You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@mesos.apache.org by "Patrick Reilly (JIRA)" <ji...@apache.org> on 2014/08/26 23:33:59 UTC

[jira] [Updated] (MESOS-350) Explore disk io isolation in cgroups

     [ https://issues.apache.org/jira/browse/MESOS-350?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Patrick Reilly updated MESOS-350:
---------------------------------

    Shepherd: Benjamin Hindman

> Explore disk io isolation in cgroups
> ------------------------------------
>
>                 Key: MESOS-350
>                 URL: https://issues.apache.org/jira/browse/MESOS-350
>             Project: Mesos
>          Issue Type: Improvement
>            Reporter: Sathya Hariesh
>            Assignee: Patrick Reilly
>
> Currently there is no disk isolation in place and this affects an executor to be starved of disk when another disk heavy operation such as copying a multi gigabyte file is being performed by another executor.
> At Twitter, the executor performs a few fsync operations and has specific wait periods for these operations to succeed. When these operations take longer than expected during termination of tasks, it would potentially result in LOST tasks.



--
This message was sent by Atlassian JIRA
(v6.2#6252)