[
https://issues.apache.org/jira/browse/HBASE-6223?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13409222#comment-13409222
]
Jonathan Hsieh commented on HBASE-6223:
---------------------------------------
Motivation for HBASE-5360: We encountered a situation where a parent was marked
off-line in meta, where data still existed in HDFS, and where we actually
wanted to restore the data that was in the FS. I believe the old behavior was
to skip these regions with the assumption that daughter regions were good. In
this particular case, the daughter regions were not -- I believe one of the two
were missing, and we wanted an option to "revive" the parent region. This
would likely cause an overlap with the surviving daughter region -- but we
already had mechanisms to handle that (-fixHdfsOverlaps). The -fixSplitParents
option is should not normally be used (it is not part of -fixAll).
Motivation for HBASE-6173: We ran into a situation withe cluster that had a
large number of regions (~20k, for 50 nodes or so). The file system operations
for loading this data, even though it happened in parallel took a very long
time. This addition allowed us just to reduced the time for a report by only
focusing on a particular table.
> Document hbck improvements: HBASE-6173, HBASE-5360
> ---------------------------------------------------
>
> Key: HBASE-6223
> URL: https://issues.apache.org/jira/browse/HBASE-6223
> Project: HBase
> Issue Type: Task
> Components: documentation, hbck
> Reporter: Jimmy Xiang
>
> We had a couple hbck improvements recently: HBASE-6173 and HBASE-5360.
> We should document them. Especially, for HBASE-5360, it's something
> one normally doesn't do.
--
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
