[ntp:questions] refclock use causes core dump of ntpd
Richard B. gilbert
rgilbert88 at comcast.net
Thu Feb 22 13:16:51 UTC 2007
Ronan Flood wrote:
> Harlan Stenn <stenn at ntp.isc.org> wrote:
>
>
>>The downside of your approach is that you'll have to keep making those
>>changes with every upgrade.
>
>
> Maybe you should just incorporate his changes and hope for the best :-/
>
> Is it appropriate anyway for a refclock driver to terminate ntpd if it
> isn't happy? From a quick look, only refclock_true and refclock_jupiter
> call abort(), and only refclock_oncore and refclock_mx4200 call exit().
> All the others presumably mark the refclock "bad" in some way and allow
> things to carry on, ignoring it.
>
Maybe there should be a specification, perhaps in the new RFC, that says
what to do if a refclock failure is detected.
Marking the refclock "insane" comes to mind as a reasonable response.
More information about the questions
mailing list