[
https://issues.apache.org/jira/browse/MESOS-1211?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ian Downes closed MESOS-1211.
-----------------------------
Resolution: Invalid
The code does actually wait for Launcher::recover() to complete before starting
Isolator::recover(); zombies were observed because cgroups::destroy wasn't
waiting until processes had been reaped.
> MesosContainerizer should recover isolators after the launcher recovers
> ------------------------------------------------------------------------
>
> Key: MESOS-1211
> URL: https://issues.apache.org/jira/browse/MESOS-1211
> Project: Mesos
> Issue Type: Bug
> Affects Versions: 0.18.0
> Reporter: Ian Downes
> Assignee: Ian Downes
> Fix For: 0.19.0
>
>
> The current recover code does Launcher::recover and Isolator::recover(s)
> concurrently.
> Isolator::recover should wait until Launcher::recover has completed so any
> processes in orphaned containers have been killed.
> The semantics of Isolator::recover should match that of Isolator::cleanup,
> i.e., there should not be any processes running.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
