[
https://issues.apache.org/jira/browse/HDDS-194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16525316#comment-16525316
]
Elek, Marton commented on HDDS-194:
-----------------------------------
I believe that TestStorageContainerManager.testBlockDeletingThrottling is not
related. Here the MiniOzoneCluster can't be started while cluster can be
started in all the other tests.
TestCloseContainerByPipeline tests the container closing which also seems to be
independent. (The test failes where the container state is checked on the
datanode. So it's not related to the incoming node reports.)
> Remove NodePoolManager and node pool handling from SCM
> ------------------------------------------------------
>
> Key: HDDS-194
> URL: https://issues.apache.org/jira/browse/HDDS-194
> Project: Hadoop Distributed Data Store
> Issue Type: Improvement
> Components: SCM
> Reporter: Elek, Marton
> Assignee: Elek, Marton
> Priority: Major
> Fix For: 0.2.1
>
> Attachments: HDDS-194.001.patch, HDDS-194.002.patch
>
>
> The current code use NodePoolManager and ContainerSupervisor to group the
> nodes to smaller groups (pools) and handle the pull based node reports group
> by group.
> But this code is not used any more as we switch back to use a push based
> model. In the datanode the reports could be handled by the specific report
> handlers, and in the scm side the reports will be processed by the
> SCMHeartbeatDispatcher which will send the events to the EventQueue.
> As of now the NodePool abstraction could be removed from the code.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
