You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@accumulo.apache.org by "Keith Turner (JIRA)" <ji...@apache.org> on 2012/12/20 17:19:12 UTC
[jira] [Commented] (ACCUMULO-919) delay file garbage collection to
survive name node failure
[ https://issues.apache.org/jira/browse/ACCUMULO-919?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13537119#comment-13537119 ]
Keith Turner commented on ACCUMULO-919:
---------------------------------------
Would enabling HDFS trash feature work just as well? I have never used it, the property is fs.trash.interval. Setting this to some multiple of the snapshot time may work.
> delay file garbage collection to survive name node failure
> ----------------------------------------------------------
>
> Key: ACCUMULO-919
> URL: https://issues.apache.org/jira/browse/ACCUMULO-919
> Project: Accumulo
> Issue Type: Bug
> Components: gc
> Reporter: Eric Newton
> Assignee: Eric Newton
> Priority: Blocker
> Fix For: 1.5.0
>
>
> Consider the catastrophic loss of the namenode.
> Assuming you have a previous snapshot, you can recover the namenode using another node back to that snapshot.
> However, the most recently deleted files will show as corrupt, because the slightly older state will point to blocks that are now gone.
> If the gc delayed file collection for a period of time longer than the age of a snapshot, accumulo could be recovered with the snapshot.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira