WDTV Live transcoding hangs after few minutes (SOLVED)

[shagrath]

you need to keep the transcode_block_multiple_connections=true too. basically with these 2 properties, the first connection will be not used, and the second one will prevail

[ditlew]

Hi shagrath,

Okay, I actually tried that also, but I guess I was too sleepy to notice why it didn't worked either. When I do that, the new code throws the following exception in the debug log:

Code: Select all

[New I/O server worker #1-1] TRACE 06:53:19.534 java.util.ConcurrentModificationException
[New I/O server worker #1-1] TRACE 06:53:19.534    at java.util.AbstractList$Itr.checkForComodification(Unknown Source)
[New I/O server worker #1-1] TRACE 06:53:19.534    at java.util.AbstractList$Itr.next(Unknown Source)
[New I/O server worker #1-1] TRACE 06:53:19.534    at net.pms.io.BufferedOutputFile.getInputStream(BufferedOutputFile.java:171)
[New I/O server worker #1-1] TRACE 06:53:19.534    at net.pms.io.ProcessWrapperImpl.getInputStream(ProcessWrapperImpl.java:154)
[New I/O server worker #1-1] TRACE 06:53:19.534    at net.pms.dlna.DLNAResource.getInputStream(DLNAResource.java:906)
[New I/O server worker #1-1] TRACE 06:53:19.534    at net.pms.network.RequestV2.answer(RequestV2.java:196)
[New I/O server worker #1-1] TRACE 06:53:19.534    at net.pms.network.RequestHandlerV2.writeResponse(RequestHandlerV2.java:208)
[New I/O server worker #1-1] TRACE 06:53:19.534    at net.pms.network.RequestHandlerV2.messageReceived(RequestHandlerV2.java:181)
[New I/O server worker #1-1] TRACE 06:53:19.534    at org.jboss.netty.handler.stream.ChunkedWriteHandler.handleUpstream(ChunkedWriteHandler.java:117)
[New I/O server worker #1-1] TRACE 06:53:19.534    at org.jboss.netty.handler.codec.http.HttpChunkAggregator.messageReceived(HttpChunkAggregator.java:101)
[New I/O server worker #1-1] TRACE 06:53:19.534    at org.jboss.netty.channel.Channels.fireMessageReceived(Channels.java:391)
[New I/O server worker #1-1] TRACE 06:53:19.550    at org.jboss.netty.handler.codec.replay.ReplayingDecoder.unfoldAndfireMessageReceived(ReplayingDecoder.java:506)
[New I/O server worker #1-1] TRACE 06:53:19.550    at org.jboss.netty.handler.codec.replay.ReplayingDecoder.callDecode(ReplayingDecoder.java:490)
[New I/O server worker #1-1] TRACE 06:53:19.550    at org.jboss.netty.handler.codec.replay.ReplayingDecoder.messageReceived(ReplayingDecoder.java:427)
[New I/O server worker #1-1] TRACE 06:53:19.550    at org.jboss.netty.channel.Channels.fireMessageReceived(Channels.java:345)
[New I/O server worker #1-1] TRACE 06:53:19.550    at org.jboss.netty.channel.Channels.fireMessageReceived(Channels.java:332)
[New I/O server worker #1-1] TRACE 06:53:19.550    at org.jboss.netty.channel.socket.nio.NioWorker.read(NioWorker.java:323)
[New I/O server worker #1-1] TRACE 06:53:19.565    at org.jboss.netty.channel.socket.nio.NioWorker.processSelectedKeys(NioWorker.java:275)
[New I/O server worker #1-1] TRACE 06:53:19.565    at org.jboss.netty.channel.socket.nio.NioWorker.run(NioWorker.java:196)
[New I/O server worker #1-1] TRACE 06:53:19.565    at org.jboss.netty.util.internal.IoWorkerRunnable.run(IoWorkerRunnable.java:46)
[New I/O server worker #1-1] TRACE 06:53:19.565    at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(Unknown Source)
[New I/O server worker #1-1] TRACE 06:53:19.565    at java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)
[New I/O server worker #1-1] TRACE 06:53:19.565    at java.lang.Thread.run(Unknown Source)

[shagrath]

could you post the whole log ?

also, you can try this little patch, drop it next to PMS.exe and restart:

http://ps3mediaserver.org/files/patchwd/update.jar

[ditlew]

Hi shagrath,

Sure thing, I don't get home until tomorrow sadly. But I'll get back to you asap then.

[Siutsch]

Hi shagrath,

had the same problem with PS3 as Client.

Most worked fine, except DVD Images, Video_TS or ISO.

Buffer filled up, but after 1 oder 2 minutes buffer goes to 0 and Playback stopped.

Had a second Client on network, Samsung A956 TV, but after removing the LAN cable noting changed,
had still the same problems.

Tried the new version 369 with update.jar and the 2 new property parameters and

IT WORKS NOW!

Thank you,
Siutsch.

[ditlew]

I'm back home.

Tested the patch. It doesn't produce any video on the client, the clients timer runs though, but is shows notthing but black screen and doesnt stop when it reaches the end timewize. I've just noticed that memcoder goes zombie after a few seconds. I've attached 2 debug.logs one with and one without the patch.

[ditlew]

I've found more ways to get around the input: 2 problem, but every solution I've found leads to the opposit of the original problem, now the stream gets suspended due to the buffer gets drained instead. Memcoder are still working but the buffer goes below the minimum limit and stays there.

Got any ideas?

*update* I've attached the debug.log from such a run (ends with me pressing stop after the video has frozen)

I've been coding and debuggin on this for quite some days now, hope you can give me an new idea for a solution.

[shagrath]

Got any ideas?

yes, I should read my buggy code :p

try this one: http://ps3mediaserver.org/files/patchwd2/update.jar

[ditlew]

Your new patch result in the original problem, an overfilled suspended buffer ..

This is one of my ways to get the opposit .. an empty suspended buffer

Code snippet from BufferedOutputFile.java

Code: Select all

      WaitBufferedInputStream atominputStream = null;
      if (!PMS.getConfiguration().getTrancodeBlocksMultipleConnections() || getCurrentInputStream() == null) {
         //atominputStream = new WaitBufferedInputStream(this);
         //inputStreams.add(atominputStream);
         
         // Ditlew
         if (inputStreams.size() == 0)
         {
            atominputStream = new WaitBufferedInputStream(this);
            inputStreams.add(atominputStream);
         }
         else
            atominputStream = (WaitBufferedInputStream)inputStreams.get(0);
               
      }


*update* I'll be debuggin NamedWindowsPipe tonight. From what I can see that is the class that should be feeding the buffer.


*update2* The following change instead of the above will get just as far.

Code snippet from BufferedOutputFile.java

Code: Select all

   public WaitBufferedInputStream getCurrentInputStream() {
      if (inputStreams.size() > 0)
      {
         // Ditlew
         //return inputStreams.get(0);
         return inputStreams.get(inputStreams.size() - 1);
      }
      else
         return null;
   }


[ditlew]

Hi,

My finding so far is that it is the follow line in NamedWindowsPipe that halts and hence the buffer goes empty:

boolean fSuccess = Kernel32.INSTANCE.ReadFile(handle1,
buffer, BUFSIZE, ibr, null);

.. now I need to find out why it halts.