[
https://issues.apache.org/jira/browse/HDFS-2124?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13059856#comment-13059856
]
Justin Joseph commented on HDFS-2124:
-------------------------------------
1) I understand that many design decisions match with HDFS-1623. The major
difference is that HDFS-1623 adopts Shared Storage approach where as this
solution uses Backup Namenode for data synchronization from Active to Standby.
(Backup Namenode approach is mentioned as the future direction in HDFS-1623).
Also, there are some differences in the way clients handle the Namenode
failover.
To avoid any conflict / duplication with HDFS-1623, I will create sub-tasks for
each of the logically independent parts in our implementation. It will help the
community in reviewing & committing those parts which compliment the
implementation done for HDFS-1623.
2) Any comments on creating an additional patch for a 0.20.* version ?
> Namenode HA using Backup Namenode as Hot Standby
> ------------------------------------------------
>
> Key: HDFS-2124
> URL: https://issues.apache.org/jira/browse/HDFS-2124
> Project: Hadoop HDFS
> Issue Type: New Feature
> Components: data-node, hdfs client, name-node
> Affects Versions: 0.20.2, 0.21.0
> Reporter: Justin Joseph
> Attachments: Namenode HA using Backup Namenode as Hot Standby.pdf
>
>
> This is to share our experiences on building Automatic, Hot High Availability
> for Namenode, leveraging the existing Backup Namenode.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
