----- Original Message -----
> From: "Olle E. Johansson" <[email protected]>
> To: "Asterisk Developers Mailing List" <[email protected]>

> So we have yet another internal resolver? Is that a good thing? Why
> are we not using the system resolver?
> 
> We need to have some direction here. I think adding yet another DNS
> resolver is bad and will make it hard to add functions like
> DNSsec/DANE support. Making it possible for one piece of asterisk to
> use another DNS resolver is poor design, we should not open up for
> that.  The number of debug problems and runtime issues I see is
> reaching infinity. "I can ping this server, but the sip channel
> can't reach it" is just a start.
> 
> Is there a way we can either avoid using yet another resolver, or
> since this is propably better - switch all of Asterisk to use it and
> we'll get asynch DNS everywhere?
> 
> I can understand why a library like PJsip have support for this, in
> some cases it's the only way when writing clients like soft phones.
> That doesn't mean we have to expose it. We can still select and
> decide for ourselves about our design.
> 
> Summary
> - Don't add yet another DNS resolver randomly
> - Don't allow using different DNS servers in parts of a server app
> and please don't allow anyone to configure anything else than the
> system resolvers in a server application.

I understand what Olle is getting at and I agree with it.  Why do we want 
res_pjsip using something different from the rest of Asterisk?  Everything 
should be using the same DNS resolver.  Otherwise, we are opening up a can of 
worms here when it comes to supporting Asterisk and trying to troubleshoot DNS 
issues.

This just doesn't seem like the right direction to go here.

Michael

-- 
_____________________________________________________________________
-- Bandwidth and Colocation Provided by http://www.api-digital.com --

asterisk-dev mailing list
To UNSUBSCRIBE or update options visit:
   http://lists.digium.com/mailman/listinfo/asterisk-dev

Reply via email to