[c-nsp] NTP synchronization problems C2801

[Rodney Dunn]

Gotcha.

We may have to wait for M3 rebuild and try that one.

Rodney



On 6/30/10 6:10 PM, Peter Rathlev wrote:
> On Wed, 2010-06-30 at 17:43 -0400, Rodney Dunn wrote:
>> Can you try the iburst option to see if it changes it?
>
> I tried adding "iburst", and though it behaves as expected by sending 8
> rapid (~4 secs apart it seemed) after initialization I still end up
> relatively unsynchronized:
>
>   10.83.247.20 configured, authenticated, our_master, sane, valid, stratum 2
>   ref ID 10.85.8.130   , time CFD638BF.C765F6DB (23:42:23.778 CEST Wed Jun 30 2010)
>   our mode client, peer mode server, our poll intvl 64, peer poll intvl 64
>   root delay 1.70 msec, root disp 25.60, reach 1, sync dist 30.50
>   delay 0.92 msec, offset -14.8808 msec, dispersion 0.15
>   precision 2**20, version 4
>   org time CFD63B35.5F8F925C (23:52:53.373 CEST Wed Jun 30 2010)
>   rec time CFD63B35.637A41FB (23:52:53.388 CEST Wed Jun 30 2010)
>   xmt time CFD63B35.633885DA (23:52:53.387 CEST Wed Jun 30 2010)
>   filtdelay =     0.98    0.95    1.58    1.41    0.99    0.93    0.92    0.94
>   filtoffset =  -14.80  -14.81  -14.96  -15.05  -14.87  -14.88  -14.88  -14.88
>   filterror =     0.00    0.03    0.06    0.09    0.12    0.15    0.18    0.21
>   minpoll = 6, maxpoll = 10
>
> That's from shortly after the initial burst ended. Status:
>
>   Clock is synchronized, stratum 3, reference is 10.83.247.20
>   nominal freq is 250.0000 Hz, actual freq is 250.0256 Hz, precision is 2**24
>   reference time is CFD63B2B.63AE2585 (23:52:43.389 CEST Wed Jun 30 2010)
>   clock offset is -14.8808 msec, root delay is 2.63 msec
>   root dispersion is 978.06 msec, peer dispersion is 0.15 msec
>   loopfilter state is 'CTRL' (Normal Controlled Loop), drift is -0.000102725 s/s
>   system poll interval is 64, last update was 13 sec ago.
>
>> I don't see how going back to 24T would change it as it has v4 there
>> too...unless there was something later that made it even worse.
>
> I tried to "harass" the 24T device as much as I could: delete NTP
> config, reset drift file ("ntp clear drift"), set the clock to somewhere
> in 1993 and reinsert config. It recovers (~10 usec offset) within
> several minutes.
>
> I'd like it to run 15.0M, so the downgrade is just to see if it makes
> any difference.
>