You are viewing a plain text version of this content. The canonical link for it is here.

Posted to hdfs-dev@hadoop.apache.org by "guodongdong (JIRA)" <ji...@apache.org> on 2014/03/25 08:25:42 UTC

[jira] [Created] (HDFS-6154) Improve the speed of saveNameSpace，making HDFS restart and checkPoint faster

guodongdong created HDFS-6154:
---------------------------------

             Summary: Improve the speed of saveNameSpace，making HDFS restart and checkPoint faster
                 Key: HDFS-6154
                 URL: https://issues.apache.org/jira/browse/HDFS-6154
             Project: Hadoop HDFS
          Issue Type: Improvement
    Affects Versions: 2.3.0
            Reporter: guodongdong


There are two stage In namenode savenamespace,  serializing INode, calculate MD5 and write to disk.  Now, two stage is doing serially, In this improvement, it is doing  parallel, one thread do serializing INode, other thread do calculating MD5 and writing to disk, it double speed of savenamespace, Detail is show in table:

Testing environment:
  only test namenode savenamespace, dfsadmin -saveNamespace
    machine: 144GB, Intel(R) Xeon(R) CPU  E5645  @ 2.40GHz, 12 cpu, Raid 5 SAS Disk,  jdk 1.7.0
 
||image size||before optimizing||after optimizing ||
|1.2GB|22sec|11sec|
|4.3GB|66sec|36sec|
|22GB|406sec|250sec|



--
This message was sent by Atlassian JIRA
(v6.2#6252)