hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mukul Kumar Singh (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDDS-298) Implement SCMClientProtocolServer.getContainerWithPipeline for closed containers
Date Tue, 31 Jul 2018 16:58:01 GMT

    [ https://issues.apache.org/jira/browse/HDDS-298?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16563995#comment-16563995

Mukul Kumar Singh commented on HDDS-298:

Thanks for updating the patch [~ajayydv]. Please find my review comment as following.

1) ContainerMapping:208, I feel this line should be removed as call to getReplicationPipeline
will create a new pipeline.
2) ContainerMapping:475, this call should be removed as well. This change isn't part of this
patch, but can be removed
3) ContainerMapping: 532, this line should be removed as well because of above reason.
I feel that getReplicationPipeline should only be called while allocating containers.

4) ContainerMapping: 217, lets keep the pipeline name as "Close-pipline-<containerid>".
let discard the pipeline name maintained for open container once the container is closed.

> Implement SCMClientProtocolServer.getContainerWithPipeline for closed containers
> --------------------------------------------------------------------------------
>                 Key: HDDS-298
>                 URL: https://issues.apache.org/jira/browse/HDDS-298
>             Project: Hadoop Distributed Data Store
>          Issue Type: Improvement
>          Components: SCM
>            Reporter: Elek, Marton
>            Assignee: Ajay Kumar
>            Priority: Critical
>             Fix For: 0.2.1
>         Attachments: HDDS-298.00.patch, HDDS-298.01.patch
> As [~ljain] mentioned during the review of HDDS-245 SCMClientProtocolServer.getContainerWithPipeline
doesn't return with good data for closed containers. For closed containers we are maintaining
the datanodes for a containerId in the ContainerStateMap.contReplicaMap. We need to create
fake Pipeline object on-request and return it for the client to locate the right datanodes
to download data. 

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org

View raw message