[jira] [Updated] (AURORA-175) thermos runner should discriminate failures using exit status

["Chris Lambert (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/AURORA-175?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Chris Lambert updated AURORA-175:
---------------------------------
    Sprint: Aurora Q3 Sprint 3, Aurora Q4 Sprint 1  (was: Aurora Q3 Sprint 3)

> thermos runner should discriminate failures using exit status
> -------------------------------------------------------------
>
>                 Key: AURORA-175
>                 URL: https://issues.apache.org/jira/browse/AURORA-175
>             Project: Aurora
>          Issue Type: Task
>          Components: Executor, Thermos
>            Reporter: brian wickman
>            Assignee: brian wickman
>            Priority: Critical
>
> We do the correct thing on the executor side when there is a configuration/interpolation problem (report FAILURE.)
> On the thermos_runner side, we don't exit with a separate exit status for bad configuration, nor do we even pay attention to the exit status from the thermos executor.  So when the runner exits unexpectedly, that's always treated as a LOST.  Instead there should be a contract between the thermos_runner and thermos_executor about certain classes of failures indicated by exit statuses so that we can differentiate between legit LOST and FAILURE, for example if the user no longer exists on the box.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)