問題描述

public class TimeServer { public void bind(int port) {try { EventLoopGroup bossGroup = new NioEventLoopGroup(); EventLoopGroup workGroup = new NioEventLoopGroup(); ServerBootstrap b = new ServerBootstrap(); b.group(bossGroup, workGroup).channel(NioServerSocketChannel.class).option(ChannelOption.SO_BACKLOG, 1024) .childHandler(new ChannelInitializer<SocketChannel>() {@Overrideprotected void initChannel(SocketChannel arg0) throws Exception { System.out.println('初始化'); arg0.pipeline().addLast(new TimeHandler());} }); ChannelFuture future = b.bind(port).sync(); System.out.println('執(zhí)行這里'); future.channel().closeFuture().sync(); System.out.println('執(zhí)行這里');} catch (InterruptedException e) { e.printStackTrace();} } public static void main(String[] args) {new TimeServer().bind(10000); }}public class TimeHandler extends ChannelInboundHandlerAdapter {@Override public void channelRead(ChannelHandlerContext ctx, Object msg) throws Exception {super.channelRead(ctx, msg); ByteBuf buf=(ByteBuf) msg;////byte[] butfs = buf.array();//報錯 System.out.println(buf.readableBytes());byte[] butfs = new byte[buf.readableBytes()];buf.readBytes(butfs);System.out.println(new String(butfs,'UTF-8'));System.out.println(msg); }}

客戶端使用的是BIO的模型：

public static void main(String[] args) throws Exception {final int port = 10000;//NioServer server = new NioServer(port);//server.init();/// ========================================================// 接下來模擬3個Client并發(fā)訪問服務(wù)器int poolsize = 1;ExecutorService pool = Executors.newFixedThreadPool(poolsize);Collection<Callable> tasks = new ArrayList<Callable>(10);final String clientname = 'clientThread';for (int i = 0; i < poolsize; i++) { final int n = i; // 若每一個Client都保持使用BIO方式發(fā)送數(shù)據(jù)到Server，并讀取數(shù)據(jù)。 tasks.add(new Callable() {@Overridepublic Object call() throws Exception { Socket socket = new Socket('127.0.0.1', port); final InputStream input = socket.getInputStream(); final OutputStream out = socket.getOutputStream(); final String clientname_n = clientname + '_' + n; // BIO讀取數(shù)據(jù)線程 new Thread(clientname_n + '_read') {@Overridepublic void run() { byte[] bs = new byte[1024]; while (true) {try { Thread.sleep(1000);} catch (InterruptedException e) { e.printStackTrace();}int len = 0;try { while ((len = input.read(bs)) != -1) {System.out.println('Clinet thread ' + Thread.currentThread().getName()+ ' read: ' + new String(bs, 0, len)); }} catch (IOException e) { e.printStackTrace();} }} }.start(); // BIO寫數(shù)據(jù)線程 new Thread(clientname_n + '_write') {@Overridepublic void run() { int a = 0; while (true) {try { Thread.sleep(100);} catch (InterruptedException e) { e.printStackTrace();}String str = Thread.currentThread().getName() + ' hello, ' + a;try { out.write(str.getBytes()); out.flush(); a++;} catch (IOException e) { e.printStackTrace();} }} }.start(); return null;} });}pool.invokeAll((Collection<? extends Callable<Object>>) tasks);//server.go(); }

結(jié)果運行的時候出現(xiàn)了以下錯誤：

月 13, 2017 5:52:56 下午 io.netty.channel.DefaultChannelPipeline onUnhandledInboundException警告: An exceptionCaught() event was fired, and it reached at the tail of the pipeline. It usually means the last handler in the pipeline did not handle the exception.io.netty.util.IllegalReferenceCountException: refCnt: 0 at io.netty.buffer.AbstractByteBuf.ensureAccessible(AbstractByteBuf.java:1408) at io.netty.buffer.AbstractByteBuf.checkReadableBytes0(AbstractByteBuf.java:1394) at io.netty.buffer.AbstractByteBuf.checkReadableBytes(AbstractByteBuf.java:1383) at io.netty.buffer.AbstractByteBuf.readBytes(AbstractByteBuf.java:850) at io.netty.buffer.AbstractByteBuf.readBytes(AbstractByteBuf.java:858) at test.netty.TimeHandler.channelRead(TimeHandler.java:17) at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:362) at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:348) at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:340) at io.netty.channel.DefaultChannelPipeline$HeadContext.channelRead(DefaultChannelPipeline.java:1334) at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:362) at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:348) at io.netty.channel.DefaultChannelPipeline.fireChannelRead(DefaultChannelPipeline.java:926) at io.netty.channel.nio.AbstractNioByteChannel$NioByteUnsafe.read(AbstractNioByteChannel.java:134) at io.netty.channel.nio.NioEventLoop.processSelectedKey(NioEventLoop.java:624) at io.netty.channel.nio.NioEventLoop.processSelectedKeysOptimized(NioEventLoop.java:559) at io.netty.channel.nio.NioEventLoop.processSelectedKeys(NioEventLoop.java:476) at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:438) at io.netty.util.concurrent.SingleThreadEventExecutor$5.run(SingleThreadEventExecutor.java:858) at io.netty.util.concurrent.DefaultThreadFactory$DefaultRunnableDecorator.run(DefaultThreadFactory.java:144) at java.lang.Thread.run(Thread.java:745)

請問這是為什么呢？

問題解答

在netty4中，對象的生命周期由引用計數(shù)器控制，ByteBuf就是如此，每個對象的初始化引用計數(shù)為1，調(diào)用一次release方法，引用計數(shù)器會減1，當嘗試訪問計數(shù)器為0的，對象時，會拋出IllegalReferenceCountException，正如ensureAccessible的實現(xiàn)，更加詳細的解釋可以參考官方文檔

AbstractByteBuf.java

protected final void ensureAccessible() {if (refCnt() == 0) { throw new IllegalReferenceCountException(0);} }

注意TZ的TimeHandler類中的 super.channelRead(ctx, msg);這行代碼。追蹤調(diào)用路徑，

private void invokeChannelRead(Object msg) {try { ((ChannelInboundHandler) handler()).channelRead(this, msg);} catch (Throwable t) { notifyHandlerException(t);} }

最終調(diào)用的代碼是：ReferenceCountUtil.release(msg)

public static boolean release(Object msg) {if (msg instanceof ReferenceCounted) { return ((ReferenceCounted) msg).release();}return false; }

也就是每次super.channelRead(ctx, msg);后，ByteBuf就會調(diào)用release()方法，計數(shù)器減一，然后在buf.readBytes(butfs);這行代碼就會校驗ensureAccessible(),計數(shù)器為0，netty認為ByteBuf對象已經(jīng)釋放，就拋出異常。

去掉TimeHandler中這行代碼 super.channelRead(ctx, msg);ByteBuf對象誰處理誰釋放。

An exceptionCaught() event was fired, and it reached at the tail of the pipeline. It usually means the last handler in the pipeline did not handle the exception.缺少exceptionCaught(),在在server端最后一個Handler中增加exceptionCaught()