17:20.68Z     dispersion=16.709213ms
unreach=0, hmode=3, pmode=4, hpoll=10, ppoll=4, headway=7925, flash=0, keyid=0
filtdelay  =    0.53    0.63    0.55    0.60    0.57    0.51    0.63    0.60 ms
filtoffset =    0.00   -0.07    0.03    0.05    0.01   -0.05    0.03    0.03 ms
filtdisp   =    0.00   15.33   30.66   45.99   61.32   76.65   91.98  107.31 ms

Last contacted + headway gives 8192s or poll=13?  What is hpoll?  The
other raspPi that doesn't have that problem (at the moment anyway, has
this as 16s and correspondingly ppoll=hpoll=4 as configured.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Wavetables for the Waldorf Blofeld:
http://Synth.Stromeko.net/Downloads.html#BlofeldUserWavetables