You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by "Bill Farner (JIRA)" <ji...@apache.org> on 2013/09/12 21:19:54 UTC
[jira] [Commented] (MESOS-691) Slave should not crash on failure to
launch/kill executors
[ https://issues.apache.org/jira/browse/MESOS-691?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13765790#comment-13765790 ]
Bill Farner commented on MESOS-691:
-----------------------------------
You may also consider failing early when a task ID is passed which is too long for different parts of the system. Or, you could hash variable-length inputs where long inputs could be problematic.
> Slave should not crash on failure to launch/kill executors
> ----------------------------------------------------------
>
> Key: MESOS-691
> URL: https://issues.apache.org/jira/browse/MESOS-691
> Project: Mesos
> Issue Type: Bug
> Reporter: Vinod Kone
> Fix For: 0.15.0
>
>
> At Twitter, we have seen tasks having very long ids causing the cgroups isolator to crash because the resultant cgroup name exceeded NAME_MAX (255) bytes.
> We should carefully revisit places where we do LOG(FATAL) and CHECKs in the slave (and isolators) to ensure we handle the error cases as gracefully as possible instead of crashing hard.
--
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