[
https://issues.apache.org/jira/browse/HADOOP-1083?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12478963
]
Nigel Daley commented on HADOOP-1083:
-------------------------------------
+1. I have manually tested this patch. I now see replications taking place
right away for missing blocks. If I bring the missing node back on line,
replications stop and over replicated blocks are deleted. Very cool! It would
be nice to have a unit test for this case.
> Replication not occuring after cluster restart when datanodes missing
> ---------------------------------------------------------------------
>
> Key: HADOOP-1083
> URL: https://issues.apache.org/jira/browse/HADOOP-1083
> Project: Hadoop
> Issue Type: Bug
> Components: dfs
> Affects Versions: 0.11.2
> Reporter: Nigel Daley
> Assigned To: Hairong Kuang
> Priority: Blocker
> Fix For: 0.12.1
>
> Attachments: replicateBlock.patch
>
>
> When a cluster is restarted, but one or more datanodes do not start, the
> blocks on those missing datanodes are not replicated.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
