hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Todd Lipcon (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-1974) HA: Introduce active and standby states to the namenode
Date Wed, 20 Jul 2011 06:05:58 GMT

    [ https://issues.apache.org/jira/browse/HDFS-1974?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13068163#comment-13068163
] 

Todd Lipcon commented on HDFS-1974:
-----------------------------------

Taking a quick look at the patch. A few thoughts:
- rather than adding more things to NN, would it be possible to get the state management in
a separate class? NameNode.java already has a lot of "stuff" in it, which can make it hard
to mock, etc.
- I see ACTIVE_STATE and STANDBY_STATE are static variables. Does that imply that these State
classes will have no member variables? If they do have member variables, it will make testing
very difficult, since we need multiple NNs in a Minicluster
- I've been thinking about NameNode.java a bit with regard to HA: I think it might make this
work a little cleaner if we were to do a bit of refactoring first. I'm imagining it would
be broken up so that NameNode itself becomes just a container class, which contains NameNodeClientService
(implements ClientProtocol), NameNodeDatanodeService (implements DatanodeProtocol), NameNodeHttpService
(starts web server), etc. Then NameNode itself would do state management and transitions,
and move the actual RPC handling code out to other spots. What do you think?


> HA: Introduce active and standby states to the namenode
> -------------------------------------------------------
>
>                 Key: HDFS-1974
>                 URL: https://issues.apache.org/jira/browse/HDFS-1974
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>          Components: name-node
>            Reporter: Suresh Srinivas
>            Assignee: Suresh Srinivas
>         Attachments: HDFS-1974.patch
>
>
> Currently namenode supports active, secondary and backup roles. To support namenode high
availability, active and standby states are needed. Note that this is different from the existing
notion of namenode role, where a namenode cannot transition from one role to the other.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message