flink 致命错误 Akka
搭建 flink standalone 集群后，taskmanager 总是不明原因的挂掉，导致任务失败。查询日志发现：
2018-10-25 03:27:05,197 INFO org.apache.flink.runtime.blob.PermanentBlobCache - Shutting down BLOB cache
2018-10-25 03:27:05,197 INFO org.apache.flink.runtime.blob.TransientBlobCache - Shutting down BLOB cache
2018-10-25 03:27:05,197 INFO org.apache.flink.runtime.state.TaskExecutorLocalStateStoresManager - Shutting down TaskExecutorLocalStateStoresManager.
2018-10-25 03:27:05,200 ERROR akka.actor.ActorSystemImpl - Uncaught error from thread [flink-akka.actor.default-dispatcher-31]: Compressed class space, shutting down JVM since 'akka.jvm-exit-on-fatal-error' is enabled for ActorSystem[flink]
java.lang.OutOfMemoryError: Compressed class space
2018-10-25 03:27:05,200 ERROR akka.actor.ActorSystemImpl - Uncaught error from thread [flink-akka.actor.default-dispatcher-212]: Compressed class space, shutting down JVM since 'akka.jvm-exit-on-fatal-error' is enabled for ActorSystem[flink]
java.lang.OutOfMemoryError: Compressed class space
2018-10-25 03:27:05,200 ERROR akka.actor.ActorSystemImpl - Uncaught error from thread [flink-akka.actor.default-dispatcher-22]: Compressed class space, shutting down JVM since 'akka.jvm-exit-on-fatal-error' is enabled for ActorSystem[flink]
java.lang.OutOfMemoryError: Compressed class space
2018-10-25 03:27:05,225 INFO org.apache.flink.runtime.io.disk.iomanager.IOManager - I/O manager removed spill file directory /tmp/flink-io-d7a472b1-6c12-4149-aa56-997c0963ceb5发生了致命错误导致 JVM 进程被 shut down，原因 java.lang.OutOfMemoryError: Compressed class space，
flink 使用 akka 框架作为分布式 RPC 处理，详情查看参考资料，猜测是因为下面这两个参数：
我的任务是流式任务，开启了 checkpoint，在任务过程中报错时会不断的进行重试，重试就会与 jobmanager 进行通信，使用 akka 方式，很容易就达到瓶颈（程序主要做数据处理，遇到异常数据没有进行异常捕捉，所以就一直在抛异常，一直在重试）。
解决方法，适当调大默认配置值，其次，主要还是要在应用中捕获异常，目前已正常运行 20 天，后续出现问题再补充。
参考 ：https://cwiki.apache.org/confluence/display/FLINK/Akka+and+Actors
https://ci.apache.org/projects/flink/flink-docs-release-1.7/ops/config.html