[
https://issues.apache.org/jira/browse/HADOOP-1141?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12482984
]
Konstantin Shvachko commented on HADOOP-1141:
---------------------------------------------
Don't know what happened, but since the issue has been created, I'm going to
use it even though the Summary is not very informative.
Summary should read:
reportWrittenBlock() in DFSClient should retry
Description:
DFSOutputStream.internalClose() calls reportWrittenBlock().
If reportWrittenBlock() fails because of e.g. a SocketTimeoutException the
client will call
abandonBlock() and repeating of allocating and transferring the block from the
very beginning.
The client should retry sending reportWrittenBlock() since at that time it
knows the data-nodes
already have their copies.
> P
> -
>
> Key: HADOOP-1141
> URL: https://issues.apache.org/jira/browse/HADOOP-1141
> Project: Hadoop
> Issue Type: Bug
> Reporter: Konstantin Shvachko
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
