[
https://issues.apache.org/jira/browse/HDFS-9676?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15111958#comment-15111958
]
Vinayakumar B edited comment on HDFS-9676 at 1/22/16 5:39 AM:
--------------------------------------------------------------
Can you check namenode logs?
In fact this seems to be some cluster configuration issue, as this is the basic
operation which is not working for you.
May be we can take this on mailing list? If we find any issue, then you can
report issue in Jira.
was (Author: vinayrpet):
Can you check namenode logs?
> “hdfs dfs -put”,the return value is 0, but the size of $hdfsPath is 0.
> ----------------------------------------------------------------------
>
> Key: HDFS-9676
> URL: https://issues.apache.org/jira/browse/HDFS-9676
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: auto-failover, datanode, hdfs-client
> Affects Versions: 2.5.0
> Environment: Linux version 3.0.101
> 2 namenode, 54 datanode
> Reporter: DuRuiling
> Assignee: DuRuiling
> Original Estimate: 168h
> Remaining Estimate: 168h
>
> I find the bug when I backup about 240T data from local file system to HDFS,
> I use command “hdfs dfs -D dfs.replication=3 -put $localPath $hdfsPath” to
> backup each file one by one, the return value of this command is 0, then
> illustrate the command execute right, I find file path is right, but the size
> of $hdfsPath is 0 and the size of $loaclPath is non-zero. I reproduce the bug
> when I backup data again.
> I write a bash shell to backup all data, according to the $? of the command
> to judge the file is backup success and record a detailed log.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
