[
https://issues.apache.org/jira/browse/HBASE-21084?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16587502#comment-16587502
]
Toshihiro Suzuki commented on HBASE-21084:
------------------------------------------
We won't online the split parent region for the original table. We will online
the region for the cloned table. The Description may be a little bit confusing.
I'll rewrite that.
In this case, we take a snapshot right after splitting a region and before
CatalogJanitor runs. So the split parent region is included in the snapshot.
When cloning the snapshot, we should restore this split parent region as a
region with a SPLIT state. But currently we will restore the region as a online
region.
> The split parent region becomes online after cloning a snapshot
> ---------------------------------------------------------------
>
> Key: HBASE-21084
> URL: https://issues.apache.org/jira/browse/HBASE-21084
> Project: HBase
> Issue Type: Bug
> Reporter: Toshihiro Suzuki
> Assignee: Toshihiro Suzuki
> Priority: Major
> Attachments: HBASE-21084.master.001.patch
>
>
> Investigating HBASE-21015, I found another issue. It seems like after
> HBASE-20881, the split parent region becomes online after cloning a snapshot.
> Steps to reproduce are as follows, which is the same as the steps in
> HBASE-21015:
> 1. Create a table
> {code}
> create "test", "cf"
> {code}
> 2. Put some data into the table
> {code}
> (0...2000).each{|i| put "test", "row#{i}", "cf:col", "val"}
> {code}
> 3. Split the table
> {code}
> split "test"
> {code}
> 4. Take a snapshot of the table before CatalogJanitor cleans up the split
> parent region
> {code}
> snapshot "test", "snap"
> {code}
> 5. Clone the snapshot
> {code}
> clone_snapshot "snap", "cloned_table"
> {code}
> After following the above steps, the split parent region becomes online.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
