[
https://issues.apache.org/jira/browse/MESOS-5874?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15388533#comment-15388533
]
Benjamin Mahler commented on MESOS-5874:
----------------------------------------
This cannot be done safely until we correctly reconcile executors between the
master and the agent. Linking in the related ticket (MESOS-1961).
> Only send ShutdownFrameworkMessage to agents associated with framework.
> -----------------------------------------------------------------------
>
> Key: MESOS-5874
> URL: https://issues.apache.org/jira/browse/MESOS-5874
> Project: Mesos
> Issue Type: Improvement
> Components: master
> Reporter: Jacob Janco
> Assignee: Jacob Janco
> Priority: Minor
> Labels: mesosphere
> Fix For: 1.1.0
>
>
> slave.cpp:2079] Asked to shut down framework ${framework} by master@${master}
> slave.cpp:2094] Cannot shut down unknown framework ${framework}
> For high framework/churn clusters this saturates agent logs with these
> messages. When a framework terminates a ShutdownFrameworkMessage is sent to
> every registered slave in a for loop. This patch proposes sending this
> message to agents with executors associated with the framework.
> Also proposed is moving the logline to VLOG(1).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
