[ntp:questions] NTP makes a time jump

[E-Mail Sent to this address will be added to the BlackLists]

karpekin at gmail.com wrote:
> Some news from testing. Setup:
>  - server 10.10.52.104, with timeshift 14 sec, config:
> tos maxdist 16
> tinker step 0
> server 10.10.2.101
> server 10.9.2.80
> server 127.127.1.0 iburst minpoll 4
> fudge 127.127.1.0 stratum 12
> driftfile /etc/ntp/ntp.drift
>
>  - two clients, config:
> tos maxdist 16
> tinker step 0
> server 10.10.52.104 iburst
> driftfile  /tmp/ntp.drift
>
> On the boot, server has no access to uplink servers.
>  as soon as clients are paired to server, I connect "island"
>  to main network. Server discovers it, connects to one of
>  uplink servers and starts slewing (as expected).
> However, the problem comes from my clients.
> After 2-5 minutes, they reject my "crazy slewing" server
>  with flash=1440 (pkt_dist, peer_dist, peer_unfit).
>  here is ntpq -rv of one of them after it rejected server:
> ...
> offset=158.878, delay=4.959, dispersion=15937.500, jitter=0.000,
> ...

Nothing should trust that dispersion !
 The ntp clients are working as they are designed to?


> The question is:
> How can I force now the clients to follow the server?

Use something other than normal ntpd ?

 ... or wait for about 9 to 12 minutes with the minpoll 4
   on the server, and the default minpoll 6 on the clients;
  On the other hand, if the clients had already ramped the
   poll for the server up to the default maxpoll (10)
    you might have to wait about 2.5hrs.

-- 
E-Mail Sent to this address <BlackList at Anitech-Systems.com>
  will be added to the BlackLists.