hadoop-hdfs-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Suresh Srinivas (Created) (JIRA)" <j...@apache.org>
Subject [jira] [Created] (HDFS-3047) DU usage in FSDataset should use a single thread, instead of thread per directory
Date Mon, 05 Mar 2012 19:25:57 GMT
DU usage in FSDataset should use a single thread, instead of thread per directory
---------------------------------------------------------------------------------

                 Key: HDFS-3047
                 URL: https://issues.apache.org/jira/browse/HDFS-3047
             Project: Hadoop HDFS
          Issue Type: Improvement
          Components: data-node, tools
    Affects Versions: 0.24.0
            Reporter: Suresh Srinivas


Currently there is a DU thread per volume and with federation a thread per block pool per
volume that tracks the DFS used space. In a large installs the number of threads started =
Number of FSVolumes * Number of block pools. This results unnecessarily huge number of threads.

I propose a single thread which performs DU for a set of directories.


--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message