[Users] Re: get_redirects("*"), serialize_branches(0), next_branches() not working for me

Bogdan-Andrei Iancu bogdan at voice-system.ro
Mon Jun 5 17:18:18 CEST 2006


Greg,

that's great if it works now.. regarding the SRV lookup - as the log 
says, no SRV and NAPTR lookups are done since the SIP URI specify a 
port. Get rid of the port number from the contact uri if you want to 
have naptr / srv lookup also.

regards,
bogdan

Greg Fausak wrote:

> Bogdan,
>
> I recompiled openser and reinstalled, it's working for
> me now.  I think I hacked something up and caused it to break!
>
> One thing I still am curious about.
> I've got a domain name with an SRV record: ab5350.node-1.com, the
> SRV (_sip._udp.ab5350.node-1.com) points to a single A record
> proxy.ab5350.node-1.com.  When I make that the Contact in
> the redirect I get this in the syslog:
>
> Jun  5 09:39:51 www openser: DEBUG: mk_proxy: doing DNS lookup...
> Jun  5 09:39:51 www openser: DEBUG:sip_resolvehost2: has port -> do A
> record lookup!
> Jun  5 09:39:51 www openser: ERROR: mk_proxy: could not resolve
> hostname: "ab5350.node-1.com"
> Jun  5 09:39:51 www openser: ERROR: uri2proxy: bad host name in URI
> <sip:+12143357976 at ab5350.node-1.com:5060>
> Jun  5 09:39:51 www openser: ERROR:tm:t_forward_nonack: failure to
> add branches
> Jun  5 09:39:51 www openser: ERROR:tm:w_t_relay: t_forward_nonack failed
> Jun  5 09:39:51 www openser: DEBUG:tm:relay_reply: branch=0, save=0,
> relay=0
>
> My code does a next_branches() followed by t_relay().  I thought that
> t_relay() did the necessary SRV lookups?
>
> If I change my redirect to return proxy.ab5350.node-1.com in the Contact
> list, everything works as expected.
>
> -g
>
>
> On 6/5/06, Bogdan-Andrei Iancu <bogdan at voice-system.ro> wrote:
>
>> yes - get_redirects("*") loads the contacts from 3xx reply into branches
>> (as for parallel forking) ; serialize_branches(0) converts the branches
>> into AVPs for serial forking ; next_branches() gets the first AVP for
>> serial forking and push it as branch.
>>
>> the message from next_branches() says all AVPs fro serial branches were
>> consumed - or none was present from the beginning.
>>
>> regards,
>> Bogdan
>





More information about the sr-users mailing list