You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-issues@hadoop.apache.org by "Xiao Chen (JIRA)" <ji...@apache.org> on 2018/05/31 22:07:00 UTC
[jira] [Commented] (HADOOP-15507) Add MapReduce counters about EC
bytes read
[ https://issues.apache.org/jira/browse/HADOOP-15507?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16497247#comment-16497247 ]
Xiao Chen commented on HADOOP-15507:
------------------------------------
(Write counters are calculated at FSDataOutputStream, and it's pretty difficult to bring HDFS information from DFSOutputStream up here. So for this Jira the proposal is to only do read stats)
> Add MapReduce counters about EC bytes read
> ------------------------------------------
>
> Key: HADOOP-15507
> URL: https://issues.apache.org/jira/browse/HADOOP-15507
> Project: Hadoop Common
> Issue Type: Improvement
> Reporter: Xiao Chen
> Assignee: Xiao Chen
> Priority: Major
>
> HDFS has added Erasure Coding support in HDFS-7285. There are HDFS level [ReadStatistics|https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/ReadStatistics.java] so from DFSClient we can know how much reads are EC/replication.
> In order for users to have a better view of how much of their workload is impacted by EC, we can expose EC read bytes to File System Counters, and to MapReduce's job counters. This way, end users can tell from MR jobs directly.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org