[asterisk-ha-clustering] load balance with openser

Edgar Guadamuz eguadamuz at gmail.com
Sat Apr 19 16:07:00 CDT 2008


So, if Asterisk boxes work as registrar servers, how should be the
handling for REGISTER transactions?

Let's suppose we have 2 asterisk boxes behind an openser server. We
are load-balancing the INVITE requests, but how would be the handling
for REGISTER messages in this case?

if (method =="REGISTER"){

????????? What would be here????????

}




On Fri, Feb 22, 2008 at 2:15 AM, Siller Janosch
<janosch.siller at badenit.de> wrote:
> Hi Edgar,
>
>  > -----Ursprüngliche Nachricht-----
>  > Von: asterisk-ha-clustering-bounces at lists.digium.com
>  > [mailto:asterisk-ha-clustering-bounces at lists.digium.com] Im
>  > Auftrag von Edgar Guadamuz
>  > Gesendet: Donnerstag, 21. Februar 2008 20:11
>  > An: Asterisk High Availability and Clustering List -
>  > Non-CommercialDiscussion
>  > Betreff: [asterisk-ha-clustering] load balance with openser
>
>
> >
>  > I was looking at
>  > http://www.openser.org/dokuwiki/doku.php/asterisk:load-balanci
>  > ng-and-ha
>  > to see how do load balance with OpenSER.
>  >
>  >
>  > As there can be seen, the balance is reach through the INVITE
>  > messages:
>  >
>  > if ( method=="INVITE" ) {
>  > # dst_select( "GROUP", "HASH METHOD")
>  >   dst_select("1","4");
>  >   sl_send_reply("100","Trying");
>  >   forward(uri:host, uri:port);
>  >   exit();
>  > }
>  >
>  >
>  > My question is: where the user register? in the openser box or in the
>  > asterisk boxes?
>
>  Asterisk is the registrar in this example.
>  If openser should be the registrar, you have to authorize/challange the user before you forward the request to asterisk boxes.
>
>  br,
>  janosch
>
>  _______________________________________________
>  --Bandwidth and Colocation Provided by http://www.api-digital.com--
>
>  asterisk-ha-clustering mailing list
>  To UNSUBSCRIBE or update options visit:
>    http://lists.digium.com/mailman/listinfo/asterisk-ha-clustering
>



More information about the asterisk-ha-clustering mailing list