[
https://issues.apache.org/jira/browse/MESOS-1739?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15259867#comment-15259867
]
Klaus Ma commented on MESOS-1739:
---------------------------------
Agree to send notification to the the frameworks.
For the executors/tasks, will we kill them if slave's attributes changed? I
think it's better to keep them running, two proposal in my mind:
Option 1:
1. operator updates slave's attributes on restart
2. master notify frameworks on attributes changed and rescind offers
3. frameworks update tasks to master according to slave's attributes, which
tasks should be killed and which tasks keep running
4. slave killed tasks if necessary
Option 2:
1. operator updates slave's attributes on restart
2. slave send new attributes to executor
3. executor decide which tasks should be terminated
We need to handle race condition to both options: launching tasks vs. killing
tasks. I think we can pending slave's action until state is stable.
Any comments :).
> Allow slave reconfiguration on restart
> --------------------------------------
>
> Key: MESOS-1739
> URL: https://issues.apache.org/jira/browse/MESOS-1739
> Project: Mesos
> Issue Type: Epic
> Reporter: Patrick Reilly
> Labels: external-volumes, mesosphere, myriad
>
> Make it so that either via a slave restart or a out of process "reconfigure"
> ping, the attributes and resources of a slave can be updated to be a superset
> of what they used to be.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
