On Thu, 2009-03-26 at 16:34 -0400, M. Ranganathan wrote:
> On Thu, Mar 26, 2009 at 4:31 PM, M. Ranganathan <[email protected]> wrote:
> > On Thu, Mar 26, 2009 at 2:12 PM, Scott Lawrence
> > <[email protected]> wrote:
> >> On Thu, 2009-03-26 at 11:27 -0400, Andy Spitzer wrote:
> >>>
> >>> I disagree.  The idea solution would be to allow sipXrelay to start
> >>> even if it cannot get STUN.  It should continue to try to get STUN
> >>> periodically, and alarm if it cannot after some period (and alarm
> >>> again when it actually gets an answer, and alarm again if it loses it
> >>> again).  That gets rid of the dependency problem yet still allows
> >>> sipXrelay to be started before sipXproxy.
> >>>
> >>> Using sipXsupervisor as a retry mechanism...I don't like it.
> >>
> >> +2
> >>
> >
> > Is it appropriate to send an alarm if stun server does not work during
> > configtest ? I assume ( based upon this discussion ) that failing
> > configtest is not the right thing to do.
> 
> There are actually two conditions here.
> 
> 1. The stun server address is plain wrong ( bad ip address or cannot
> even be resolved)
> 2. The address is OK but there is no stun server running.
> 
> I assume that in case 1, the right thing to do is to FAIL configtest.
> I assume in case 2 the right thing to do is PASS configtest but send an alarm.

I don't believe that in practice those two are distinguishable, so you
have to treat them identically.  The alarm gives the user a heads-up to
check and figure it out.


_______________________________________________
sipx-dev mailing list
[email protected]
List Archive: http://list.sipfoundry.org/archive/sipx-dev
Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev

Reply via email to