[
https://issues.apache.org/jira/browse/YARN-1897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14936077#comment-14936077
]
Xuan Gong commented on YARN-1897:
---------------------------------
bq. The patch already prints all signals supported if you don't specify any
parameter. Do you want an explicit option called "-all"?
Okay, i think that it is good enough right now.
bq. That also brings up the issue where RM and NM continue to use
NodeHeartbeatResponse's ContainersToCleanup to kill containers due to
preemption. Should we migrate it to ContainersToSignalList? But that could be a
separate jira.
Yes, this is the issue. Let us do it separately. Let us focus on the public API
here.
bq. Regarding the diagnosis, do you want to allow the end user to specify the
reason from CLI/YarnClient? If it is generated only by YARN components, we can
also use enum similar to CMgrCompletedContainersEvent's reason.
For example, if we go to RM web ui/ATS ui, we check the status of all
containers, it is better to show more details,such as "kill by RM because of
Preemption", "kill by the user for testing", etc, in instead of just simply
showing "KILL BY RESOURCEMANAGER". Probably, this can let users better
understand their application. But, right now, I think that it is fine to skip
this. We can do separately if needed.
> CLI and core support for signal container functionality
> -------------------------------------------------------
>
> Key: YARN-1897
> URL: https://issues.apache.org/jira/browse/YARN-1897
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: api
> Reporter: Ming Ma
> Assignee: Ming Ma
> Attachments: YARN-1897-2.patch, YARN-1897-3.patch, YARN-1897-4.patch,
> YARN-1897-5.patch, YARN-1897-6.patch, YARN-1897-7.patch, YARN-1897.1.patch
>
>
> We need to define SignalContainerRequest and SignalContainerResponse first as
> they are needed by other sub tasks. SignalContainerRequest should use
> OS-independent commands and provide a way to application to specify "reason"
> for diagnosis. SignalContainerResponse might be empty.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
