VoIP Mailing List Archives

[woodydickson at gmail.com]

Hi,

I tried to performance test freeswitch with media proxy thur fs.  With 400 cps, I start to see 2000 channels remaining in Freeswitch, and then "no read codec" error starts to pop up.  With only 1875 channels, how come freeswitch is complaining about no read codec?  Also, I am using media_proxy = true, whey should it need a codec anyway?


freeswitch@mycom.com (freeswitch@mycom.com)> 2009-09-14 20:59:16.777675 [ERR] switch_core_io.c:118 sofia/external/12323232@192.168.1.116:5911 has no read codec.
2009-09-14 20:59:30.815547 [ERR] switch_core_io.c:118 sofia/external/12323232@192.168.1.116:5911 has no read codec.
2009-09-14 20:59:30.815547 [ERR] sofia_glue.c:2566 AUDIO RTP REPORTS ERROR: []
2009-09-14 20:59:45.349181 [ERR] switch_core_io.c:118 sofia/external/12323232@192.168.1.116:5911 has no read codec.
2009-09-14 20:59:45.350179 [ERR] sofia_glue.c:2566 AUDIO RTP REPORTS ERROR: []
2009-09-14 21:00:00.104559 [ERR] switch_core_io.c:118 sofia/external/12323232@192.168.1.116:5911 has no read codec.
2009-09-14 21:00:04.495545 [ERR] switch_core_io.c:118 sofia/external/12323232@192.168.1.116:5911 has no read codec.
2009-09-14 21:00:16.996438 [ERR] switch_core_io.c:118 sofia/external/12323232@192.168.1.116:5911 has no read codec.
2009-09-14 21:00:25.905617 [ERR] switch_core_io.c:118 sofia/external/12323232@192.168.1.116:5911 has no read codec.

freeswitch@mycom.com (freeswitch@mycom.com)> show channels count
API CALL [show(channels count)] output:

1875 total.

freeswitch@mycom.com (freeswitch@mycom.com)> 2009-09-14 21:00:36.212767 [ERR] switch_core_io.c:118 sofia/external/12323232@192.168.1.116:5911 has no read codec.


After I paused the traffic from sipp and when sipp finished, I still got a bunch of zombie channels that are in CONSUME_MEDIA stage.  What is the cause of these zombies and how can I fix it?


uuid,direction,created,created_epoch,name,state,cid_name,cid_num,ip_addr,dest,application,application_data,dialplan,context,read_codec,read_rate,write_codec,write_rate,secure
5f013ece-d8a0-4ee0-bce0-3a56c05dc225,outbound,2009-09-14 20:53:55,1252932835,sofia/external/12323232@192.168.1.116:5911,CS_CONSUME_MEDIA,sipp,sipp,192.168.1.116,12323232@192.168.1.116:5911,,,XML,default,PROXY,8000,PROXY,8000,
9347ea12-7ffb-4e96-8ee9-b2ac1ee57752,outbound,2009-09-14 20:53:57,1252932837,sofia/external/12323232@192.168.1.116:5911,CS_CONSUME_MEDIA,sipp,sipp,192.168.1.116,12323232@192.168.1.116:5911,,,XML,default,PROXY,8000,PROXY,8000,
5dc2f6f2-7d58-4582-b680-40bb3ed330ef,outbound,2009-09-14 20:54:02,1252932842,sofia/external/12323232@192.168.1.116:5911,CS_CONSUME_MEDIA,sipp,sipp,192.168.1.116,12323232@192.168.1.116:5911,,,XML,default,PROXY,8000,PROXY,8000,
66361571-b022-42b9-9507-87d1bfc01b03,outbound,2009-09-14 20:54:03,1252932843,sofia/external/12323232@192.168.1.116:5911,CS_CONSUME_MEDIA,sipp,sipp,192.168.1.116,12323232@192.168.1.116:5911,,,XML,default,PROXY,8000,PROXY,8000,
7a075ea5-7c01-4951-b826-cf982df03501,outbound,2009-09-14 20:54:03,1252932843,sofia/external/12323232@192.168.1.116:5911,CS_CONSUME_MEDIA,sipp,sipp,192.168.1.116,12323232@192.168.1.116:5911,,,XML,default,PROXY,8000,PROXY,8000,
bd8c9905-dbab-4ab7-a9af-b7e2b87cc0b0,outbound,2009-09-14 20:54:05,1252932845,sofia/external/12323232@192.168.1.116:5911,CS_CONSUME_MEDIA,sipp,sipp,192.168.1.116,12323232@192.168.1.116:5911,,,XML,default,PROXY,8000,PROXY,8000,
3af84a6c-02f2-44ae-8a4b-1c7940522005,outbound,2009-09-14 20:54:06,1252932846,sofia/external

[anthony.minessale at g...]

[tculjaga at gmail.com]

Hi Woody,

well, it is quite hard to answer you back with this logs...

you didn't tell us:

1. what machine are you running (CPU/RAM)
2. what distro are you running - 32 or 64 bit (i had some lets say "experience" with a wrong selection )
   
3. what is your configuration (dialplan/sip_profiles)
4. did you disable all logging?
   
5. what modules are loaded (you should load minimal modules - at least disable conference)
   
6. if you moved the db files to a ram disk
7. how do you start the calls (slowly with 10 - 20 CPS or you are DDoS-ing it with 400 right away).
   
8. how long do you keep the calls going?
9. what is the the current CPU usage when stresstesting
10. what is the amount of read/writes to from/to your HDD

... i could go on and on with the list.



Here Anthony has nothing to tell you except that you reached the maximum and you are currently killing the machine/application

Also, what are you looking for? ... A machine that can do a lot of simultaneous calls or a machine that can do a lot of CPS? you should decide at some point.


I have a lot of experience with commercial SoftSwitches and i can tell you that FreeSWITCH performance is something outstanding ... i'm getting a reliable 500 CPS on just one FS machine (dualcore xeon 2.33 GHz - bogomips 4670). just to compare: NetCentrex as a comercial SoftSwitch can do only 480 CPS (distribuited on 10 nodes)... and of course signaling only. And this is enough to run my 12.000 calls.


Tihomir.




On Mon, Sep 14, 2009 at 3:08 PM, Woody Dickson <woodydickson@gmail.com (woodydickson@gmail.com)> wrote: