[
https://issues.apache.org/jira/browse/HADOOP-8365?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13423262#comment-13423262
]
Todd Lipcon commented on HADOOP-8365:
-
I think you're misinterpreting HDFS-3731: the bug there is that data which was
in-progress during the upgrade will be lost in the upgraded cluster. But
without durable sync, all the blocks being written would have been lost anyway.
That is to say, the bug makes durable sync just regress back to the old
non-durable behavior. Anyone who is not using the feature would have lost the
same amount of data.
Or am I the one misinterpreting?
> Add flag to disable durable sync
>
>
> Key: HADOOP-8365
> URL: https://issues.apache.org/jira/browse/HADOOP-8365
> Project: Hadoop Common
> Issue Type: Improvement
>Affects Versions: 1.1.0
>Reporter: Eli Collins
>Assignee: Eli Collins
>Priority: Blocker
> Fix For: 1.1.0
>
> Attachments: hadoop-8365.txt, hadoop-8365.txt
>
>
> Per HADOOP-8230 there's a request for a flag to disable the sync code paths
> that dfs.support.append used to enable. The sync method itself will still be
> available and have a broken implementation as that was the behavior before
> HADOOP-8230. This config flag should default to false as the primary
> motivation for HADOOP-8230 is so HBase works out-of-the-box with Hadoop 1.1.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira