最近发现系统中出现了很多 IOException: Connection reset by peer 与 ClosedChannelException: null
深入看了看代码, 做了些测试, 发现 Connection reset 会在客户端不知道 channel 被关闭的情况下, 触发了 eventloop 的 unsafe.read() 操作抛出
而 ClosedChannelException 一般是由 Netty 主动抛出的, 在 AbstractChannel 以及 SSLHandler 里都可以看到 ClosedChannel 相关的代码
AbstractChannel
static final ClosedChannelException CLOSED_CHANNEL_EXCEPTION = new ClosedChannelException(); ... static { CLOSED_CHANNEL_EXCEPTION.setStackTrace(EmptyArrays.EMPTY_STACK_TRACE); NOT_YET_CONNECTED_EXCEPTION.setStackTrace(EmptyArrays.EMPTY_STACK_TRACE); } ... @Override public void write(Object msg, ChannelPromise promise) { ChannelOutboundBuffer outboundBuffer = this.outboundBuffer; if (outboundBuffer == null) { // If the outboundBuffer is null we know the channel was closed and so // need to fail the future right away. If it is not null the handling of the rest // will be done in flush0() // See https://github.com/netty/netty/issues/2362 safeSetFailure(promise, CLOSED_CHANNEL_EXCEPTION); // release message now to prevent resource-leak ReferenceCountUtil.release(msg); return; } outboundBuffer.addMessage(msg, promise); }
在代码的许多部分, 都会有这个 ClosedChannelException, 大概的意思是说在 channel close 以后, 如果还调用了 write 方法, 则会将 write 的 future 设置为 failure, 并将 cause 设置为 ClosedChannelException, 同样 SSLHandler 中也类似
-----------------
回到 Connection reset by peer, 要模拟这个情况比较简单, 就是在 server 端设置一个在 channelActive 的时候就 close channel 的 handler. 而在 client 端则写一个 Connect 成功后立即发送请求数据的 listener. 如下
client
public static void main(String[] args) throws IOException, InterruptedException { Bootstrap b = new Bootstrap(); b.group(new NioEventLoopGroup()) .channel(NioSocketChannel.class) .handler(new ChannelInitializer<NioSocketChannel>() { @Override protected void initChannel(NioSocketChannel ch) throws Exception { } }); b.connect("localhost", 8090).addListener(new ChannelFutureListener() { @Override public void operationComplete(ChannelFuture future) throws Exception { if (future.isSuccess()) { future.channel().write(Unpooled.buffer().writeBytes("123".getBytes())); future.channel().flush(); } } });
server
public class SimpleServer { public static void main(String[] args) throws Exception { EventLoopGroup bossGroup = new NioEventLoopGroup(1); EventLoopGroup workerGroup = new NioEventLoopGroup(); ServerBootstrap b = new ServerBootstrap(); b.group(bossGroup, workerGroup) .channel(NioServerSocketChannel.class) .option(ChannelOption.SO_REUSEADDR, true) .childHandler(new ChannelInitializer<NioSocketChannel>() { @Override protected void initChannel(NioSocketChannel ch) throws Exception { ch.pipeline().addLast(new SimpleServerHandler()); } }); b.bind(8090).sync().channel().closeFuture().sync(); } } public class SimpleServerHandler extends ChannelInboundHandlerAdapter { @Override public void channelActive(ChannelHandlerContext ctx) throws Exception { ctx.channel().close().sync(); } @Override public void channelRead(ChannelHandlerContext ctx, final Object msg) throws Exception { System.out.println(123); } @Override public void channelInactive(ChannelHandlerContext ctx) throws Exception { System.out.println("inactive"); } }
这种情况之所以能触发 connection reset by peer 异常, 是因为 connect 成功以后, client 段先会触发 connect 成功的 listener, 这个时候 server 段虽然断开了 channel, 也触发 channel 断开的事件 (它会触发一个客户端 read 事件, 但是这个 read 会返回 -1, -1 代表 channel 关闭, client 的 channelInactive 跟 channel active 状态的改变都是在这时发生的), 但是这个事件是在 connect 成功的 listener 之后执行, 所以这个时候 listener 里的 channel 并不知道自己已经断开, 它还是会继续进行 write 跟 flush 操作, 在调用 flush 后, eventloop 会进入 OP_READ 事件里, 这时候 unsafe.read() 就会抛出 connection reset 异常. eventloop 代码如下
NioEventLoop
private static void processSelectedKey(SelectionKey k, AbstractNioChannel ch) { final NioUnsafe unsafe = ch.unsafe(); if (!k.isValid()) { // close the channel if the key is not valid anymore unsafe.close(unsafe.voidPromise()); return; } try { int readyOps = k.readyOps(); // Also check for readOps of 0 to workaround possible JDK bug which may otherwise lead // to a spin loop if ((readyOps & (SelectionKey.OP_READ | SelectionKey.OP_ACCEPT)) != 0 || readyOps == 0) { unsafe.read(); if (!ch.isOpen()) { // Connection already closed - no need to handle write. return; } } if ((readyOps & SelectionKey.OP_WRITE) != 0) { // Call forceFlush which will also take care of clear the OP_WRITE once there is nothing left to write ch.unsafe().forceFlush(); } if ((readyOps & SelectionKey.OP_CONNECT) != 0) { // remove OP_CONNECT as otherwise Selector.select(..) will always return without blocking // See https://github.com/netty/netty/issues/924 int ops = k.interestOps(); ops &= ~SelectionKey.OP_CONNECT; k.interestOps(ops); unsafe.finishConnect(); } } catch (CancelledKeyException e) { unsafe.close(unsafe.voidPromise()); } }
这就是 connection reset by peer 产生的原因
------------------
再来看 ClosedChannelException 如何产生, 要复现他也很简单. 首先要明确, 并没有客户端主动关闭才会出现 ClosedChannelException 这么一说. 下面来看两种出现 ClosedChannelException 的客户端写法
client 1, 主动关闭 channel
public class SimpleClient { private static final Logger logger = LoggerFactory.getLogger(SimpleClient.class); public static void main(String[] args) throws IOException, InterruptedException { Bootstrap b = new Bootstrap(); b.group(new NioEventLoopGroup()) .channel(NioSocketChannel.class) .handler(new ChannelInitializer<NioSocketChannel>() { @Override protected void initChannel(NioSocketChannel ch) throws Exception { } }); b.connect("localhost", 8090).addListener(new ChannelFutureListener() { @Override public void operationComplete(ChannelFuture future) throws Exception { if (future.isSuccess()) { future.channel().close(); future.channel().write(Unpooled.buffer().writeBytes("123".getBytes())).addListener(new ChannelFutureListener() { @Override public void operationComplete(ChannelFuture future) throws Exception { if (!future.isSuccess()) { logger.error("Error", future.cause()); } } }); future.channel().flush(); } } }); } }
只要在 write 之前主动调用了 close, 那么 write 必然会知道 close 是 close 状态, 最后 write 就会失败, 并且 future 里的 cause 就是 ClosedChannelException
--------------------
client 2. 由服务端造成的 ClosedChannelException
public class SimpleClient { private static final Logger logger = LoggerFactory.getLogger(SimpleClient.class); public static void main(String[] args) throws IOException, InterruptedException { Bootstrap b = new Bootstrap(); b.group(new NioEventLoopGroup()) .channel(NioSocketChannel.class) .handler(new ChannelInitializer<NioSocketChannel>() { @Override protected void initChannel(NioSocketChannel ch) throws Exception { } }); Channel channel = b.connect("localhost", 8090).sync().channel(); Thread.sleep(3000); channel.writeAndFlush(Unpooled.buffer().writeBytes("123".getBytes())).addListener(new ChannelFutureListener() { @Override public void operationComplete(ChannelFuture future) throws Exception { if (!future.isSuccess()) { logger.error("error", future.cause()); } } }); } }
服务端
public class SimpleServer { public static void main(String[] args) throws Exception { EventLoopGroup bossGroup = new NioEventLoopGroup(1); EventLoopGroup workerGroup = new NioEventLoopGroup(); ServerBootstrap b = new ServerBootstrap(); b.group(bossGroup, workerGroup) .channel(NioServerSocketChannel.class) .option(ChannelOption.SO_REUSEADDR, true) .childHandler(new ChannelInitializer<NioSocketChannel>() { @Override protected void initChannel(NioSocketChannel ch) throws Exception { ch.pipeline().addLast(new SimpleServerHandler()); } }); b.bind(8090).sync().channel().closeFuture().sync(); } }
这种情况下, 服务端将 channel 关闭, 客户端先 sleep, 这期间 client 的 eventLoop 会处理客户端关闭的时间, 也就是 eventLoop 的 processKey 方法会进入 OP_READ, 然后 read 出来一个 -1, 最后触发 client channelInactive 事件, 当 sleep 醒来以后, 客户端调用 writeAndFlush, 这时候客户端 channel 的状态已经变为了 inactive, 所以 write 失败, cause 为 ClosedChannelException
Netty 中 IOException: Connection reset by peer 与 java.nio.channels.ClosedChannelException: null