Hello,
o Randy Jackson [01/01/09 23:47]:
Hello Stefan,
Thanks for the info in the reply below. Apologies for sending the debug
dump attached to the email. Wasn't aware that I should have used pastebin.
I took your advice and upgraded to the Ser 2.0 rc1 and Sems 1.0.0. I could
not figure out how to get the voicemail app to work without the Ser to add
the email address. Maybe I just don't understand everything well enough
yet. This is ok because I think I would prefer this configuration for
scalability. You were right to suspect a NAT, the RTP was expected on a
private ip address which could not be reachable by my client. I finally made
this correlation when I started the sems server and the 'local media IP' was
set to a non-routeable ip. So to fix this, in the sems.conf file there is a
parameter called media_ip which should be set to the ip_addr or nic device
the sems will use to send and receive media. Setting this parameter did not
work for me, however when I added the listen parameter from an older config
file I was able to get the RTP streamed media via the public ip address. Is
the media_ip parameter suppose to work? If I've done something wrong please
let me know? Thanks for your help.
in sems 1.0.0 rc1 you can set listen=xyz which will be the interface
used for both media and signaling.
If you get a newer version from the 1.0.0 branch (svn co
https://svn.berlios.de/svnroot/repos/sems/branches/1.0.0), or, better,
the 1.1.0 RC1
(http://ftp.iptel.org/pub/sems/1.1/rc1/src/sems-1.1.0-rc1.tar.gz) you
can configure media_ip and sip_ip separately. This way you can have the
SEMS server listen only for traffic on the private interface from your
proxy, but still send and receive RTP on the public interface.
media_ip was introduced in 1.0.0 branch after 1.0.0 RC1 - sorry for the
version confusion. This reminds me that we still need to release 1.1.0
and 1.0.1, which will hopefully happen "real soon now".
You can always find the configuration parameters with explanations in
the source tree at core/etc/sems.conf.sample, or if you do 'make doc'
you will find a link from the main documentation page
(doc/doxygen_doc/html/index.html).
Best Regards
Stefan
Regards,
-Randy
-----Original Message-----
From: Stefan Sayer [mailto:[email protected]]
Sent: Monday, December 15, 2008 10:04 AM
To: [email protected]
Cc: sems
Subject: Re: SER->SEMS Config Not recording the voice message ->
AmAudioEvent::noAudio event!
Hello,
this list has a message size limit of 40k; if you have large debug logs to
post, you can use pastebin or similar.
Subject:
SER->SEMS Config Not recording the voice message -> AmAudioEvent::noAudio
event!
From:
"Randy Jackson" <[email protected]>
Date:
Thu, 11 Dec 2008 15:16:24 -0500
To:
<[email protected]>
Hello,
I have sems-0.10.0-rc2 & ser-0.9.6-sems loaded on a Linux centos 5
system. I am trying to run the voicemail application in the default
configuration as identified in the instructions for the ser/sems combo
that I have installed. Per the instructions I have downloaded the
source to /usr/local/src in their respective directories and installed
Ser for Sems in /opt/ser-sems. Right now I'm trying to understand and
resolve a number of issues. The first being passing the P-App-Name
and P-App-Param headers with the INVITE. I have tried to pass these
values a number of ways with no success because the avpops:load_avps:
db_load fails on the language. I have worked around this by hardcoding
the email and language with the avp_write before the call is made from
ser to sems. When doing so the voicemail application answers the call
and presumably records the message. I say presumably because I'm
seeing the following indicators that the recording of the message and
the sending of the e
mail are not really taking place. First indicator in the sems debug output
is after the AmSession::SipRequest->onSessionStart I'm getting a "posting
AmAudioEvent::noAudio event". Additionally when the sendMailNotification is
called the log shows that the record size is 0. I'm expecting the recorded
voice file to be in /tmp/<sessionid>.wav, but I'm not seeing the file show
up. I've started to trace the code, but I sincerely believe that there is
some that I have configured wrong. I have included the configuration files
and debug output to assist in clarifying what I have stated. Can you help
me get the right configuration so that I can successfully record a message
and send it via email? Thanks in advance.
-Randy
About passing of parameters to SEMS, you will find string composition much
simpler with ser 2.1 functions.
As you already have noticed SEMS did not record any audio:
(4373) DEBUG: sendMailNotification (AnswerMachine.cpp:343): recorded data
size: 0 I suppose you are behind a NAT and your client sends RTP to some
public address, thus it does not reach the voicemail server. SEMS expects
RTP from 10.x.x.228/43436 . If you run tcpdump, do you see RTP flowing to
SEMS?
The AmAudioEvent::noAudio is sent to notify the application of the end of
the welcome message, so that it knows it should start recording.
I would also suggest trying SEMS 1.0 or 1.1 RC1 as it is way easier to use
(no need for SER as SIP stack any more etc),
http://ftp.iptel.org/pub/sems/1.1/rc1/src/sems-1.1.0-rc1.tar.gz
BR
Stefan Sayer
-------- Original Message --------
Subject: Forward of moderated message
Date: Mon, 15 Dec 2008 15:47:27 +0100
From: [email protected]
To: [email protected]
--
Stefan Sayer
VoIP Services
[email protected]
www.iptego.com
IPTEGO GmbH
Am Borsigturm 40
13507 Berlin
Germany
Amtsgericht Charlottenburg, HRB 101010
Geschaeftsfuehrer: Alexander Hoffmann
_____
avast! Antivirus <http://www.avast.com> : Outbound message clean.
Virus Database (VPS): 081228-0, 12/28/2008
Tested on: 1/1/2009 5:46:01 PM
avast! - copyright (c) 1988-2009 ALWIL Software.
--
Stefan Sayer
VoIP Services
[email protected]
www.iptego.com
IPTEGO GmbH
Am Borsigturm 40
13507 Berlin
Germany
Amtsgericht Charlottenburg, HRB 101010
Geschaeftsfuehrer: Alexander Hoffmann
_______________________________________________
Sems mailing list
[email protected]
http://lists.iptel.org/mailman/listinfo/sems
