January 2013 - freeswitch-dev@lists.freeswitch.org

Re: [Freeswitch-dev] transfer_history performance
by David Brazier 18 Jan '13

18 Jan '13

François Thanks for the suggestion, will try re-working with execute_extension. David From: François Delawarde Sent: 17/01/2013 11:26 To: freeswitch-dev(a)lists.freeswitch.org Subject: Re: [Freeswitch-dev] transfer_history performance Hi David, The "transfer" app is quite heavy to be chained 1000s of times as it does many things to keep track of what happened (handle max-fw, dup caller profile, history var, ...). You might end-up with other issues trying to scale, maybe disk bottleneck writing huge xml-cdr files, ... What about making your application use "execute_extension" instead? Note that I still vote yes for your optimization ideas! François. On Thu, 2013-01-17 at 10:48 +0000, David Brazier wrote: > Hi again > > A bit more information on reproducing it: each call using > test_transfer will fully occupy one CPU, so on a 4 core machine you > need to start 4 of these calls to tie it up, then audio on other calls > is broken or completely stopped, CLI stops responding etc. As a > workaround I have commented out from switch_ivr.c line 1808 near end > of switch_ivr_session_transfer: > > /*switch_channel_add_variable_var_check(channel, > SWITCH_TRANSFER_HISTORY_VARIABLE, new_profile->transfer_source, > SWITCH_FALSE, SWITCH_STACK_PUSH);*/ > > Of course that means transfer_history is no longer populated, but it > does remove the performance problem, the test_transfer calls now run > through so fast I had to start the loop off at 9999 and although they > do use a bit of CPU (it is a tight loop with no sleep/IO etc after > all) 4 simultaneously do not tie up the server and other calls > continue fine. I can think of a few more complete solutions: > 1. Improve the performance of appending to variables with big > values - I think this is ultimately done in > switch_event_base_add_header in switch_event.c, though that > looks like quite a complex function. In the end, repeated > appending to long strings is never going to be efficient. > 2. Maintain the transfer history as a stack/list/array structure > and either render it to the variable string whenever it is > changed (still expensive?) or generate the string only on > demand - though I'm not sure how that would work with info, > uuid_dump and verbose events in event sockets etc. > 3. Allow channels to opt out of having transfer_history tracked, > perhaps by setting it to "suppress" or "false", this could be > checked before the line above. Would put the onus on the > dialplan to know if it was going to do lots of transfers, but > it would be backward-compatible with anyone who needs this > variable. Out of interest why was it added? > Once I have managed to log in to Jira I will raise all this there. > > Cheers > > David > > > On Thu, Jan 17, 2013 at 12:57 AM, David Brazier > <davidjbrazier(a)gmail.com> wrote: > Hi > > Our application makes use of long-running calls with a lot of > diaplan transfers to bridge them to other calls, put them on > hold etc. We've noticed a slow down in performance (high CPU, > audio dropped, delayed) after several hours. I can reproduce > it with some dialplan that does 1000 transfers: > > <extension name="test_transfer"> > <condition field="destination_number" > expression="^test_transfer$"> > <action application="set" data="max_forwards=9999"/> > <action application="transfer" data="test_transfer_999"/> > </condition> > </extension> > <extension name="test_transfer_0" continue="false"> > <condition field="destination_number" > expression="^test_transfer_0$"> > <action application="hangup"/> > </condition> > </extension> > <extension name="test_transfer_n"> > <condition field="destination_number" > expression="^test_transfer_(\d+)$"> >  > <action application="transfer" data="test_transfer_${expr > $1-1}"/> > </condition> > </extension> > > and then > > originate user/1000 test_transfer XML blah > > On a Windows laptop (git head) it slows down to more than 1 > sec per iteration after it gets past 800, on a bigger Linux > server (FS 1.2.5.1) it gets a bit futher but not much. On a > server with an old version (git-1086cba 2011-05-23 22-51-43 > -0500) it zips through to the end with no slow down. If you > put the info call in you can see that > variable_transfer_history is growing, and I think this is the > source of the problem - this variable is missing from the > older version but the newer ones are apparently appending to > it on each iteration which becomes a very expensive > operation. > > I'll raise a Jira. Any suggestion for a workaround? (Apart > from rewriting our dialplan!) > > David > > > > > > > _________________________________________________________________________ > Professional FreeSWITCH Consulting Services: > consulting(a)freeswitch.org > http://www.freeswitchsolutions.com > > FreeSWITCH-powered IP PBX: The CudaTel Communication Server > http://www.cudatel.com > > Official FreeSWITCH Sites > http://www.freeswitch.org > http://wiki.freeswitch.org > http://www.cluecon.com > > 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 _________________________________________________________________________ Professional FreeSWITCH Consulting Services: consulting(a)freeswitch.org http://www.freeswitchsolutions.com FreeSWITCH-powered IP PBX: The CudaTel Communication Server http://www.cudatel.com Official FreeSWITCH Sites http://www.freeswitch.org http://wiki.freeswitch.org http://www.cluecon.com 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 1

[Freeswitch-dev] deliver_vm function in mod_voicemail.c does not pass all user's variables and parameters
by Vik Killa 17 Jan '13

17 Jan '13

I'm experimenting with implementing a voicemail call page in mod_voicemail.c I have some semi-working test code, I'm stuck though, on the fact that the deliver_vm function does not get passed all of the mailbox user's domain variables and parameters. It only receives the specific user's variables and parameters. In my setup (and probably a lot of other setups) in order for a call to originate out a carrier you probably need some necessary directory variables (accountcode, default_areacode, etc...) I've been trying to hack a way to pass all the domain variables & parameters to deliver_vm (specifically in the voicemail_inject function) but I've gotten no where. I can see the parameters are passed using "switch_event_t *my_params" which apparently retrieves the data using 'switch_event_create(&my_params, SWITCH_EVENT_REQUEST_PARAMS);" I have no clue how this function works or if there is a way for SWITCH_EVENT_REQUEST_PARAMS to return ALL the user's (including the domain's) variables and parameters. Please help me so I can improve FreeSWITCH for everyone but adding this nice feature. Any tips or assistance would be greatly appreciated. Thank you again. Emmanuel

1 0

[Freeswitch-dev] transfer_history performance
by David Brazier 17 Jan '13

17 Jan '13

Hi Our application makes use of long-running calls with a lot of diaplan transfers to bridge them to other calls, put them on hold etc. We've noticed a slow down in performance (high CPU, audio dropped, delayed) after several hours. I can reproduce it with some dialplan that does 1000 transfers: <extension name="test_transfer"> <condition field="destination_number" expression="^test_transfer$"> <action application="set" data="max_forwards=9999"/> <action application="transfer" data="test_transfer_999"/> </condition> </extension> <extension name="test_transfer_0" continue="false"> <condition field="destination_number" expression="^test_transfer_0$"> <action application="hangup"/> </condition> </extension> <extension name="test_transfer_n"> <condition field="destination_number" expression="^test_transfer_(\d+)$">  <action application="transfer" data="test_transfer_${expr $1-1}"/> </condition> </extension> and then originate user/1000 test_transfer XML blah On a Windows laptop (git head) it slows down to more than 1 sec per iteration after it gets past 800, on a bigger Linux server (FS 1.2.5.1) it gets a bit futher but not much. On a server with an old version (git-1086cba 2011-05-23 22-51-43 -0500) it zips through to the end with no slow down. If you put the info call in you can see that variable_transfer_history is growing, and I think this is the source of the problem - this variable is missing from the older version but the newer ones are apparently appending to it on each iteration which becomes a very expensive operation. I'll raise a Jira. Any suggestion for a workaround? (Apart from rewriting our dialplan!) David

2 2

[Freeswitch-dev] Jira Login
by David Brazier 17 Jan '13

17 Jan '13

Very sorry, but I have forgotten my Jira password (username brazierd) and when I try to recover it, the email never arrives, not in spam/junk as far as I can see. I wanted to report the transfer_history problem. Could a kind admin reset it? Thanks David

1 0

[Freeswitch-dev] limit memory usage of freeswitch
by rentmycoder rentmycoder 17 Jan '13

17 Jan '13

The question is how many concurrent calls are you actually doing? I'm using a test system, inteconnect two FS boxes via sip, and start 50 concurent calls... Average call duration is 45 sec, a random number between 30-60sec... The dialplan does not do nothing special, just a streamfile action... After a night FS 1.2.5.3 eats up all the 900Mb ram of the system... The same test using Asterisk 1.8 eats up 50Mb... Is this normal? This means I cant use FS on a small virtual mashine vith 1GB ram together with other services on the same VM for a small company handling 30 concurrent calls... This is normal, or a memeory leak, or I make some mistake during testing?

3 2

[Freeswitch-dev] limit memory usage of freeswitch
by rentmycoder rentmycoder 16 Jan '13

16 Jan '13

*Hi All,****According to my experineces and some info found on FS lists it is the**normal behaviour that freeswitch eats up all of the available memory...**It's not a memory leak, but the standard behaviour of FS by design...**Correct me if I'm wrong...* *Is there any way to limit FS memory usage???*** *I need another services to run on the FS box...* * * *According to the answer on the users list, there is no built in way to limit the usage...* *I'm going to add this feature request to fs and add as an option to freeswitch.conf.xml* *Maybe I have to just modify the function where FS asks for available system memory...*** *** Can anyone advice me a direction, how to code it? **Thanks, **John*

4 3

[Freeswitch-dev] FreeSWITCH Community Conference Call Today
by Michael Collins 16 Jan '13

16 Jan '13

Hello all! Today's conference call agenda is here: http://wiki.freeswitch.org/wiki/FS_weekly_2013_01_16 We have updates on FreeSWITCH 1.2.6, the new FreeSWITCH book, and ClueCon 2012 videos. We are also going to discuss some techniques for how to validate bugs and update Jira tickets with the information. Also, if anyone is familiar with git bisect we would like to invite you to spend a few minutes giving us some pointers as that is a very useful tool for locating the commit that "broke something." Talk to you shortly! -- Michael S Collins Twitter: @mercutioviz http://www.FreeSWITCH.org http://www.ClueCon.com http://www.OSTAG.org

1 0

[Freeswitch-dev] FreeSWITCH Weekly News and Notes
by Michael Collins 14 Jan '13

14 Jan '13

Greetings! It has been a light week when it comes to news items. However, the FreeSWITCH team has been working hard on bug fixes and various projects. More information on those will come out once the details are available. Last week we had a community discussion on the weekly conference call<http://wiki.freeswitch.org/wiki/FS_weekly_2013_01_09>, mostly discussing the ways that we can all help out with getting Jira's handled. Although we did not formally adopt the "Jira karma" idea we decided that we (as a project and community) would do our best to reward those who help out with bug marshaling, verifying, documenting, etc. Every little bit helps. To that end this week<http://wiki.freeswitch.org/wiki/FS_weekly_2013_01_16>we will be spending some time as a group going over a few selected bugs and as a group trying our best to figure what's going on. Ken and Brian will supply us with a few bugs to tackle. Oh, and if you use Cisco phones, please be very careful<http://arstechnica.com/security/2013/01/hack-turns-the-cisco-phone-on-your-…> ! -- Michael S Collins Twitter: @mercutioviz http://www.FreeSWITCH.org http://www.ClueCon.com http://www.OSTAG.org

1 0

[Freeswitch-dev] Github tags
by Tom Parrott 14 Jan '13

14 Jan '13

Hi, Is it possible to get the tag branches pushed to the github mirror as well? https://github.com/FreeSWITCH/FreeSWITCH Thanks Tom

1 0

[Freeswitch-dev] Ability to cancel leg_timeout when using group_confirm from b-leg (fund)
by Tom Parrott 14 Jan '13

14 Jan '13

Hi, The company I work for would like to sponsor the development of a feature to be able to cancel the leg_timeout during a bridge from a lua script that is run on the b-leg as part of the group_confirm functions. I have opened an improvement request here: http://jira.freeswitch.org/browse/FS-4955 The reason we would like this feature is so that we can provide a "press 1 to accept" feature to our callees, and when they press 1 proceed to play a series of call announcements to them. The caller should continue to hear ring tones until the announcements are finished, however the leg_timeout should be cancelled from the lua script, maybe by setting a channel variable, e.g. cancel_timeout = 1 Thanks Tom Parrott

1 2