hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Yoram Arnon (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-993) Namenode does not need to store any data node info persistently.
Date Thu, 15 Mar 2007 23:58:09 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-993?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12481382
] 

Yoram Arnon commented on HADOOP-993:
------------------------------------

I don't understand the motivation. Is it to save space in the fsimage file? Is it simplicity?
Seems like we're losing potential goodness:
* namenode can know what to expect on startup
* in case of a misconfiguration of a datanode, namenode can tell which datanode "should" have
a certain storageid
* lastly - UI can report on missing datanodes, which proved immensely useful for administration

- not very strong points but then, where's the value in removing this information?

what would be good is to allow the admin to declare a datanode 'gone', where currently if
a node ever connects it stays on the list forever.


> Namenode does not need to store any data node info persistently.
> ----------------------------------------------------------------
>
>                 Key: HADOOP-993
>                 URL: https://issues.apache.org/jira/browse/HADOOP-993
>             Project: Hadoop
>          Issue Type: Improvement
>          Components: dfs
>            Reporter: Raghu Angadi
>         Assigned To: Sameer Paranjpye
>         Attachments: noDatanodesInFsimage.patch
>
>
> Namenode does not need to serialize datanode info. It will map datanode to storageID
when datanode register.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message