Good to know about both -T and -W options. In many of my rtpproxy installations I migrated to 2.0, because the default packaged rtpproxy version from distros (mainly stable debian or centos -- which come with some 1.2.1+patches), seems to be a problem with re-INVITEs changing the RTP port. However, rtpproxy v2.0 doesn't seem to be packaged yet in the stable distros out there, hopefully they will update it soon, it will make the life easier to deploy.

Cheers,
Daniel

On 12.07.17 15:14, Maxim Sobolev wrote:
Yes, Daniel is correct you can bump overall timeout with the -T option. There is also another option -W now since 3 years ago (2.0+) to set timeout specifically on early sessions, so that you can have bigger timeout during call setup phase which might be relevant in this case. Hope it helps.

-Max

On Jul 12, 2017 4:42 PM, "Daniel-Constantin Mierla" <miconda@gmail.com> wrote:

Hello,

knowing that rtpproxy v2.0 has -T is very useful. v1.2 is very old, so I think its time to migrate to v2.0.

With this occasion I checked rtpengine -- it has a bunch of options related to timeouts, so it's a good alternative to go for it as well.

Cheers,
Daniel


On 12.07.17 09:24, Marian Piater wrote:
Hi,
works fine, after recompile RTPProxy with newer version (2.2.alpha.20160822) and using option -T 150. There was no necessary to modify the source code.

/usr/bin/rtpproxy -s udp:127.0.0.1 7722 -u rtpproxy rtpproxy -p /var/run/rtpproxy/rtpproxy.pid -F -l XXX.XXX.XXX.XXX -d DBUG LOG_LOCAL0 -T 150

Thank you again,
Marian


Dňa 12.7.2017 o 09:18 Daniel-Constantin Mierla napísal(a):

Hello,

ok -- it would be good to know if that solved the problem, so others can find the answer in the archive when facing a similar situation.

Cheers,
Daniel


On 11.07.17 15:55, Marian Piater wrote:

Hello,

thank you for quick reply, I tried to find some options in the man pages, but I didn't read HOWITWORKS section.

I will try to compile RTPProxy from source.

Regards Marian


Dňa 11.7.2017 o 15:36 Daniel-Constantin Mierla napísal(a):

Hello,

doing 'man rtpproxy' on debian jessie reveals:

"The proxy tracks idle time for each of existing sessions (i.e. the time within which there were no packets relayed), and automatically cleans up a sessions whose idle times exceed the value specified at compile time (60 seconds by default)."

Which seems to explain exactly what happens in your case.

I knew there was an inactivity cleanup routine, but I expected to be higher than 60sec. Interesting that it didn't hit me badly so far. I haven't checked the source code to see if it still matches the docs, though, but I see that master branch of rtpproxy still has this remark in the manual:

  - https://github.com/sippy/rtpproxy/blob/master/rtpproxy.8#L335

An issue should be created for rtpproxy to at least increase this value, or better make it an option for command line.

Cheers,
Daniel

On 11.07.17 15:21, Marian Piater wrote:

Hi Daniel,

we are using 1.2.1-2.1 version from Debian Jessie repository.

rtpproxy                        1.2.1-2.1                          amd64        Relay for Real-time Transport Protocol (RTP) media streams

Regards,

Marian


Dňa 11.7.2017 o 13:09 Daniel-Constantin Mierla napísal(a):

Hello,

what is the version for rtpproxy?

Cheers,
Daniel



-- 
Daniel-Constantin Mierla
www.twitter.com/miconda -- www.linkedin.com/in/miconda
Kamailio Advanced Training - www.asipto.com
Kamailio World Conference - www.kamailioworld.com

-- 
Marian Piater

VoIPsun s.r.o.
+420 608 24 58 42
marian.piater@voipsun.cz
www.voipsun.cz

-- 
Daniel-Constantin Mierla
www.twitter.com/miconda -- www.linkedin.com/in/miconda
Kamailio Advanced Training - www.asipto.com
Kamailio World Conference - www.kamailioworld.com

_______________________________________________
Kamailio (SER) - Users Mailing List
sr-users@lists.kamailio.org
https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users



-- 
Daniel-Constantin Mierla
www.twitter.com/miconda -- www.linkedin.com/in/miconda
Kamailio Advanced Training - www.asipto.com
Kamailio World Conference - www.kamailioworld.com