[Freeswitch-users] Choppy VM when using loopback (on Debian lenny)

Tue Feb 1 23:11:00 MSK 2011

Yes, using 1f1541b our calls that use mod_loopback into mod_conference
are unusable. The audio is choppy and delay increases as time passes.

I thought it was an issue relating to the default 100Hz kernel on EC2
so I spent some time yesterday putting together a 1000Hz kernel but it
didn't make any difference. In the meantime I've compiled mod_loopback
from 4c5426f and loaded it with my 1f1541b build which eliminates the
issue.

I haven't added it to Jira yet as I want to spend some time debugging
it (and I also owe you some debug info for FS-2934) but the problem is
definitely there.

On Tue, Feb 1, 2011 at 3:58 PM, Anthony Minessale
<anthony.minessale at gmail.com> wrote:
> Are you saying you have better results on that version than you do on
> the latest?
> What conditions do you have that cause you trouble, what is the
> endpoint on the other side.
>
> If the last commit to mod_loopback intended to improve audio quality
> actually makes it worse I need to investigate it.
>
>
> On Tue, Feb 1, 2011 at 5:53 AM, Dan Lane <null at invalid.name> wrote:
>> FWIW we've been also been having audio issues with loopback recently
>> on EC2 (with a 1000Hz kernel).
>>
>> We worked around it in the short term by reverting mod_loopback to
>> git-4c5426f during the build process.
>>
>> For anyone else who wants to try this just run "git checkout 4c5426f"
>> in src/mod/endpoints/mod_loopback then build as usual. This is NOT a
>> long term solution though.
>>
>> On Mon, Jan 31, 2011 at 8:02 AM, Øyvind Albrigtsen <oa at estation.dk> wrote:
>>> Thanks for all your feedback. I'll keep on trying and inform you what
>>> worked for me.
>>>
>>>
>>> Regards
>>> Oyvind
>>>
>>> On 2011-01-29 21:48, Anthony Minessale wrote:
>>>> Everyone should try latest GIT before pondering any further because I
>>>> added a patch like 2 days ago to adress this issue.
>>>>
>>>>
>>>> On Sat, Jan 29, 2011 at 2:12 PM, Frank Park<frank at telonium.com>  wrote:
>>>>
>>>>> Yeah. I, too, don't see the correlation between the NTP and hw timer..
>>>>> I am not familiar with the timer_test command and what it's measuring, but
>>>>> of the 50 tests it ran, min is 19089 and max is 20713.
>>>>> Frank
>>>>>
>>>>>
>>>>> On Fri, Jan 28, 2011 at 2:33 PM, David Ponzone<david.ponzone at ipeva.fr>
>>>>> wrote:
>>>>>
>>>>>> Frank,
>>>>>> I fail to see the relationship between the hw timer and NTP.
>>>>>> Can you please elaborate ?
>>>>>> David Ponzone  Direction Technique
>>>>>> email: david.ponzone at ipeva.fr
>>>>>> tel:      01 74 03 18 97
>>>>>> gsm:   06 66 98 76 34
>>>>>> Service Client IPeva
>>>>>> tel:      0811 46 26 26
>>>>>> www.ipeva.fr  -   www.ipeva-studio.com
>>>>>> Ce message et toutes les pièces jointes sont confidentiels et établis à
>>>>>> l'intention exclusive de ses destinataires. Toute utilisation ou diffusion
>>>>>> non autorisée est interdite. Tout message électronique est susceptible
>>>>>> d'altération. IPeva décline toute responsabilité au titre de ce message s'il
>>>>>> a été altéré, déformé ou falsifié. Si vous n'êtes pas destinataire de ce
>>>>>> message, merci de le détruire immédiatement et d'avertir l'expéditeur.
>>>>>>
>>>>>>
>>>>>>
>>>>>> Le 28/01/2011 à 19:49, Frank Carmickle a écrit :
>>>>>>
>>>>>> Hi
>>>>>>
>>>>>> On Jan 28, 2011, at 7:39 AM, Øyvind Albrigtsen wrote:
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> I'm using latest git-version of Freeswitch, and when I go to voicemail
>>>>>>
>>>>>> when calling a number the sound playback is choppy and it skips some of
>>>>>>
>>>>>> the digits in the number I called.
>>>>>>
>>>>>> What kind of results do you get from timer_test at the fs_cli?  Are you
>>>>>> running on hardware or are you virtualized?  What is your clock source set
>>>>>> to and what are your available clock source options?  See
>>>>>> /sys/devices/system/clocksource/clocksource0/available_clocksource and
>>>>>> /sys/devices/system/clocksource/clocksource0/current_clocksource.  I am
>>>>>> running virtualized with the 2.6.26-2-xen-amd64 and I can get timer_test to
>>>>>> hang at 19998/19999 which works very well for me.  When I was having problem
>>>>>> it was reporting numbers all over the map from 17400 to 22600 with lots of
>>>>>> randomness in between.  I have my clocksource set to jiffies and xen
>>>>>> independent wallclock set to 1.  Of course at that point you need to have
>>>>>> ntp running against a bunch of servers to drive your clock nice and steady.
>>>>>>   I know my set up is probably a lot different than yours but I thought I'd
>>>>>> toss it out there to show that some of the harshest conditions can be dealt
>>>>>> with and don't give up trying.  If you are running on hardware with a cpu
>>>>>> that doesn't have constant_tsc then you might have some problems.  Just play
>>>>>> with the different timer options until you find the one that works.
>>>>>>
>>>>>> HTH
>>>>>> --FC
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> FreeSWITCH-users mailing list
>>>>>> FreeSWITCH-users at 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 at lists.freeswitch.org
>>>>>> http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
>>>>>> UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
>>>>>> http://www.freeswitch.org
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>>
>>>>> ----=======================----
>>>>> Frank Park
>>>>> Telonium Communications, LLC
>>>>> frank at telonium.com
>>>>> http://www.telonium.com
>>>>> Follow Us on Twitter: @GetTelonium
>>>>> 404-566-8888 x1001 Office
>>>>> 404-939-4242 Cell
>>>>> ----=======================----
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> FreeSWITCH-users mailing list
>>>>> FreeSWITCH-users at 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 at 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 at lists.freeswitch.org
>> http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
>> UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
>> http://www.freeswitch.org
>>
>
>
>
> --
> Anthony Minessale II
>
> FreeSWITCH http://www.freeswitch.org/
> ClueCon http://www.cluecon.com/
> Twitter: http://twitter.com/FreeSWITCH_wire
>
> AIM: anthm
> MSN:anthony_minessale at hotmail.com
> GTALK/JABBER/PAYPAL:anthony.minessale at gmail.com
> IRC: irc.freenode.net #freeswitch
>
> FreeSWITCH Developer Conference
> sip:888 at conference.freeswitch.org
> googletalk:conf+888 at conference.freeswitch.org
> pstn:+19193869900
>
> _______________________________________________
> FreeSWITCH-users mailing list
> FreeSWITCH-users at lists.freeswitch.org
> http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
> UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
> http://www.freeswitch.org
>