[
https://issues.apache.org/jira/browse/MESOS-284?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vinod Kone closed MESOS-284.
----------------------------
Resolution: Fixed
https://reviews.apache.org/r/7270/
> Short-term fix for fire-walling slave shutdown and lost slave messages from
> the 'wrong' master
> ----------------------------------------------------------------------------------------------
>
> Key: MESOS-284
> URL: https://issues.apache.org/jira/browse/MESOS-284
> Project: Mesos
> Issue Type: Bug
> Reporter: Vinod Kone
> Assignee: Vinod Kone
>
> A recent network partition at Twitter exposed a bug, where 2 masters were
> running as leaders. And even though the slaves and framework re-registered
> with a new master, they still accepted shutdown/lost slave messages from the
> old master, resulting in an instantaneous roll of the entire cluster.
> While in the longer-term we need a Firewall abstraction to allow messages
> from specific entities, in the short-term we will just drop the above
> messages.
--
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
