[
https://issues.apache.org/jira/browse/HBASE-10111?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13981871#comment-13981871
]
Esteban Gutierrez commented on HBASE-10111:
-------------------------------------------
[~tychang] yes, it was committed for 0.94.15
(https://svn.apache.org/repos/asf/hbase/branches/0.94@1549820) or via github:
https://github.com/apache/hbase/commit/8215c58511a964680e9842c34ce61356a4f24756
> Verify that a snapshot is not corrupted before restoring it
> -----------------------------------------------------------
>
> Key: HBASE-10111
> URL: https://issues.apache.org/jira/browse/HBASE-10111
> Project: HBase
> Issue Type: Bug
> Components: snapshots
> Affects Versions: 0.98.0, 0.96.0, 0.94.14
> Reporter: Matteo Bertozzi
> Assignee: Matteo Bertozzi
> Priority: Minor
> Fix For: 0.98.0, 0.96.1, 0.94.15
>
> Attachments: HBASE-10111-v0.patch, HBASE-10111-v1.patch
>
>
> To avoid assigning/opening regions with missing files, verify that the
> snapshot is not corrupted before restoring/cloning it.
> In 96 a corrupted file in a region is "not a problem" since the assignment
> will give up after awhile.
> In 94 having a corrupted file in a region means looping forever, on "enable",
> until manual intervention. (Easy way to test this is create a table, disable
> it, add a corrupted reference file and enable the table to start looping: you
> can use echo "foo" >
> aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa.aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa)
--
This message was sent by Atlassian JIRA
(v6.2#6252)
