Opensips 1.5.0 Issue with LCR next_contacts

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Opensips 1.5.0 Issue with LCR next_contacts

Amit Sharma-2
Hi,
  I was planning to upgrade to version 1.5.0 from our current version
1.2.2 which is quite old. We are currently using lcr module
functionality to fork to contacts based on q-values

  However, it seems that opensips keeps trying the same contact in
case only a single contact is registered. I guess one of the reasons
is that next_contacts is returning a positive value in case
contact_avp doesn't exist even when called from failure route. The
changes to unify the ruri and branch handling seem to have changed
this behavior. Is this a intentional change in behavior of
next_contacts?

  Are there any additional changes required (apart from defining the
avp's)  to opensips.cfg to use the above functionality in 1.5



Thanks,
Amit

_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
Reply | Threaded
Open this post in threaded view
|

Re: Opensips 1.5.0 Issue with LCR next_contacts

Bogdan-Andrei Iancu
Hi Amit,

Amit Sharma wrote:
> Hi,
>   I was planning to upgrade to version 1.5.0 from our current version
> 1.2.2 which is quite old. We are currently using lcr module
> functionality to fork to contacts based on q-values
>  

more suitable and proofed are the core functions:
    serialize_branches() + next_branches()

their were added especially to get rid of the lcr dependency in other
modules.
>   However, it seems that opensips keeps trying the same contact in
> case only a single contact is registered. I guess one of the reasons
> is that next_contacts is returning a positive value in case
> contact_avp doesn't exist even when called from failure route. The
> changes to unify the ruri and branch handling seem to have changed
> this behavior. Is this a intentional change in behavior of
> next_contacts?
>  
I think this is the result of a bit of inconsistency in prev lcr version
- in request_route, if no AVP was found, true was returned; in
failure_route, in the same case, false is return.

Right now there is no difference between request or failure route (there
is the same behaviour), but I think "false" should be returned if
noother AVP is available
>   Are there any additional changes required (apart from defining the
> avp's)  to opensips.cfg to use the above functionality in 1.5
>  
if you use LCR only for q-based forking, batter switch to the core
functions.

Regards,
Bogdan

>
>
> Thanks,
> Amit
>
> _______________________________________________
> Users mailing list
> [hidden email]
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>
>  


_______________________________________________
Users mailing list
[hidden email]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users