[
https://issues.apache.org/jira/browse/HADOOP-3006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12578006#action_12578006
]
Raghu Angadi commented on HADOOP-3006:
--------------------------------------
This affects pipelining while replicating data, and not when client writes the
data first time. That explains why benchmark performance was not affected.
> DataNode sends wrong length in header while pipelining.
> -------------------------------------------------------
>
> Key: HADOOP-3006
> URL: https://issues.apache.org/jira/browse/HADOOP-3006
> Project: Hadoop Core
> Issue Type: Bug
> Components: dfs
> Affects Versions: 0.17.0
> Reporter: Raghu Angadi
> Assignee: Raghu Angadi
> Fix For: 0.17.0
>
>
> Bug introduced by HADOOP-2758. This part is removed in HADOOP-1702, but 1702
> is not yet ready. I will submit a patch.
> The first datanode in the pipeline receives data with non-interleaving
> checksum and sends interleaving checksum to mirror (this mismatch is
> temporary till HADOOP-1702). The packet size differs in these two cases and
> DataNode did not correct the packet size in the header.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
