ClosedChannelException thrown when stopping Channel

["이희승 (Trustin Lee)"]

I modified Netty NIO transport implementation to raise
ClosedChannelException only once even if there were many pending write
requests, because I couldn't think of any negative side effect but
speedup in connection cleanup.  Let me know if this change breaks your
application.

Thanks for a good idea,
Trustin

On 07/08/2009 09:31 PM, neilson9 wrote:
> Sorry - I realised my mistake after the post - throwing an exception makes
> sense when a write event is pending...
> Regards Neil.
> 
> 
> Trustin Lee wrote:
>> If a channel has been closed and there are pending writes which were not
>> written yet, ClosedChannelException is raised for each write request.
>> You can simply discard the exception if it was intentional.
>>
>> One idea though is it might be a better idea to fire exceptionCaught
>> event only once while notifying all futures.  I'm still not sure about
>> the negative side effect of this change.  WDYT?
>>
>> Thanks,
>> Trustin
>>
>> On 07/08/2009 09:21 PM, neilson9 wrote:
>>> Hi,
>>> In my connection pool Im invalidating and closing connections when they
>>> are
>>> no longer needed (i.e. pools hightwater mark is reached) and I seem to
>>> get
>>> many ClosedChannelExceptions. Looking at the Stack and the Source Code
>>> [below] it looks like I will either get a NotYetConnectedException OR a
>>> ClosedChannelException - when stopping a channel.
>>>
>>> Is it possible to prevent this from happening? 
>>> Regards Neil.
>>>
>>> >From NioWorker.cleanUpWriteBuffer
>>>              // Create the exception only once to avoid the excessive
>>> overhead
>>>                 // caused by fillStackTrace.
>>>                 if (channel.isOpen()) {
>>>                     cause = new NotYetConnectedException();
>>>                 } else {
>>>                     cause = new ClosedChannelException();
>>>                 }
>>>                 evt.getFuture().setFailure(cause);
>>>
>>>                 fireExceptionCaught(channel, cause);
>>>
>>> The exception picked up by my Handler
>>> 2009-07-08 13:10:16,272 WARN 3:10000-10-10 (netty.NettyClientHandler)	
>>> com.liquidlabs.transport.netty.NettyClientHandler at 3f84e8 m:null
>>> Client[null]] Unexpected exception from downstream, latch[null]
>>> java.nio.channels.ClosedChannelException
>>> 	at
>>> org.jboss.netty.channel.socket.nio.NioWorker.cleanUpWriteBuffer(NioWorker.java:609)
>>> 	at
>>> org.jboss.netty.channel.socket.nio.NioWorker.close(NioWorker.java:568)
>>> 	at
>>> org.jboss.netty.channel.socket.nio.NioClientSocketPipelineSink.eventSunk(NioClientSocketPipelineSink.java:97)
>>> 	at org.jboss.netty.channel.Channels.close(Channels.java:1043)
>>> 	at
>>> org.jboss.netty.channel.AbstractChannel.close(AbstractChannel.java:185)
>>> 	at com.ll.transport.netty.NettySenderImpl.stop(NettySenderImpl.java:118)
>>>
>>>
>> _______________________________________________
>> netty-users mailing list
>> netty-users at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/netty-users
>>
>>
>