[
https://issues.apache.org/jira/browse/HADOOP-5007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12662560#action_12662560
]
Eric Yang commented on HADOOP-5007:
-----------------------------------
The original design had all adaptors execute inside a loop. For this reason,
removing adaptor from the adaptor hash would be sufficient. However, new
adapators may be multi-threaded, hence, adding "adaptor.hardStop()" is ideal.
There might be some changes required for the adaptor to stop itself, and remove
the code for removing from the adaptor hash.
> can't hard-stop Chukwa adaptors
> -------------------------------
>
> Key: HADOOP-5007
> URL: https://issues.apache.org/jira/browse/HADOOP-5007
> Project: Hadoop Core
> Issue Type: Bug
> Components: contrib/chukwa
> Reporter: Ari Rabkin
>
> There are two ways to stop a Chukwa adaptor -- gracefully, via
> Adaptor.shutdown(), and abruptly, via Adaptor.hardStop(). Agent.stopAdaptor()
> should call hardStop() when the user passes in the appropriate flag, instead
> of just leaving the adaptor running.
> This got broken by HADOOP-4709. I assume this was inadvertent; if not, could
> someone explain the rationale for this change?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
