June 2011 - FreeSWITCH-users - FreeSWITCH Mailing Lists

NAT Traversal on SFLphone - FS not Auto Changing port
by Avi Marcus 30 Jun '11

30 Jun '11

http://pastebin.freeswitch.org/16627 I've got a Linksys ATA behind NAT and a softphone behind NAT. Both seem to register with the same UDP-NAT string and both have the same type of contact string.. but on one FS is rewriting the RTP IP to work properly - the 1102 works, but on 1000 I don't seem to be getting any audio. Is there some hidden parameter I can't see affecting this? Both are dialing the same extension 9664 default MOH stuff... This softphone actually seems to have a responsive gui in linux. If I can get the audio to actually work, that would great! (Oh, it has tls/srtp too, it seems) Thanks! <http://pastebin.freeswitch.org/16627> -Avi Marcus

3 2

instant messaging via freeswitch
by ran zhang 30 Jun '11

30 Jun '11

hi All: I got the instant messaging working between 2 linphone clients via freeswitch, but i dont see any debug info while messages are sent. which module enables instant messaging in freeswitch as i'm trying to understand more about how freeswitch is handling this.

3 3

Optimize Freeswitch build
by Stephen Wilde 30 Jun '11

30 Jun '11

Hi all, it's possible to optimize the Freeswitch build to maximize the performance? Or this is the default configuration? I have not found any reference to this argument. Stephen

5 7

significant delay waiting for response after playAndGetDigits()
by Wes 30 Jun '11

30 Jun '11

I am experiencing a delay waiting for the system to respond in a lua script that is using playAndGetDigits. I press a key while the voice is still speaking, and it continues to speak for a few seconds before moving to the part of the script for the particular keypress. Also, related: sometimes it speaks the "invalid" message even though the key pressed is a valid one, and then immediately after speaking the invalid message, it continues with the processing for the key that was pressed. So I know I didn't press an invalid key. -- session:execute('flush_dtmf'); session:flushDigits() local digits = session:playAndGetDigits(1, 1, 3, 2000, "#", "phrase:play_submit_rerecord:1:2:3", invalid, "\\d{1}") if ( digits == "1") then ETC here is the phrase: <macro name="play_submit_rerecord"> <input pattern="^([0-9#*]):([0-9#*]):([0-9#*])$"> <match> <action function="speak-text" data="To listen to your dictation, press $1, To submit your dictation , press $2, To discard your dictation and start over, press $3."/> </match> </input> </macro> Is the speak-text just too cpu intensive and causing it to be slow to respond to the DTMF? Any help would be appreciated... thanks!

4 8

Join Us For A ClueCon Party!
by marketing＠cluecon.com 30 Jun '11

30 Jun '11

1 0

UniMRCP Server as daemon won't connect RTSP
by Jason Moran 30 Jun '11

30 Jun '11

I've been playing around with IVONA's new unimrcp plugin with unimrcp 1.0.0 (r1725) on OpenSuse 11.4 Anyway, when I start it up with the following, it stays in the foreground of my SSH w/ the following command: ./unimrcpserver That works! However, since it's in the foreground when I close the SSH window, it kills the process. If I use the so-called daemon mode (-d or --daemon) it says it is going into daemon mode, but FS will immediately return that it "Failed to Connect to RTSP Server..." at the IP:port I specified. The unimrcpserver process is running, but doesn't seem to respond to anything. ./unimrcpserver -d ./unimrcpserver --daemon If I attempt to background it by either using "&" or ctrl-z it says [1]+ Stopped ./unimrcpserver and FS will successfully make the RTSP connection but then nothing will happen. Mod_unimrcp will spit out a warning about "MRCP session has not opened after 5000 ms" ./unimrcpserver & Those are hard to kill. But when I kill -9 it then FS will finally remove the MRCP handle and tell me that it couldn't allocate the speech engine. Lastly, I can nohup it, but then unimrcpserver eats up 95%+ of the CPU (instead of 1-3% as it does when I have it in the foreground), which it does not do when it runs in any other modes. It also makes a very, very large nohup.out file that keeps on growing. Even if I tell it to not log. nohup ./unimrcpserver & nohup ./unimrcpserver -o 0 & Ideas?? Thanks, Jason Moran

3 4

SIP 100 response and media
by Boris Kovalenko 30 Jun '11

30 Jun '11

Hello! I have a customers with Cisco 2821 which want to use IVR on it. He has configured Cisco and told me there is no voice when number is called. The configuration is: Cisco 2821 <-- FreeSwitch (full proxy mode) --> PSTN I looked at tcpdump and the call flow is: FS (INVITE) --> Cisco 2821 Cisco 2821 (100 Trying) --> FS Cisco 2821 (RTP) --> FS I think the call flow is wrong in that case and SIP100 can not go throught Freeswitch (http://jdrosen.net/papers/draft-donovan-mmusic-183-00.txt) and there should be SIP180 to get it work. But customers said this is OK and he has no troubles with same configuration with other provider. What may I do in this sutiation? Any specific Freeswitch configuration? P.S. FreeSWITCH Version 1.0.head (git-1c95ad9 2011-01-20 22-43-50 -0300) -- Regards, Boris

3 10

segfault after 407 response
by covici＠ccs.covici.com 30 Jun '11

30 Jun '11

Hi. I have an fs install from about a week ago, and I am using an extension on the external sip port of 5060. Its default context is public and I have uncommented the last section where it asks for a sip authorization. When it receives the 407 challenge, I am getting a seg fault. I have traces, but they are with the default configs, so a lot of things are optimized out. Should I do them again, file a bug, or is this something which is known? Thanks in advance for any suggestions. -- Your life is like a penny. You're going to lose it. The question is: How do you spend it? John Covici covici(a)ccs.covici.com

4 5

Public users with internal sip profiles
by Bryan Smart 30 Jun '11

30 Jun '11

I'd like both authenticated and unauthenticated users to connect to the same IP and port. I will separate them by dialplan contexts. At the moment, only authenticated users can connect to the internal sip profile. I have a "default" user in the directory. Its password is blank, and I've set it to use the public dialplan context. I thought that would cause unauthenticated users connecting to the internal sip profile to be authenticated as the "default" user. Instead, the console says: Rejected by ACL "domains". Falling back to digest auth. This is all I'm told, even at log level 9. I read about the acls. They seem to cause authentication based on the network address, so I thought they might be blocking connections as the "default" user. They don't seem to block connections to the external sip profile, though. The default internal sip profile is an extremely long read. I've been through it more than a few times, but still don't understand a lot of it. I can disable authentication completely, but I need it to be optional. Are the settings that I want in the sip profile? Thanks for any clues. Bryan

2 1

SIP Profile
by sharad 30 Jun '11

30 Jun '11

Hi, I want my FS to entertain only those calls which are coming from a specific IP address . For this, I have defined a profile in vars.xml as <X-PRE-PROCESS cmd="set" data="profile_00=192.168.4.10"/>. & another line in public.xml as <extension name="transfer_to_context_01"> <condition field="${sip_from_host}" expression="$${profile_00}"> <action application="transfer" data="${destination_number} XML sharad"/> </condition> </extension> As per my understanding, if call is coming from 192.168.4.10, that call should be processed as per the context `sharad'.....correct ? Yes, it works like this. Upto here all is ok. But if the call is coming from 192.168.100 - 192.168.4.109, these calls also are processed as per context `sharad'. Means entire IP is not matched by condition field. Is there any solution for this. Requesting all of you to help me in this regard. Regards Sharad

4 4