- FreeSWITCH-users - FreeSWITCH Mailing Lists

Single Server having two sangoma cards and can't handle full capacity
by bawajan 23 Jul '14

23 Jul '14

Hi, We are having two A108-8port & A116-16port Sangoma cards connected with 23 E1 links on Single Server, each E1 link has 30 channels although we have 24 ports we are using 23 so totally 23*30=690 channels We have integrated ESL library with java and making outbound calls, not using any dialplan. The issue what we are facing is when we use all 690 channels, we see segmentation_fault in Freeswitch and big dump file is written (ie., > 15gb) and which would crash the system, because of high CPU/IO. But if we reduce the channels say, 450 we don't see any issue at all. What could be the reason for this issue. Appreciate any Help... Thanks & Regards, Bawajan -- View this message in context: http://freeswitch-users.2379917.n2.nabble.com/Single-Server-having-two-sang… Sent from the freeswitch-users mailing list archive at Nabble.com.

2 1

FreeSWITCH Week in Review(Master Branch) July 13th-19th
by Kathleen King 22 Jul '14

22 Jul '14

Hello, again. This week in the FreeSWITCH master branch we had 69 commits. FreeSWITCH's git URLs were updated to Stash this week, mod_skinny added the ability to set flash or ring on inbound calls, and mod_verto saw some more tweaks and improvements this week and, if you haven't already, go check it out! You can access the demo by following this link using the latest Chrome: https://conference.freeswitch.org/verto/ <http://conference.freeswitch.org/verto/> The following bugs were squashed: 17d6c5b FS-6578 Update windows build to use openssl 1.0.1h. Jira: http://jira.freeswitch.org/browse/FS-6578 e7700c0 FS-6665 Fix for garbled audio using verto without stereo for mod_verto. Jira: http://jira.freeswitch.org/browse/FS-6665 91d405a Call recovery_track on recovering channels once the recovery has completed and fix race condition with repeated recovery in FS core and mod_sofia. 3df080a FS-6447 #close SUBSCRIBER_ABSENT also needs handling. Thanks to Ben Langfeld. Jira: http://jira.freeswitch.org/browse/FS-6447 3a00260 FS-6615 Fix possible crash with api_execute response in language modules. Jira: http://jira.freeswitch.org/browse/FS-6615 New features that were added: 3ce4ae9 FS-6540 Added notify functionality in mod_sofia. Please test. Jira: http://jira.freeswitch.org/browse/FS-6540 997e921 Add transfer command for mod_verto. 502cd3f Fix verto compile error on arm in mod_verto. 4e34aa6 Added hash auto dialing to verto.js in the demo js for mod_verto. 41ea9c5 Updated FS git URLs to point to Stash. 835fae4 Any browser (Firefox, Chrome, Opera) that can run Verto can do TLS1.2, We shouldn\'t lower the bar when we don\'t need too for mod_verto. 3526ca5 Allow setting threshold for RTP auto adjust in FS core. b6566b8 FS-5384 Enabled the flash indicator on ring in mod_skinny. Jira: http://jira.freeswitch.org/browse/FS-5384 6f12de9 Adding hash auto dialing in verto.js in the demo js for mod_verto. 025fe24 Added and improved error checking for invalid data in jquery.verto.js in the demo js of mod_verto. 74262ee Add `session::consoleLog2()` to FS core. d1ec643 Improve name of new console log function in FS core. 8d11e8a Removed git.freeswitch.org URL and replaced it with the updated Stash URL. 58f7376 Improvments to hash auto dialing in verto.js in the demo js for mod_verto. db63dc5 Improvements to dropped connection handling in mod_verto. ff2edeb Add blind-reg param to mod_verto. c58e64b Added some more headers to sip messages in mod_sofia. 236233d Switch to incall page on early media too in mod_verto. 1b73601 Associate "sending early media" log with session in mod_sofia. a2921e2 Update verto with some fixes and enhancements with reconnection in mod_verto. 8d33af4 Clean up verto demo and add examples of moderator controls in mod_verto. ab292dc Add some more signal tracking in mod_verto. 958d9d9 FS-5384 Implement support for ring-on-active and ring-on-idle attributes in mod_skinny. Jira: http://jira.freeswitch.org/browse/FS-5384 Improvements in cross platform build supports: 730cd20 Remove js and mod_spidermonkey from Windows build. 62404c6 Remove auto generation of files that already exist in git for Windows. bf19719 OpenBSD Auto Build. a0e19e1 Removed libportaudio and use system portaudio in mod_portaudio. b186ca9 FS-6542 Add pkg_config path for portaudio on mac via homebrew. Jira: http://jira.freeswitch.org/browse/FS-6542 58a3398 Add missing include for alloc.h in spandsp. 7180513 FS-6542 Add missing download to windows build Jira: http://jira.freeswitch.org/browse/FS-6542 In an effort to determine how many people read this review, I've decided to buy a drink at ClueCon for one person that responds with feedback by Friday. You must be going to ClueCon to redeem this offer. Feedback is welcome at kathleen.king(a)quentustech.com and the commits for this week are in the attached file along with the corresponding Jira links. -- Kathleen King Quentus Technologies, INC 1037 NE 65th St Suite 273 Seattle, WA 98115 Main: (253) 617-3210 Cell: (703) 859-3757 kathleen.king(a)quentustech.com

1 0

mod_callcenter - multiple queues with shared agents?
by Nathan Neulinger 22 Jul '14

22 Jul '14

Looking for possible ways of implementing call center with "expertise areas" for agents. One possibility was simply having multiple queues - with agents signing in to each queue for which they are qualified. However, it seems like this might have issues with the agent's "busy" status not being shared across the queues. Is there any other mechanism that would be suitable? It looked like tiers might also be a way to accomplish it, but from the docs sounded like completely different than how they were intended to be used. Intent is to flag the incoming call based on progression through an IVR, picking a particular group of agents based on the selection. (Generally a subset for lower skilled.) Thoughts? -- Nathan ------------------------------------------------------------ Nathan Neulinger nneul(a)mst.edu Missouri S&T Information Technology (573) 612-1412 System Administrator - Architect

1 0

RTP port rewriting stops working when combined with transcoding?
by Kees Jan Koster 22 Jul '14

22 Jul '14

Dear All, I am trying to figure out why I see what I see. When I make a call using the same codec (say, PCMA/8000 on both sides), FreeSwitch rewrites the RTP stream's IP address and port for me. That I like, since my FreeSwitch is behind NAT. From the logs: 2014-07-16 12:32:36.815907 [INFO] switch_rtp.c:3738 Auto Changing port from 1.2.3.4:4040 to 5.6.7.8:4040 2014-07-16 12:32:36.995906 [INFO] switch_rtp.c:3738 Auto Changing port from 9.10.11.12:12345 to 13.14.15.16:41730 Now, when I try to use two different codecs, this behaviour stops. From the logs and without changing the configuration or restarting FreeSwitch: 2014-07-16 12:35:36.495902 [INFO] switch_rtp.c:3738 Auto Changing port from 1.2.3.4:4042 to 5.6.7.8:4042 The other line is missing and ngrep reveals how FreeSwitch stubbornly sends RTP audio to the wrong port and IP address. Further testing shows: PCMA -> PCMA: IP/port is rewritten SILK -> PCMA: IP/port is *not* rewritten GSM -> PCMA: IP/port is rewritten Speex -> PCMA: IP/port is *not* rewritten How can I have transcoding *and* RTP IP/port rewriting work at the same time, please? Is this a per-codec thing? -- Kees Jan http://java-monitor.com/ kjkoster(a)kjkoster.org +31651838192 Change is good. Granted, it is good in retrospect, but change is good.

4 6

ZRTP failure after upgrading
by sangdrax8 22 Jul '14

22 Jul '14

I have been using freeswitch with ZRTP successfully, and recently decided to do a pull from master. We haven't updated in a while, and I am testing how our current configurations work with the latest git pull. After the pull, my clients can't seem to establish a ZRTP session. Should I be doing this differently, or is there a new variable I need to be setting? I was under the impression that proxy_media would pass the media through unaltered, so I am a bit confused as to why this would change between versions. Previous version: (Version 1.5.8b git beb0673 2013-11-27 00:48:49Z 64bit) In this version I would look for a zrtp-hash, and if present I would simply proxy_media. <condition field="${endpoint_disposition}" expression="^(DELAYED NEGOTIATION)"/> <condition field="${switch_r_sdp}" expression="(zrtp-hash)" break="never"> <action application="set" data="proxy_media=true"/> </condition> Current version: (Version 1.5.14b git 5ea1729 2014-07-21 11:48:10Z 64bit) Using the same code as above, my endpoints do not negotiate ZRTP. Eventually my clients issue a hangup, since I have them set to require ZRTP for testing purposes. 2014-07-22 16:44:10.891255 [INFO] switch_channel.c:3304 sofia/shared/ sip:7kkklf@192.168.10.134:33588 Activating ZRTP passthru mode. 2014-07-22 16:44:10.891255 [NOTICE] sofia.c:7181 Channel [sofia/shared/ sip:7kkklf@192.168.10.134:33588] has been answered 2014-07-22 16:44:10.891255 [INFO] switch_core_media.c:5152 Activating RTCP PORT 0 2014-07-22 16:44:10.891255 [INFO] switch_core_media.c:5333 Activating ZRTP PROXY MODE 2014-07-22 16:44:10.891255 [INFO] switch_core_media.c:5152 Activating RTCP PORT 0 2014-07-22 16:44:10.891255 [INFO] switch_core_media.c:5333 Activating ZRTP PROXY MODE 2014-07-22 16:44:10.891255 [NOTICE] switch_ivr_originate.c:3494 Channel [sofia/shared/cueyeo@testcust] has been answered 2014-07-22 16:44:11.331267 [INFO] switch_rtp.c:5622 Auto Changing port from 192.168.10.134:60402 to 1.1.1.1:50944 2014-07-22 16:44:16.391257 [NOTICE] sofia.c:947 Hangup sofia/shared/cueyeo@testcust [CS_EXECUTE] [NORMAL_CLEARING] The only suggestions I can find on the wiki are at " http://wiki.freeswitch.org/wiki/ZRTP" which indicates that zrtp_secure_media and rtp-enable-zrtp needs to be set, which I have done.

1 1

ClueCon Weekly Update! And ClueCon reminder!
by Ken Rice 22 Jul '14

22 Jul '14

Hey FreeSWITCHers! Just a Few Reminders! This week, Wed July 23rd, Fred Dixon from BigBlueButton will be joining us on ClueCon Weekly! Tomorrow, 1PM Eastern, 10A Pacific! Join us at http://conference.freeswitch.org/ and see what the BBB guys have been working on! Also if you have not yet booked your Hotel Room for ClueCon, rooms are running very low! Get those reservations in today at http://goo.gl/biKxyy (at last report from the hotel, we¹re breaking the 90% capacity mark for ClueCon! If you have issues booking a room contant me or Brian West and we¹ll get you sorted! Don¹t Forget to Follow us on Twitter @FreeSWITCH and @ClueCon for the latest updates on whats going on! I¹m going to ClueCon are you? There still some slots left for the PizzaParty on Tuesday Night! Get those registrations in today so you don¹t miss out on all the fun! -- Ken http://www.FreeSWITCH.org http://www.ClueCon.com http://www.OSTAG.org irc.freenode.net #freeswitch Twitter: @FreeSWITCH

1 0

Dropped and duplicated DTMF digits
by Chris Young 22 Jul '14

22 Jul '14

Hi all, We are experiencing problems with DTMF collection using FreeSWITCH 1.4.6. The symptoms are missed or duplicated digits when decoding DTMF sent using RFC 2833 from Cisco phones; for example, given the sequence: 1 2 3 4 5 6 7 8 9 0 # we might typically end up with something like this after FreeSWITCH has processed the received RTP: 1 2 3 5 7 7 8 9 9 # Using Wireshark, we can see that all the expected RTP events are present on the wire but FreeSWITCH seems to have difficulty in processing them all consistently. I've tried setting the following manual RTP bugs (bit of guesswork here I'm afraid, these may or may not be relevant): IGNORE_DTMF_DURATION ACCEPT_ANY_PACKETS but unfortunately neither made any difference. Are there any other settings we could try which might help? Or is this a known problem with some other workaround? Regards, Chris Chris Young Software Engineer [cid:image05853e.PNG@376d53fe.46b309f7] t: +44 118 943 9249 e: Chris.Young(a)enghouse.com<mailto:Chris.Young@enghouse.com> w: www.enghouseinteractive.co.uk<http://www.enghouseinteractive.co.uk> [cid:image3d627d.PNG@c805b5df.4cb095da]

3 3

Server having two sangoma cards and can't handle full capacity
by Shaik Bawajan 22 Jul '14

22 Jul '14

We are having two (Sangoma) Voice Cards, one with 8-ports and other with 16-ports. We use Freeswitch with ESL library to connect and process the Outbound calls. The issue what we are facing is when we use all 690 channels, we see segmentation_fault in Freeswitch and big dump file is written (ie., > 10gb) and which would crash the system, because of high CPU\IO. But if we reduce the channels say, 450 we don't see any issue at all. What could be the reason for this issue. Appreciate any Help...

3 2

(no subject)
by Nick Zaitsev 22 Jul '14

22 Jul '14

Good day to you. I set the freeswitch to work with skinny endpoints,i can make calls from sip devic3e to skinny device but i can not vice versa. I can only see pushing digits in fs debug?but no call processing. Could you help please? -- Nick Zaitsev

2 1

Conference auto outcall questions
by Chris Martineau 22 Jul '14

22 Jul '14

Hi, Just setting up conferencing to use auto outcall. Q1. Is it possible to not hear anything from the called user until connected and then just hear the entry sound? Currently if I outcall to more than 1 person I get a mess of ringing until each person answers. Q2. If you auto outcall and it goes into voicemail is there anyway to kick the offending conference leg? This is especially a pain with mobiles as invariably the only way to stop the voicemail from just sitting there recording your conference is to hang up? Thanks Chris

2 4