I also have a go at fixing this bug last night but got lost in the stream
code. Basically I think the problem is ConnectionConsumer grabs some
channels, and does not release them. There is even a todo in there about
releasing them. However the event to tell the connection consumer to that
the provider has disconnected doesnt seem to make it to the connection
consumer. I will ask steven for help as he should know.
- Luke
On 1/25/07, Nankun Huang <[EMAIL PROTECTED]> wrote:
oops, I sent the previous email prematured.. damn google mail enter
key.. anyway to continue where I left off...
1, When a stream is deleted by the
deleteStreamById(streamId) function, it did not call the
unreserveStreamId(int streamId) function!! Therefore, it causes the
reservedStreams variable to show that the stream is still reserved
EVEN THOUGH its actually gone from the streams[] Array.
2, I tried to trace who is the one that is called the
deleteStreamById() function and came to the
org.red5.server.stream.StreamService class. From there, I found that
the function which calls the delete stream is the: closeStream()
function on line 59. So I thought maybe, by changing
deleteStreamById(streamId) to unreserveStreamId(streamId) ( since
unreserve stream function calls delete stream anyways ), it'll remove
the offending problem.
However when I did that, it choose to give me a DIFFERENT error
instead!! Here's the error
[ERROR] 233079 SocketAcceptorIoProcessor-0.0:(
org.red5.server.net.rtmp.codec.RTMPProtocolDecoder.error ) Error in
packet at 140
java.lang.NullPointerException
at org.red5.server.net.rtmp.codec.RTMPProtocolDecoder.decodeHeader
(RTMPProtocolDecoder.java:295)
at org.red5.server.net.rtmp.codec.RTMPProtocolDecoder.decodePacket
(RTMPProtocolDecoder.java:217)
at org.red5.server.net.rtmp.codec.RTMPProtocolDecoder.decode(
RTMPProtocolDecoder.java:114)
at org.red5.server.net.rtmp.codec.RTMPProtocolDecoder.decodeBuffer
(RTMPProtocolDecoder.java:87)
at org.red5.server.net.rtmp.codec.RTMPMinaProtocolDecoder.decode(
RTMPMinaProtocolDecoder.java:48)
at
org.apache.mina.filter.codec.ProtocolCodecFilter.messageReceived(
ProtocolCodecFilter.java:157)
at
org.apache.mina.common.support.AbstractIoFilterChain.callNextMessageReceived
(AbstractIoFilterChain.java:501)
at
org.apache.mina.common.support.AbstractIoFilterChain.access$1000(
AbstractIoFilterChain.java:51)
at
org.apache.mina.common.support.AbstractIoFilterChain$EntryImpl$1.messageReceived
(AbstractIoFilterChain.java:787)
at
org.apache.mina.common.support.AbstractIoFilterChain$1.messageReceived(
AbstractIoFilterChain.java:110)
at
org.apache.mina.common.support.AbstractIoFilterChain.callNextMessageReceived
(AbstractIoFilterChain.java:501)
at
org.apache.mina.common.support.AbstractIoFilterChain.messageReceived(
AbstractIoFilterChain.java:492)
at org.apache.mina.transport.socket.nio.SocketIoProcessor.read(
SocketIoProcessor.java:284)
at org.apache.mina.transport.socket.nio.SocketIoProcessor.process(
SocketIoProcessor.java:244)
at
org.apache.mina.transport.socket.nio.SocketIoProcessor.access$400(
SocketIoProcessor.java:41)
at
org.apache.mina.transport.socket.nio.SocketIoProcessor$Worker.run(
SocketIoProcessor.java:565)
It seems to me that its calling the decoding function over and over
again with some kind of a null pointer. blah. I dunno. hope this
helps. I wish I could be more of a help but I dunno any of this spring
mina stuff =( so this is as far as I'll go.
NK
On 1/24/07, Nankun Huang <[EMAIL PROTECTED]> wrote:
> I did a bit more of an investigation into this with the wild hope that
> I might debug this ( with my noob java skillz ). So I set up the
> eclipse debugger and fired it away. Now I finally give up after 3
> hours. However I did have some interesting finds.
>
> Firstly, I started noticing the problem in:
> org.red5.server.net.rtmp.RTMPConnection.java
> As i used my program to subscribe and unsubscribe to my own streams, I
> noticed something:
>
> 1, When a stream is deleted by the
> deleteStreamById(streamId) function, it did not call the
> unreserveStreamId(int streamId) function!! therefore, this causes the
>
> On 1/24/07, Nankun Huang <[EMAIL PROTECTED]> wrote:
> > Hi,
> > Ok, sorry I didn't look at the wiki more carefully =) This
> > actually happens on the playback side too. When a client subscribe to
> > more than 12 streams ( one after another ), the client will not be
> > able to subscribe to any new ones anymore. Thnx
> >
> > NK
> >
> >
> > On 1/24/07, Chris Allen <[EMAIL PROTECTED]> wrote:
> > > Sorry for the confusion. I thought that it had been fixed already.
It is a
> > > priority to have it fixed before the rc2 release though.
> > >
> > > -Chris
> > >
> > >
> > > On 1/24/07, Thijs Triemstra <[EMAIL PROTECTED]> wrote:
> > > > There's a bug for this issue on
> > > http://jira.red5.org/browse/SN-14
> > > >
> > > > Thijs
> > > >
> > > >
> > > > Op 24-jan-2007, om 11:27 heeft Nankun Huang het volgende
geschreven:
> > > >
> > > > > Hi all,
> > > > > According to an email which I got from this list yesterday, it
was
> > > > > stated that the 12 connection limit bug is fixed. I was
overjoyed and
> > > > > rushed to test out if this was the case. However after
downloading the
> > > > > lattest trunk and compiling, this does not seem to be the case.
There
> > > > > also isn't a bug report on the wiki about this and when the bug
was
> > > > > reported unto the mailing list earlier ( by someone else, not me
), it
> > > > > was December and the issue didn't seem to have gotten much
attention
> > > > > from the dev team. Here's a simple adobe flex program which I
made
> > > > > that illustrates the bug. the MXML source file is below:
> > > > >
> > > > > The bug happens when a user publishes a stream, unpublish it,
publish
> > > > > a new stream with a different stream name and then unpublish it
> > > > > again... rinse and repeat about 12 times and it stops working (
I no
> > > > > longer get the publish success message ). I guess the intended
> > > > > functionality is that each user shouldnt' publish more than 12
> > > > > simultaneous streams which makes sense.. however in this case,
once a
> > > > > stream is unpublished, it should just get deleted and no longer
> > > > > "count" toward the user. Thanks for your attention.
> > > > >
> > > > > It seems that when I reach the limit of the number of streams
that can
> > > > > be published, red5 gives me this error in the console
> > > > >
> > > > > [java] [WARN] 575703 pool-1-thread-2:
> > > > > ( org.red5.server.net.rtmp.RTMPHandler
> > > > > .warn ) Unhandled ping: Ping: 3, 0, 0, -1
> > > > > [java] 00 03 00 00 00 00 00 00 00 00
> > > > >
> > > > >
> > > > >
> > > > > ------ and here's the mxml file
> > > > >
> > > > > <?xml version="1.0" encoding="utf-8"?>
> > > > > <mx:Application xmlns:mx=" http://www.adobe.com/2006/mxml";
> > > > > layout="absolute">
> > > > > <mx:Script>
> > > > > <![CDATA[
> > > > > import mx.utils.UIDUtil ;
> > > > > import flash.net.NetStream;
> > > > > import flash.net.NetConnection;
> > > > > import flash.net.ObjectEncoding;
> > > > > import flash.media.Microphone;
> > > > > import flash.events.NetStatusEvent;
> > > > > import mx.controls.Alert;
> > > > >
> > > > > //set to AMF0 for red5
> > > > > NetConnection.defaultObjectEncoding =
> > > ObjectEncoding.AMF0;
> > > > >
> > > > > private var nc: NetConnection;
> > > > > private var ns:NetStream;
> > > > > private var ns_play:NetStream;
> > > > > private var stream_name:String;
> > > > >
> > > > > //connect to server
> > > > > private function connect():void{
> > > > > nc = new NetConnection();
> > > > > nc.client = this;
> > > > > nc.addEventListener(NetStatusEvent.NET_STATUS,
> > > > > netStatusHandler);
> > > > > nc.connect ( "rtmp://localhost/fitcDemo");
> > > > > }
> > > > >
> > > > > //useless callbcks from fitcDemo
> > > > > public function setId( id:String):void{
> > > > > trace("server invoked set ID: "+ id);
> > > > > }
> > > > >
> > > > > public function newStream( id:String):void{
> > > > > trace("server invoked newStream: "+ id);
> > > > > }
> > > > > //start to publish the stream
> > > > > public function startPub():void{
> > > > > var mic:Microphone = Microphone.getMicrophone();
> > > > > // generate a unique stream name every time with
the
> > > > > unique ID generator
> > > > > stream_name = UIDUtil.createUID();
> > > > >
> > > > > this.btnPub.enabled = false;
> > > > > ns = new NetStream(nc);
> > > > > ns.attachAudio( mic )
> > > > > ns.addEventListener( NetStatusEvent.NET_STATUS,
> > > > > netStreamStatusHandler);
> > > > > ns.publish( stream_name );
> > > > >
> > > > > }
> > > > >
> > > > > //stop publishing the stream
> > > > > public function stopPub():void{
> > > > > ns_play.close();
> > > > > ns.close();
> > > > > this.btnPub.enabled = true;
> > > > > this.btnStop.enabled= false;
> > > > > }
> > > > >
> > > > >
> > > > >
> > > > > private function netStatusHandler
> > > > > (event:NetStatusEvent):void{
> > > > > trace(event.info.code);
> > > > > if(event.info.code ==
> > > > > "NetConnection.Connect.Success") {
> > > > > this.btnPub.enabled = true;
> > > > > Alert.show('connected to server!!');
> > > > > }
> > > > > }
> > > > >
> > > > >
> > > > > //when publish is successful, play the stream
> > > > > private function netStreamStatusHandler
> > > > > (event:NetStatusEvent):void{
> > > > > if( event.info.code == "NetStream.Publish.Start"
) {
> > > > > trace(event.info.code);
> > > > > ns_play = new NetStream(nc);
> > > > > ns_play.play( stream_name );
> > > > > this.btnStop.enabled = true;
> > > > > }
> > > > > }
> > > > >
> > > > >
> > > > > ]]>
> > > > > </mx:Script>
> > > > > <mx:Panel x="10" y="10" width="284" height="256"
layout="absolute"
> > > > > title="Red5 12 connection limit bug ">
> > > > > <mx:Button x="10" y="10" label="Connect"
click="connect()"/>
> > > > > <mx:Button x="10" y="40" label="Publish & Play"
> > > > > mouseDown="startPub()" enabled="false" id="btnPub"/>
> > > > > <mx:Button x="125" y="40" label="Stop"
mouseDown="stopPub()"
> > > > > enabled="false" id="btnStop"/>
> > > > > <mx:Text x="10" y="70" text="Press connect to connect to
> > > > > server (pop up will inform you that its connected), when
connected
> > > > > press publish & play to publish with your own mic and play
the
> > > > > same stream you published. Press stop to stop the publishing and
> > > > > playing. Do this for 12 times, it'll die on the 13th time,
> > > > > illustrating the bug. Somtiems it fails before 12 times."
> > > > > width="251"/>
> > > > > </mx:Panel>
> > > > > </mx:Application>
> > > > >
> > > > > _______________________________________________
> > > > > Red5 mailing list
> > > > > [email protected]
> > > > > http://osflash.org/mailman/listinfo/red5_osflash.org
> > > >
> > > >
> > > > _______________________________________________
> > > > Red5 mailing list
> > > > [email protected]
> > > > http://osflash.org/mailman/listinfo/red5_osflash.org
> > > >
> > >
> > >
> > > _______________________________________________
> > > Red5 mailing list
> > > [email protected]
> > > http://osflash.org/mailman/listinfo/red5_osflash.org
> > >
> > >
> > >
> >
>
_______________________________________________
Red5 mailing list
[email protected]
http://osflash.org/mailman/listinfo/red5_osflash.org
_______________________________________________
Red5 mailing list
[email protected]
http://osflash.org/mailman/listinfo/red5_osflash.org
