标签：log cep pac status bash === 重启 ted worker
yarn启动spark进行测试时候发现，几台机器上启动executor都失败了，日志如下2018-11-22 14:32:24 WARN YarnAllocator:66 - Container marked as failed: container_1516236189600_0229_01_000011 on host: cloud3. Exit status: 127. Diagnostics: Exception from container-launch.
Container id: container_1516236189600_0229_01_000011 Exit code: 127 Stack trace: ExitCodeException exitCode=127: at org.apache.hadoop.util.Shell.runCommand(Shell.java:601) at org.apache.hadoop.util.Shell.run(Shell.java:504) at org.apache.hadoop.util.Shell$ShellCommandExecutor.execute(Shell.java:786) at org.apache.hadoop.yarn.server.nodemanager.DefaultContainerExecutor.launchContainer(DefaultContainerExecutor.java:213) at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch.call(ContainerLaunch.java:302) at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch.call(ContainerLaunch.java:82) at java.util.concurrent.FutureTask.run(FutureTask.java:266) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) at java.lang.Thread.run(Thread.java:748) Container exited with a non-zero exit code 127 2018-11-22 14:32:26 INFO YarnAllocator:54 - Driver requested a total number of 0 executor(s).
在google上查了很久也没解决问题，直到看到有个人建议用yarn logs -applicationId <APP_ID>查日志才发现如下情况
Container: container_1516236189600_0231_01_000001 on cloud3_22681
===================================================================
LogType:stderr
Log Upload Time:星期四 十一月 22 14:48:13 +0800 2018
LogLength:75
Log Contents:
/bin/bash: /data/platform/jdk1.8.0_144/bin/java: No such file or directory
LogType:stdout
Log Upload Time:星期四 十一月 22 14:48:13 +0800 2018
LogLength:0
Log Contents:
已经很明显了，然后连到相应机器上看，确实没有jdk1.8.0_144，增加jdk1.8.0_144。重启job发现没问题了。
标签：log cep pac status bash === 重启 ted worker
原文地址：https://www.cnblogs.com/bainily/p/10001369.html
