一、错误日志

org.apache.flink.runtime.io.network.netty.exception.RemoteTransportException: Error at remote task manager 'xx.xxx.xxx.xxx/xxx.xxx.xxx.xxx:34750'.
at org.apache.flink.runtime.io.network.netty.CreditBasedPartitionRequestClientHandler.decodeMsg(CreditBasedPartitionRequestClientHandler.java:325) ~[flinkjob:?]
at org.apache.flink.runtime.io.network.netty.CreditBasedPartitionRequestClientHandler.channelRead(CreditBasedPartitionRequestClientHandler.java:214) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:379) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:365) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:357) ~[flinkjob:?]
at org.apache.flink.runtime.io.network.netty.NettyMessageClientDecoderDelegate.channelRead(NettyMessageClientDecoderDelegate.java:112) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:379) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:365) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:357) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.DefaultChannelPipeline$HeadContext.channelRead(DefaultChannelPipeline.java:1410) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:379) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:365) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.DefaultChannelPipeline.fireChannelRead(DefaultChannelPipeline.java:919) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.epoll.AbstractEpollStreamChannel$EpollStreamUnsafe.epollInReady(AbstractEpollStreamChannel.java:792) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.epoll.EpollEventLoop.processReady(EpollEventLoop.java:475) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.epoll.EpollEventLoop.run(EpollEventLoop.java:378) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.util.concurrent.SingleThreadEventExecutor$4.run(SingleThreadEventExecutor.java:989) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.util.internal.ThreadExecutorMap$2.run(ThreadExecutorMap.java:74) ~[flinkjob:?]
at java.lang.Thread.run(Thread.java:748) [?:1.8.0_144]
Caused by: org.apache.flink.runtime.io.network.partition.ProducerFailedException: java.lang.NullPointerException
at org.apache.flink.runtime.io.network.netty.PartitionRequestQueue.writeAndFlushNextMessageIfPossible(PartitionRequestQueue.java:224) ~[flinkjob:?]
at org.apache.flink.runtime.io.network.netty.PartitionRequestQueue.enqueueAvailableReader(PartitionRequestQueue.java:110) ~[flinkjob:?]
at org.apache.flink.runtime.io.network.netty.PartitionRequestQueue.userEventTriggered(PartitionRequestQueue.java:173) ~[flinkjob:?]
at org.apache.flink.shaded.netty4.io.netty.channel.AbstractChannelHandlerContext.invokeUserEventTriggered(AbstractChannelHandlerContext.java:346) ~

二、原因分析

　　由于启动参数中只设置了一个插槽，而多个Task任务都来共享这个插槽，导致网络出现繁忙、一些Task容器无法使用到插槽，从而出现network.partition.ProducerFailedException

三、解决方案

　　发生报错之前，我的启动参数如下：

bin/flink run -d -m yarn-cluster -yjm 2048m -ytm 4096m -p 1 -ys 1 -yD taskmanager.memory.managed.size=265M -c 全类名 flinkjob.jar

　　解决方案是：合理设置并行度以及插槽数，在启动参数中指定即可

bin/flink run -d -m yarn-cluster -yjm 2048m -ytm 4096m -p 5 -ys 3 -yD taskmanager.memory.managed.size=265M -c 全类名 flinkjob.jar

参数说明：

　　-m：指定yarn集群

　　-yjm：指定jobmanager的内存

　　-ytm：指定taskmanager的内存

　　-p：指定并行度

　　-ys：指定插槽的数量

　　-yD：后台运行

　　taskmanager.memory.managed.size：taskmanager直接管理内存