hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Bharath Mundlapudi (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HDFS-1061) Memory footprint optimization for INodeFile object.
Date Mon, 22 Mar 2010 23:07:27 GMT

    [ https://issues.apache.org/jira/browse/HDFS-1061?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12848428#action_12848428
] 

Bharath Mundlapudi commented on HDFS-1061:
------------------------------------------

INodeFile objects are proportional to the number of files created. By eliminating short data
field, we get 8 bytes of saving due to byte code alignment on a 64bit JVM. We can have 40
bits for preferred block Size and 24 bit for block replication. We probably don't need 24
bit for replication. In future, we could use these extra bits for storing some other meta
data without creating new data types in this memory structure.

> Memory footprint optimization for INodeFile object. 
> ----------------------------------------------------
>
>                 Key: HDFS-1061
>                 URL: https://issues.apache.org/jira/browse/HDFS-1061
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: name-node
>    Affects Versions: 0.22.0
>            Reporter: Bharath Mundlapudi
>            Priority: Minor
>             Fix For: 0.22.0
>
>
> I am proposing a footprint optimization to merge blockReplication and preferredBlockSize
fields into one 'long header' field in INodeFile class. This saves 8 bytes per INodeFile object
on a 64 bit JVM. This memory optimization is transparent and changes are very minimal.

-- 
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