Hi Emmanuel, Tried that, no success. To overcome my problem I had to take out Mina completely and write a custom implementation based on non-blocking queues for the UDP interface and use a direct mechanism to java.net.Socket for TCP interface. I guess this limitation that mina has for high data rates is a performance related bug?
Cheers, Brendan -----Original Message----- From: Emmanuel Lécharny [mailto:elecha...@gmail.com] Sent: 14 August 2012 17:21 To: users@mina.apache.org Subject: Re: Thread priorities between request and response Le 8/14/12 5:45 PM, Brendan Crowley a écrit : > Hi Emmanuel, > > Thanks for that, that makes sense, can you give me guidance as to how to > replace the Executor Filter from the chain and use more IoProcessor instead? This is pretty easy : - first, get rid of the executor (simply not nject it in the chain) - second, when you declare your Acceptor, you can provide the number of IoProcessor to use : /** * Constructor for {@link NioSocketAcceptor} using default parameters, and * given number of {@link NioProcessor} for multithreading I/O operations. * * @param processorCount the number of processor to create and place in a * {@link SimpleIoProcessorPool} */ public NioSocketAcceptor(int processorCount) { -- Regards, Cordialement, Emmanuel Lécharny www.iktek.com
