[OpenSIPS-Users] Questions about initial setup

Kurtis Heimerl kheimerl at endaga.com
Thu May 15 19:46:49 CEST 2014


This makes *a lot* of sense. Thanks!I expected the SIP REGISTER to be
terminated at the instance rather than forwarded as I had that user/pw
combo in the subscribers table. I assume it doesn't do that by default I
have to add a command to the routing logic to do that? Or, architecturally,
should I just forward that request anyhow?

Thanks!


On Thu, May 15, 2014 at 2:11 AM, Brett Nemeroff <brett at nemeroff.com> wrote:

> Normally this means that you haven't actually done anything with the call,
> but you are t_relaying it out.
>
> In other words, the RURI was destined for your opensips box.. That's how
> it got there.. it hit opensips, then you sent it back out.. But you never
> adjusted the RURI to point to the next hop by either changing the request
> user part or domain part. So it sends it back to itself (loopback) until
> Max-Forwards expires and returns the 483.
>
> If you remove the 483 check, you get a massive loop on the loopback
> interface until you start to kill server resources, which gave you the 408.
>
> Be sure you are actually doing some change to the message before you send
> it back out. Try to xlog the r-uri before you t_relay() like this:
>
> xlog("L_INFO","Sending call out to $ru");
>
> If you watch sip traffic on your loopback interface, you'll see what all
> the madness is about.. I don't really recommend it, since it very quickly
> will consume your terminal, but you'll at least understand what exactly is
> happening..
>
> Good luck!
> -Brett
>
>
>
> On Wed, May 14, 2014 at 8:36 PM, Kurtis Heimerl <kheimerl at endaga.com>wrote:
>
>> Hey Users,
>>
>> I'm a developer with a lot of experience in FreeSWITCH trying out
>> OpenSIPS for a larger, more central piece of infrastructure. I started
>> investigating it a few days ago with the following use case in mind:
>>
>> FS instance with users attached.
>> OpenSIPS in the cloud routing the FS communications to another cloud
>> provider (e.g., Twilio).
>>
>> I've got FS set up with our OpenSIPS box as an external provider and a
>> set username and password. That seems to be working. I set up OpenSIPS from
>> the opensips APT repository (http://apt.opensips.org/) and version 11.1.
>> It's running, configured with TEXTDB, and I inserted the username and
>> password (after modifying the DB schema to allow null emails, which I guess
>> is an ancient bug... http://sourceforge.net/p/openser/bugs/593/).
>>
>> The registration from FS is being rejected as a 483 "Too many hops" This
>> is strange. I removed the 483 check in the opensips cfg and now it's a 408
>> Timeout. OpenSIPS itself it complaining in the following way. The actual
>> log is also available, but too large for the mailing list.
>>
>> May 15 01:27:34 [26230] WARNING:core:fm_malloc: Not enough free memory,
>> will atempt defragmentation
>> May 15 01:27:34 [26230] ERROR:tm:relay_reply: no more share memory
>> May 15 01:27:34 [26230] WARNING:core:fm_malloc: Not enough free memory,
>> will atempt defragmentation
>> May 15 01:27:34 [26230] ERROR:tm:_reply_light: failed to allocate shmem
>> buffer
>>
>> Googling this error brings up a few hits on issues with a memory leak,
>> but that can't be the case here as it is just started.
>>
>> Ideas? What am I doing wrong? Is this is right use case for OpenSIPS
>> anyhow?
>>
>> Thanks!
>>
>> _______________________________________________
>> Users mailing list
>> Users at lists.opensips.org
>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>
>>
>
> _______________________________________________
> Users mailing list
> Users at lists.opensips.org
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20140515/def88988/attachment.htm>


More information about the Users mailing list