[
https://issues.apache.org/jira/browse/HADOOP-2705?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12562726#action_12562726
]
Doug Cutting commented on HADOOP-2705:
--------------------------------------
I'm confused too. This is an incompatible, non-blocker, performance
improvement, filed long after the freeze. Why is it included in 0.16.0? I
interpreted Chris' last comment to say that the performance improvement offered
by this patch was actually moderate. In any case, the measured performance
improvement is limited to a single place (lzo) while io.file.buffer.size is
used widely throughout the system, and the impact of the change might thus be
great.
I don't mean to make too big of a deal about this, since anyone who suffers
from this change has an easy workaround, namely to reset the value back to 4k.
But I don't follow the process here.
> io.file.buffer.size should default to a value larger than 4k
> ------------------------------------------------------------
>
> Key: HADOOP-2705
> URL: https://issues.apache.org/jira/browse/HADOOP-2705
> Project: Hadoop Core
> Issue Type: Improvement
> Components: conf
> Reporter: Chris Douglas
> Assignee: Chris Douglas
> Priority: Minor
> Fix For: 0.16.0
>
> Attachments: 2705-0.patch
>
>
> Tests using HADOOP-2406 suggest that increasing this to 32k from 4k improves
> read times for block, lzo compressed SequenceFiles by over 40%; 32k is a
> relatively conservative bump.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
