[Red5] Pause delay due Jetty buffer (or something like this)
Mamontov Sergey V.
moonlight at ua.elro.com
Mon Dec 17 08:17:49 PST 2007
Hello everyone :)
I'm just wondering, is there any possible solution for our problem? :)
We have to leave this feature for a time, and project will be launched
in beta mode :)
But I think solution will be founded not very soon :) All Red5 team
have much more important task on RTMPE, Edges and other great features
Regard Red5 team!!! Your make very important job in direction of free
source development in world where everybody is hold they hard work
sources in secret :)
> Today we try our test on FMS, he have same problem...
> Maybe problem inside of RTMP itself, because even developer of this
> protocol have same problem...
> Regards for Red5 team :)
> Currently solution for this problem will be disconnecting user from
> server at pause and then reconnect him back ( not really good
> solution, but only one )
> But if it will be somehow fixed, it will be really COOL! :)
>> Anybody, say something, please :) This is important question for us...
>> Is this bug, can negative value of pending messages can be reason of
>> picture what we can see in first video?
>> Is this fact confirmed from developers side about negative pending
>> Yesterday I've tried to find where the problem can be, but no success
>>> Tell something ;)
>>> What do you think about founded information? It can be reason what
>>> server keep sending packets even if we hi pause button?
>>>> Sergey & Outfoxer,
>>>> Can you try to tune the "underrunTrigger" property of
>>>> "playlistSubscriberStream" bean to see if it helps? The bean is
>>>> defined in red5-common.xml. The property is used as a threshold for
>>>> sending more data to the client. When the number of packets pending
>>>> inside MINA exceed this value, a bandwidth insufficient notification
>>>> will be sent to the client and the server will stop sending more
>>>> data to be piled up in MINA. I recommend you to try a smaller value
>>>> like 2 or 3 to see if it helps with the delay issue.
>>>> On Dec 5, 2007 1:25 AM, <outfoxer at gmail.com> wrote:
>>>> Hello, Steven.
>>>>> Hi Sergey,
>>>>> Thanks for the video. Yes, I guess that should be the buffering
>>>>> issue on server side, specifically the sending buffer of MINA. I
>>>>> will try to see if there are any options in MINA to invalidate all pending packets.
>>>> I tried to set the SendBufferSize for MINA for extra low values =
>>>> 32 or 64 in RTMPMinaTransport class directly (recompile the server)
>>>> or by means of option mrtmp.send_buffer_size in red5.properties
>>>> but it doesn't help, the problem remains :'(
>>>> Red5 mailing list
>>>> Red5 at osflash.org
>>> Red5 mailing list
>>> Red5 at osflash.org
>> Red5 mailing list
>> Red5 at osflash.org
> Red5 mailing list
> Red5 at osflash.org
More information about the Red5