hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "sam liu (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-7624) TestFileAppendRestart hardcode block size without considering native OS
Date Sat, 17 Jan 2015 01:41:34 GMT

    [ https://issues.apache.org/jira/browse/HDFS-7624?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14281120#comment-14281120
] 

sam liu commented on HDFS-7624:
-------------------------------

Hi Konstantin, 

The "power platform" means IBM power system/server and the Linux OS for power has some differences
from Linux OS for x86 system. In the test, 'BLOCK_SIZE = 4096' does not mean the HDFS block
size is 4096, but mean the operating system page size is 4096. The page size of operating
systems are different, so we should not hardcode the page size in the test. The correct way
is to get the page size using 'NativeIO.POSIX.getCacheManipulator().getOperatingSystemPageSize()'
instead.

> TestFileAppendRestart hardcode block size without considering native OS
> -----------------------------------------------------------------------
>
>                 Key: HDFS-7624
>                 URL: https://issues.apache.org/jira/browse/HDFS-7624
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: test
>            Reporter: sam liu
>            Assignee: sam liu
>         Attachments: HDFS-7624.001.patch
>
>
> TestFileAppendRestart hardcode block size with 'BLOCK_SIZE = 4096', however it's incorrect
on some platforms. For example, on power platform, the correct value is 65536.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message