[
https://issues.apache.org/jira/browse/YARN-1070?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13742411#comment-13742411
]
Zhijie Shen commented on YARN-1070:
-----------------------------------
This is because one more CONTAINER_KILLED_ON_REQUEST is emitted at KILLING,
which is created in the following code:
{code}
if (rContainer.cancel(false)) {
if (container.getContainerState() == ContainerState.KILLING) {
dispatcher.getEventHandler().handle(
new ContainerExitEvent(containerId,
ContainerEventType.CONTAINER_KILLED_ON_REQUEST,
ExitCode.TERMINATED.getExitCode(),
"Container terminated before launch."));
}
}
{code}
Did some local test, and found that rContainer.cancel(false) will return true
immediately when containerlaunch.call() is still running. Therefore, It's not
an exclusive indicator to determine whether containerlaunch.call() is canceled
before running.
> ContainerImpl State Machine: Invalid event: CONTAINER_KILLED_ON_REQUEST at
> CONTAINER_CLEANEDUP_AFTER_KILL
> ---------------------------------------------------------------------------------------------------------
>
> Key: YARN-1070
> URL: https://issues.apache.org/jira/browse/YARN-1070
> Project: Hadoop YARN
> Issue Type: Bug
> Components: nodemanager
> Reporter: Hitesh Shah
>
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
