hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Daryn Sharp (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-7435) PB encoding of block reports is very inefficient
Date Mon, 24 Nov 2014 23:28:13 GMT

    [ https://issues.apache.org/jira/browse/HDFS-7435?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14223757#comment-14223757
] 

Daryn Sharp commented on HDFS-7435:
-----------------------------------

While I agree it's a questionably nice to have feature - see below - decoding the block report
into a contiguous or non-contiguous array is an implementation detail that can be changed
in another jira w/o updating the layout version.

This patch simply undoes the damage caused by moving from Writables to PBs.   I changed it
internally from long[] -> dynamically growing ArrayList<Long> -> PB -> dynamically
growing ArrayList<Long> -> long[]; to simply fixed long[] -> PB -> fixed long[].
  Further changes to segment the list in ways not previously done are feature creep.

If 20-30MB is going to cause a promotion failure in a namenode servicing a hundred millions
of blocks - it's already game over.  2.x easily generates over 1GB garbage/sec at a mere ~20k
ops/sec.  That's an average of 54MB/request.  This is why I'm submitting GC friendly changes.

> PB encoding of block reports is very inefficient
> ------------------------------------------------
>
>                 Key: HDFS-7435
>                 URL: https://issues.apache.org/jira/browse/HDFS-7435
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: datanode, namenode
>    Affects Versions: 2.0.0-alpha, 3.0.0
>            Reporter: Daryn Sharp
>            Assignee: Daryn Sharp
>            Priority: Critical
>         Attachments: HDFS-7435.patch
>
>
> Block reports are encoded as a PB repeating long.  Repeating fields use an {{ArrayList}}
with default capacity of 10.  A block report containing tens or hundreds of thousand of longs
(3 for each replica) is extremely expensive since the {{ArrayList}} must realloc many times.
 Also, decoding repeating fields will box the primitive longs which must then be unboxed.



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

Mime
View raw message