- FreeSWITCH-dev - FreeSWITCH Mailing Lists

ONE WEEK LEFT - Official FreeSWITCH Training in New York City
by Darren Schreiber 06 Oct '10

06 Oct '10

Hi folks, The Official FreeSWITCH Training course in New York is just one week away and is almost full. If you are interested in learning a ton of FreeSWITCH, taught by one of the authors of the FreeSWITCH Book, you should consider signing up. Code FREEVOIP is still valid for $400 off the training. You can register at http://www.voipkb.com/ . There is also a huge discount for registering two or more people together. The course is an in-depth dive into FreeSWITCH with completely different material then is in the book. We will cover: * Understanding configuration files and the default configuration * Call authentication and routing basics * Integration modules (mod_skypiax, mod_dingaling for Skype/GTalk/XMPP integration) * Understanding presence * Load balancing and high availability * FreeSWITCH Internals * How to debug and troubleshoot FreeSWITCH * Building Custom C Modules * Advanced Modules Please let me know if you have any questions. We're excited to continue offering this course and thank the FreeSWITCH team for their continued outstanding work. See you in New York! Thanks, Darren Schreiber Join us in October for FreeSWITCH Training! Visit www.voipkb.com<http://www.voipkb.com/> for more information

1 0

ONE WEEK LEFT - Official FreeSWITCH Training in New York City
by Darren Schreiber 06 Oct '10

06 Oct '10

Hi folks, The Official FreeSWITCH Training course in New York is just one week away and is almost full. If you are interested in learning a ton of FreeSWITCH, taught by one of the authors of the FreeSWITCH Book, you should consider signing up. Code FREEVOIP is still valid for $400 off the training. You can register at http://www.voipkb.com/ . There is also a huge discount for registering two or more people together. The course is an in-depth dive into FreeSWITCH with completely different material then is in the book. We will cover: * Understanding configuration files and the default configuration * Call authentication and routing basics * Integration modules (mod_skypiax, mod_dingaling for Skype/GTalk/XMPP integration) * Understanding presence * Load balancing and high availability * FreeSWITCH Internals * How to debug and troubleshoot FreeSWITCH * Building Custom C Modules * Advanced Modules Please let me know if you have any questions. We're excited to continue offering this course and thank the FreeSWITCH team for their continued outstanding work. See you in New York! Thanks, Darren Schreiber Join us in October for FreeSWITCH Training! Visit www.voipkb.com<http://www.voipkb.com/> for more information

1 0

FreeSWITCH Conference Call Today
by Michael Collins 06 Oct '10

06 Oct '10

Hello all! Our agenda for today is here: http://wiki.freeswitch.org/wiki/FS_weekly_2010_10_06 We have a few things to discuss, such as some additions to the default configs and some Jira tasks that we could use some help with. Also, we are hoping that Darren Schreiber will be calling in to give us an update from ITExpo down in L.A. Talk to you soon, Michael

1 0

responce code 183 related functionality
by Bambr 05 Oct '10

05 Oct '10

As far as I know, when, during the calling, FS recieves 183 responce, call is considered succesful, it is marked in cdr and so on. While, there's possibility for this call to be failed, as after ringing the call might be declined. Asking in IRC on #freeswitch got me answer like "it's a question of conventions" - and it is right, but our convention (dictated by users) is that such call can not be marked as successful "a priori" - only when it really succeeded. I tried looking at sources of FS finding the place where it is judged, found something like it in switch_ivr_originate.c but i'm not sure it is the right place and tests i did - they didn't get me needed result. So here is my question: is it possible to make configurable behavior upon recieving 183? I'm aware about ignoring all early media, but this is not an option. I'm interested in managing status of call on recieving 183. -- Bambr in da house

5 13

sipX consultative trasnfer to freeswitch Conference leaves a phantom channel
by Joegen E. Baclor 30 Sep '10

30 Sep '10

Hi freeswitchers, I was recently tasked to investigate an apparent bug in sipX IVR subsystem that involves consultative transfer. The bug can be replicated as follows: 1. Extensions 5001 and 5002 are registered to sipX. 2. Extension 65001 is configured as a conference room in freeswitch. 3. Extension 5001 calls 5002 and puts 5002 on hold. 4. Extension 5001 calls 65001 and puts 65001 on hold. 5. Extention 5001 sends REFER with replaces to 5002 towards 65001. 6. Extension 5002 sends INVITE with replaces to 65001 At this point, signaling will be processed as normal. Notifications from 5002 completing the transfer looks OK and BYE was sent by 5001 to tear down the original call to which freeswitch politely responded with 200 Ok. Audio between 5002 and and 65001 conference is bidirectional. However, looking at the status of the conference bridge, there are still two participants active. The old channel that 5001 opened is still there. This would eventually timeout after a period of time probably because of media inactivity. If you think this is due to the mishandling or misbehavior of the event socket implemention of the IVR. Eg: It should manually destroy the channel after receiving a "channel replaced" event from freeswitch, I would appreciate pointers how it should be done. If you think this is a bug in freeswitch I am willing to get it patched if someone already familiar with the code point me to the right places to look at. (I am new to freeswitch but I'm not new to C nor SIP). Thanks in advance for any help. Joegen

4 8

FreeSWITCH Conference Call Today
by Michael Collins 29 Sep '10

29 Sep '10

Hey all, the FreeSWITCH conf call agenda is here: http://wiki.freeswitch.org/wiki/FS_weekly_2010_09_29 It is a light agenda today but I have several things that I want to discuss with the community and I do have a few items that I'd like to get documented and I would appreciate some input from the gang. See you at 1 Eastern/10 Pacific -MC

1 0

Re: [Freeswitch-dev] [Freeswitch-users] no audio in loopback mode
by Bernhard Suttner 29 Sep '10

29 Sep '10

Hi, I found out what happend. It's a bug within mod_loopback if loopback_bowout is set to false. If it is set to false, it will still not write/read the media because there is a missing check. The attached patch does fix the issue. Thanks a lot for your help and for all the great freeswitch stuff! Best regards, Bernhard -----Ursprüngliche Nachricht----- Von: freeswitch-users-bounces(a)lists.freeswitch.org [mailto:freeswitch-users-bounces@lists.freeswitch.org] Im Auftrag von Bernhard Suttner Gesendet: Mittwoch, 29. September 2010 10:20 An: 'FreeSWITCH Users Help' Betreff: Re: [Freeswitch-users] no audio in loopback mode Hi, I have attached the log to pastebin 14054. Any hint would be appreciated. Thanks. Best regards, Bernhard -----Ursprüngliche Nachricht----- Von: freeswitch-users-bounces(a)lists.freeswitch.org [mailto:freeswitch-users-bounces@lists.freeswitch.org] Im Auftrag von Bernhard Suttner Gesendet: Dienstag, 28. September 2010 20:33 An: FreeSWITCH Users Help Betreff: Re: [Freeswitch-users] no audio in loopback mode Hi Brian, Ok. I will create them tomorrow if I could not find a solution. I thought that I maybe miss something or that there is a good work-around. Best regards, Bernhard ----- Original Message ----- From: Brian West [mailto:brian@freeswitch.org] To: FreeSWITCH Users Help [mailto:freeswitch-users@lists.freeswitch.org] Sent: Tue, 28 Sep 2010 20:23:36 +0200 Subject: Re: [Freeswitch-users] no audio in loopback mode > Logs on pastebin? with sip trace? > > /b > > On Sep 28, 2010, at 1:12 PM, Bernhard Suttner wrote: > > > Hi, > > > > I forget something to mention: If I set the loopback_bowout to "true" than > I have media in both directions. > > > > Best regards, > > Bernhard > > > _______________________________________________ > FreeSWITCH-users mailing list > FreeSWITCH-users(a)lists.freeswitch.org > http://lists.freeswitch.org/mailman/listinfo/freeswitch-users > UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users > http://www.freeswitch.org > _______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users(a)lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org _______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users(a)lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org

1 0

Re: [Freeswitch-dev] luasql or freeswitch.dbh
by Bernhard Suttner 28 Sep '10

28 Sep '10

Hi, Great! Thanks a lot for that test. Now I am pretty sure that I will use the freeswitch.dbh technology. Thx!!! We should find a good page to store this performance test within the wiki. Best regards, Bernhard ----- Original Message ----- From: Leon de Rooij [mailto:leon@scarlet-internet.nl] To: freeswitch-dev(a)lists.freeswitch.org Sent: Tue, 28 Sep 2010 20:39:47 +0200 Subject: Re: [Freeswitch-dev] luasql or freeswitch.dbh > Good evening, > > I got the LuaSQL postgres driver working and ran the same example: > > require "luasql.postgres"; > > local env= assert(luasql.postgres()) > local con= assert(env:connect("db", "user","pass","172.31.0.5")) > local cur= assert(con:execute ("select 1 as foo, 2 as bar")) > > local row= {} > while (cur:fetch(row,"a")) do > for key, valin pairs(row) do > stream:write(string.format("%25s : %s\n", key, val)) > end > stream:write("\n") > end > > cur:close() > con:close() > env:close() > > And did the same test: > > freeswitch@fs-dev> hammer 1000 lua test_luasql_postgres.lua > +OK Finished in 40625 ms (min 30 ms, max 122 ms) > > So it's slowest of the three ! :-) - I didn't investigate any further > about the cause of this. > > I ran it a couple of times with the same result, and it was done on the > same machine. > > (for completeness sake: all tests were done on an Intel Q6600 with two > KVM guests, one containing an Ubuntu 10.04/64 with Postgres 8.4 and one > guest with CentOS 5.5/64 with FS Trunk) > > -- To answer your other question about whether it blocks: > > The FreeSWITCH cache db behaves in a way that it makes a new connection > if there are no free handles in the pool. It will only block if your db > server doesn't allow a new connection. In c, you should normally release > your db handle manually, but Lua has garbage collection, that > automatically triggers release of any db handles to the pool when they > go out of scope (for example when your script terminates). If a free > handle is available in the pool it will be re-used by another thread. > > If I understand it right, a free handle in the pool will be disconnected > from the db server if it is not used for 120 seconds. > > Kind regards, > > Leon > > > > > Op 9/28/2010 4:10 PM, Leon de Rooij schreef: > > Hi Bernhard, > > > > I tried out several ways for connecting to a database from inside > > FreeSWITCH myself before writing the freeswitch.Dbh (which is rather > > simple actually - it only exposes already existing FreeSWITCH > > functionality in mod_lua). > > > > My findings were that caching the connection as is provided by > > FreeSWITCH gives a huge speed increase compared to re-connecting all > > the time as must be done with LuaSQL. Perhaps there are ways to have a > > connection pool in ODBC but I didn't investigate that. > > > > > > Here's a test with LuaSQL with ODBC driver: > > require "luasql.odbc"; > > > > local env= assert(luasql.odbc()) > > local con= assert(env:connect("dsn","user","pass")) > > local cur= assert(con:execute ("select 1 as foo, 2 as bar")) > > > > local row= {} > > while (cur:fetch(row,"a")) do > > for key, valin pairs(row) do > > stream:write(string.format("%25s : %s\n", key, val)) > > end > > stream:write("\n") > > end > > > > cur:close() > > con:close() > > env:close() > > freeswitch@fs-dev> hammer 1000 lua test_luasql_odbc.lua > > +OK Finished in 38952 ms (min 29 ms, max 107 ms) > > > > > > The same test with freeswitch.Dbh: > > > > local dbh= assert(freeswitch.Dbh("dsn","user","pass")) > > > > assert(dbh:query("select 1 as foo, 2 as bar",function(row) > > for key, valin pairs(row) do > > stream:write(string.format("%25s : %s\n", key, val)) > > end > > stream:write("\n") > > end)) > > > > assert(dbh:release()) > > > > freeswitch@fs-dev> hammer 1000 lua test_lua_dbh.lua > > +OK Finished in 2723 ms (min 1 ms, max 25 ms) > > > > As you can see, with these tests the freeswitch.Dbh is about 14 times > > as fast as luasql.odbc. Of course this test is single threaded (the > > scripts are executed serially), so you'd get different results in > > real-world, and if you have 'long' running scripts with many queries > > on the same handle, the results would even out a bit. > > > > (btw mod_hammer is a simple app in contrib/ledr/c that executes > > another app n times and shows how long it took) > > > > I just tried compiling LuaSQL with Postgres driver to do the same > > tests, but that borked. Maybe you can give results of that ? I am > > curious how it behaves. > > > > About having a native driver or odbc: odbc is just an abstraction > > layer which uses the native drivers itself, so I'd reckon it should be > > almost as fast - correct me if I'm wrong ? > > > > I wrote some more about these tests on my blog at http://www.toyos.nl > > > > Kind regards, > > > > Leon > > > > > > On Sep 16, 2010, at 8:52 PM, Bernhard Suttner wrote: > > > >> Hi, > >> > >> from performance point of view, what should be used to connect to > >> many different databases, luasql or freeswitch.dbh which does use ODBC? > >> > >> The problem is, that is not only one database but 3-5 databases which > >> are on 2-3 different hosts. Lua will be used to fetch the directory > >> and dialplan directly from the db. Also live-data will be stored on > >> the different databases. The advantage of freeswitch.dbh would be, > >> that the db connections will be re-used. LUAsql would use the native > >> e.g. postgres or mysql driver. > >> > >> Thanks for any hint. > >> > >> Best Regards, > >> Bernhard > >> > >> _______________________________________________ > >> FreeSWITCH-dev mailing list > >> FreeSWITCH-dev(a)lists.freeswitch.org > >> <mailto:FreeSWITCH-dev@lists.freeswitch.org> > >> http://lists.freeswitch.org/mailman/listinfo/freeswitch-dev > >> UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-dev > >> http://www.freeswitch.org > > > > > > _______________________________________________ > > FreeSWITCH-dev mailing list > > FreeSWITCH-dev(a)lists.freeswitch.org > > http://lists.freeswitch.org/mailman/listinfo/freeswitch-dev > > UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-dev > > http://www.freeswitch.org > >

1 0

luasql or freeswitch.dbh
by Bernhard Suttner 28 Sep '10

28 Sep '10

Hi, from performance point of view, what should be used to connect to many different databases, luasql or freeswitch.dbh which does use ODBC? The problem is, that is not only one database but 3-5 databases which are on 2-3 different hosts. Lua will be used to fetch the directory and dialplan directly from the db. Also live-data will be stored on the different databases. The advantage of freeswitch.dbh would be, that the db connections will be re-used. LUAsql would use the native e.g. postgres or mysql driver. Thanks for any hint. Best Regards, Bernhard

3 4

Re: [Freeswitch-dev] luasql or freeswitch.dbh
by Bernhard Suttner 28 Sep '10

28 Sep '10

Hi, Well, ODBC is a wrapper and therefore there is some time needed. I found something: http://www.devtoolshed.com/content/performance-benchmarks-odbc-vs-oracle-my… I think the native driver should be faster but I have done no further testing. But its not the most important point for me. Much more interesting is, what the freeswitch.dbh does, if there are many threads try to get a connection? Best regards, Bernhard ----- Original Message ----- From: Michael Jerris [mailto:mike@jerris.com] To: freeswitch-dev(a)lists.freeswitch.org Sent: Tue, 28 Sep 2010 18:32:57 +0200 Subject: Re: [Freeswitch-dev] luasql or freeswitch.dbh > > On Sep 28, 2010, at 11:14 AM, Bernhard Suttner wrote: > > > > Currently its not possible to me to create the luasql.postges test to > access the database with the native driver but I am sure, that the postgres > driver should be much faster than the ODBC. Could someone test that? > > > > > Why does everyone seem to think this about ODBC? Is there any basis or just > mass hysteria? > > Mike

1 0