You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mesos.apache.org by "Michael Park (JIRA)" <ji...@apache.org> on 2017/02/16 17:24:41 UTC

[jira] [Updated] (MESOS-7061) Re-persist tasks/executors with allocation info during agent recovery.

     [ https://issues.apache.org/jira/browse/MESOS-7061?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Michael Park updated MESOS-7061:
--------------------------------
    Sprint: Mesosphere Sprint 51

> Re-persist tasks/executors with allocation info during agent recovery.
> ----------------------------------------------------------------------
>
>                 Key: MESOS-7061
>                 URL: https://issues.apache.org/jira/browse/MESOS-7061
>             Project: Mesos
>          Issue Type: Task
>          Components: agent
>            Reporter: Benjamin Mahler
>            Assignee: Michael Park
>
> When the agent is upgraded, it will need to re-persist resources for recovered active tasks and executors, but with Resource.allocation_info.role set to FrameworkInfo.role.
> If this agent receives new tasks from an old master (because the master has not been upgraded yet), it will also augment the resources to have Resource.allocation_info.role set prior to persisting on disk. This is necessary to ensure we continue to charge existing tasks / executors if the framework changes its role(s).
> Importantly, re-persisting will not prevent downgrading the agent since a downgraded agent will simply ignore the unknown fields.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)