[
https://issues.apache.org/jira/browse/HDDS-576?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16684659#comment-16684659
]
Yiqun Lin commented on HDDS-576:
--------------------------------
[~nandakumar131], as we have removed ALLOCATED, CREATING State for containers.
We can reword the javadoc of {{ContainerStateManager}} in a separate jira.
{noformat}
* This is how a create container happens: 1. When a container is created, the
* Server(or SCM) marks that Container as ALLOCATED state. In this state, SCM
* has chosen a pipeline for container to live on. However, the container is not
* created yet. This container along with the pipeline is returned to the
* client.
* <p>
* 2. The client when it sees the Container state as ALLOCATED understands that
* container needs to be created on the specified pipeline. The client lets the
* SCM know that saw this flag and is initiating the on the data nodes.
* <p>
....
{noformat}
> Move ContainerWithPipeline creation to RPC endpoint
> ---------------------------------------------------
>
> Key: HDDS-576
> URL: https://issues.apache.org/jira/browse/HDDS-576
> Project: Hadoop Distributed Data Store
> Issue Type: Bug
> Components: SCM
> Reporter: Mukul Kumar Singh
> Assignee: Nanda kumar
> Priority: Major
> Fix For: 0.4.0
>
> Attachments: HDDS-576.000.patch
>
>
> With independent Pipeline and Container Managers in SCM, the creation of
> ContainerWithPipeline can be moved to RPC endpoint. This will ensure clear
> separation of the pipeline Manager and Container Manager
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
