You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by "Benjamin Mahler (JIRA)" <ji...@apache.org> on 2013/12/03 01:32:35 UTC
[jira] [Commented] (MESOS-858) Ignore launch/kill requests in the
slave originating from non-leading masters.
[ https://issues.apache.org/jira/browse/MESOS-858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13837139#comment-13837139 ]
Benjamin Mahler commented on MESOS-858:
---------------------------------------
https://reviews.apache.org/r/15951/
https://reviews.apache.org/r/15952/
> Ignore launch/kill requests in the slave originating from non-leading masters.
> ------------------------------------------------------------------------------
>
> Key: MESOS-858
> URL: https://issues.apache.org/jira/browse/MESOS-858
> Project: Mesos
> Issue Type: Bug
> Components: slave
> Reporter: Benjamin Mahler
> Assignee: Benjamin Mahler
> Fix For: 0.16.0
>
>
> These were left as TODOs in the code when applying a related fix:
> void Slave::runTask(
> const FrameworkInfo& frameworkInfo,
> const FrameworkID& frameworkId,
> const string& pid,
> const TaskInfo& task)
> {
> // TODO(bmahler): Consider ignoring requests not originating from the
> // expected master.
> void Slave::killTask(const FrameworkID& frameworkId, const TaskID& taskId)
> {
> // TODO(bmahler): Consider ignoring requests not originating from the
> // expected master.
> We've observed a scenario in which a non-leading master flushed it's queue of messages before committing suicide, the framework in this case launched the same task twice, but regardless we should ignore these messages in the slave.
--
This message was sent by Atlassian JIRA
(v6.1#6144)