[
https://issues.apache.org/jira/browse/HADOOP-2447?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12553408
]
dhruba borthakur commented on HADOOP-2447:
------------------------------------------
Thanks Allen for your comments. I agree that tuning based upon heap size is
tempting, but is a more complex task. I would like to do it, but it requires
more time and would not be ready for 0.16. Instead, my proposal is to have a
*single* config parameter called "dfs.max.objects". It limits the total number
of files + directories + blocks that the system can have. Would that be useful
as a first stepping stone to the final goal?
I like the idea of having a tool that estimates the optimal setting of
"dfs.max.objects" based on a specified heap size. That tool can be in the form
of a equation (or spread-sheet) that we can provide.
> HDFS should be capable of limiting the total number of inodes in the system
> ---------------------------------------------------------------------------
>
> Key: HADOOP-2447
> URL: https://issues.apache.org/jira/browse/HADOOP-2447
> Project: Hadoop
> Issue Type: New Feature
> Reporter: Sameer Paranjpye
> Assignee: dhruba borthakur
> Fix For: 0.16.0
>
> Attachments: fileLimit.patch
>
>
> The HDFS Namenode should be capable of limiting the total number of Inodes
> (files + directories). The can be done through a config variable, settable in
> hadoop-site.xml. The default should be no limit.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
