[
https://issues.apache.org/jira/browse/HDDS-48?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Arpit Agarwal updated HDDS-48:
------------------------------
Description:
We propose refactoring the HDDS DataNode IO path to enforce clean separation
between the Container management and the Storage layers. All components
requiring access to HDDS containers on a Datanode should do so via this Storage
layer.
The proposed Storage layer would be responsible for end-to-end disk and volume
management. This involves running disk checks and detecting disk failures,
distributing data across disks as per the configured policy, collecting
performance statistics.
Attached Design Doc gives an overview of the proposed class diagram.
was:
We propose refactoring the HDDS DataNode IO path to enforce clean separation
between the Container management and the Storage layers. All components
requiring access to HDDS containers on a Datanode should do so via this Storage
layer.
The proposed Storage layer would be responsible for end-to-end disk and volume
management. This involves running disk checks and detecting disk failures,
distributing data across disks as per the configured policy, collecting
performance statistics and verifying the integrity of the data.
Attached Design Doc gives an overview of the proposed class diagram.
> ContainerIO - Storage Management
> --------------------------------
>
> Key: HDDS-48
> URL: https://issues.apache.org/jira/browse/HDDS-48
> Project: Hadoop Distributed Data Store
> Issue Type: Improvement
> Reporter: Hanisha Koneru
> Assignee: Hanisha Koneru
> Priority: Major
> Fix For: 0.2.1
>
> Attachments: ContainerIO-StorageManagement-DesignDoc.pdf, HDDS
> DataNode Disk Layout.pdf, HDDS-48.00.patch
>
>
> We propose refactoring the HDDS DataNode IO path to enforce clean separation
> between the Container management and the Storage layers. All components
> requiring access to HDDS containers on a Datanode should do so via this
> Storage layer.
> The proposed Storage layer would be responsible for end-to-end disk and
> volume management. This involves running disk checks and detecting disk
> failures, distributing data across disks as per the configured policy,
> collecting performance statistics.
> Attached Design Doc gives an overview of the proposed class diagram.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
