hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Rakesh R (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-10697) Erasure Code(6:3) can not work, using get fs shell when 2 datanode down
Date Wed, 27 Jul 2016 14:24:20 GMT

    [ https://issues.apache.org/jira/browse/HDFS-10697?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15395755#comment-15395755
] 

Rakesh R commented on HDFS-10697:
---------------------------------

I think this could be a similar issue HDFS-10586 where (number of datanode failures <=
parity number of datanodes). Could you tell me some more details:-
# file block locations of this file to see the blocks are placed in unique data nodes to ensure
that the dead datanodes have only one block each.
# the length of the file.
# please analyse the client and datanode logs to understand the failure reason.

> Erasure Code(6:3) can not work, using get fs shell when 2 datanode down
> -----------------------------------------------------------------------
>
>                 Key: HDFS-10697
>                 URL: https://issues.apache.org/jira/browse/HDFS-10697
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: erasure-coding
>    Affects Versions: 3.0.0-alpha2
>         Environment: 9 datanodes 2 HA name node and Yarn
>            Reporter: Alberic Liu
>
> I can put file into EC folder(6:3), and get the file successfully.
> But when I shutdown one datanode, I still can get the file, using hdfs dfs -get.
> But when I shutdown two datanodes, when i use hdfs dfs -get, I get following error:
> get: 4 missing blocks, the stripe is: Offset=786432, length=65536, fetchedChunksNum=5,
missingChunksNum=4; locatedBlocks is: LocatedBlocks{



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

---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org


Mime
View raw message