[
https://issues.apache.org/jira/browse/MESOS-310?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13555333#comment-13555333
]
Vinod Kone commented on MESOS-310:
----------------------------------
https://reviews.apache.org/r/8096/
> cgroups isolation module should not block on fetching executors
> ---------------------------------------------------------------
>
> Key: MESOS-310
> URL: https://issues.apache.org/jira/browse/MESOS-310
> Project: Mesos
> Issue Type: Bug
> Reporter: Vinod Kone
> Assignee: Vinod Kone
>
> With the latest refactor in isolation module, fetching of an executor is done
> by the isolation module.
> If for some reason the fetch blocks (for eg: HDFS issue), all the other
> messages to the isolation module are enqueued. This is problematic because,
> when a new task arrives, the slave currently dispatches resourcesChanged()
> call to the isolation module and immediately sends the task to the executor.
> Subsequently, the executor might be running a task without the isolation
> module updating the cgroups!
--
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
