[
https://issues.apache.org/jira/browse/YARN-4744?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15176186#comment-15176186
]
Jason Lowe commented on YARN-4744:
----------------------------------
bq. Can we use similar check like
LinuxContainerExecutor#isContainerAlive(ContainerLivenessContext ctx).
That function is implemented in terms of signalContainer (so we have the same
issue), and the process could exit between the check and the subsequent kill
attempt.
bq. My feeling is that the PrivilegedOperationExecutor should log failures
irrespective of the error code
There's always going to be a race where a container can exit before it gets
killed, and I'm not sure we accomplish much besides alarming users when we log
warnings when that occurs. IMHO PrivilegedOperationExecutor should not be the
one that decides what should and shouldn't be logged, since it doesn't have any
context on whether the error is severe enough to warrant it. Instead I think
we should ensure the same data is present in the PrivilegedOperationException
and let the code handling that error perform the logging if it is appropriate
to do so.
> Too many signal to container failure in case of LCE
> ---------------------------------------------------
>
> Key: YARN-4744
> URL: https://issues.apache.org/jira/browse/YARN-4744
> Project: Hadoop YARN
> Issue Type: Bug
> Affects Versions: 2.9.0
> Reporter: Bibin A Chundatt
> Assignee: Sidharta Seethana
>
> Install HA cluster in secure mode
> Enable LCE with cgroups
> Start server with dsperf user
> Submit mapreduce application terasort/teragen with user yarn/dsperf
> Too many signal to container failure
> Submit with user the exception is thrown
> {noformat}
> 2014-03-02 09:20:38,689 INFO
> SecurityLogger.org.apache.hadoop.security.authorize.ServiceAuthorizationManager:
> Authorization successful for testing (auth:TOKEN) for protocol=interface
> org.apache.hadoop.yarn.server.nodemanager.api.LocalizationProtocolPB
> 2014-03-02 09:20:40,158 WARN
> org.apache.hadoop.yarn.server.nodemanager.containermanager.ContainerManagerImpl:
> Event EventType: KILL_CONTAINER sent to absent container
> container_e02_1393731146548_0001_01_000013
> 2014-03-02 09:20:43,071 INFO
> org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch:
> Container container_e02_1393731146548_0001_01_000009 succeeded
> 2014-03-02 09:20:43,072 INFO
> org.apache.hadoop.yarn.server.nodemanager.containermanager.container.ContainerImpl:
> Container container_e02_1393731146548_0001_01_000009 transitioned from
> RUNNING to EXITED_WITH_SUCCESS
> 2014-03-02 09:20:43,073 INFO
> org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch:
> Cleaning up container container_e02_1393731146548_0001_01_000009
> 2014-03-02 09:20:43,075 INFO
> org.apache.hadoop.yarn.server.nodemanager.containermanager.linux.runtime.DelegatingLinuxContainerRuntime:
> Using container runtime: DefaultLinuxContainerRuntime
> 2014-03-02 09:20:43,081 WARN
> org.apache.hadoop.yarn.server.nodemanager.containermanager.linux.privileged.PrivilegedOperationExecutor:
> Shell execution returned exit code: 9. Privileged Execution Operation Output:
> main : command provided 2
> main : run as user is yarn
> main : requested yarn user is yarn
> Full command array for failed execution:
> [/opt/bibin/dsperf/HAINSTALL/install/hadoop/nodemanager/bin/container-executor,
> yarn, yarn, 2, 9370, 15]
> 2014-03-02 09:20:43,081 WARN
> org.apache.hadoop.yarn.server.nodemanager.containermanager.linux.runtime.DefaultLinuxContainerRuntime:
> Signal container failed. Exception:
> org.apache.hadoop.yarn.server.nodemanager.containermanager.linux.privileged.PrivilegedOperationException:
> ExitCodeException exitCode=9:
> at
> org.apache.hadoop.yarn.server.nodemanager.containermanager.linux.privileged.PrivilegedOperationExecutor.executePrivilegedOperation(PrivilegedOperationExecutor.java:173)
> at
> org.apache.hadoop.yarn.server.nodemanager.containermanager.linux.runtime.DefaultLinuxContainerRuntime.signalContainer(DefaultLinuxContainerRuntime.java:132)
> at
> org.apache.hadoop.yarn.server.nodemanager.containermanager.linux.runtime.DelegatingLinuxContainerRuntime.signalContainer(DelegatingLinuxContainerRuntime.java:109)
> at
> org.apache.hadoop.yarn.server.nodemanager.LinuxContainerExecutor.signalContainer(LinuxContainerExecutor.java:513)
> at
> org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch.cleanupContainer(ContainerLaunch.java:520)
> at
> org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainersLauncher.handle(ContainersLauncher.java:139)
> at
> org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainersLauncher.handle(ContainersLauncher.java:55)
> at
> org.apache.hadoop.yarn.event.AsyncDispatcher.dispatch(AsyncDispatcher.java:184)
> at
> org.apache.hadoop.yarn.event.AsyncDispatcher$1.run(AsyncDispatcher.java:110)
> at java.lang.Thread.run(Thread.java:745)
> Caused by: ExitCodeException exitCode=9:
> at org.apache.hadoop.util.Shell.runCommand(Shell.java:927)
> at org.apache.hadoop.util.Shell.run(Shell.java:838)
> at
> org.apache.hadoop.util.Shell$ShellCommandExecutor.execute(Shell.java:1117)
> at
> org.apache.hadoop.yarn.server.nodemanager.containermanager.linux.privileged.PrivilegedOperationExecutor.executePrivilegedOperation(PrivilegedOperationExecutor.java:150)
> ... 9 more
> 2014-03-02 09:20:43,113 INFO
> org.apache.hadoop.yarn.server.nodemanager.NMAuditLogger: USER=yarn
> OPERATION=Container Finished - Succeeded TARGET=ContainerImpl
> RESULT=SUCCESS APPID=application_1393731146548_0001
> CONTAINERID=container_e02_1393731146548_0001_01_000009
> 2014-03-02 09:20:43,115 INFO
> org.apache.hadoop.yarn.server.nodemanager.containermanager.container.ContainerImpl:
> Container container_e02_1393731146548_0001_01_000009 transitioned from
> EXITED_WITH_SUCCESS to DONE
> 2014-03-02 09:20:43,115 INFO
> org.apache.hadoop.yarn.server.nodemanager.containermanager.application.ApplicationImpl:
> Removing container_e02_1393731146548_0001_01_000009 from application
> application_1393731146548_0001
> {noformat}
> Checked the same scenario in 2.7.2 version (not available)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
