Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Date: Thu, 12 Feb 2015 19:51:12 +0000 (UTC)
From: "Andrew Purtell (JIRA)" <jira@apache.org>
To: issues@hbase.apache.org
Message-ID: <JIRA.12752800.1415126758000.30763.1423770672782@Atlassian.JIRA>
In-Reply-To: <JIRA.12752800.1415126758000@Atlassian.JIRA>
References: <JIRA.12752800.1415126758000@Atlassian.JIRA>
 <JIRA.12752800.1415126758461@arcas>
Subject: [jira] [Comment Edited] (HBASE-12425) Document the phases of the
 split transaction
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HBASE-12425?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14318868#comment-14318868 ] 

Andrew Purtell edited comment on HBASE-12425 at 2/12/15 7:50 PM:
-----------------------------------------------------------------

Looks great. A couple of quick suggestions:

bq. After each flush or compaction finishes, a region split request is enqueued if the _region split policy_ determines that the region should be split into two.

"After each flush or compaction finishes, the amount of data stored in the region has changed. The regionserver consults the _region split policy_ to determine if the region has grown too large or should be split for another policy-specific reason. A region split request is enqueued if the policy recommends it."

Above
bq. Since all data files in HBase are immutable,
consider:

"Logically the process of splitting a region is simple. We find a suitable point in the keyspace of the region where we should divide the region in half, then split the region's data into two new regions at that point. The details of the process however are not simple. "

Finally, I suppose it's fine but I wonder if there is a way to explicitly set numbers in a numbered list. Otherwise an edit to the text might accidentally make the list numbering go out of sync with the diagram. 


was (Author: apurtell):
Looks great. A couple of quick suggestions:

bq. After each flush or compaction finishes, a region split request is enqueued if the _region split policy_ determines that the region should be split into two.
After each flush or compaction finishes, the amount of data stored in the region has changed. The regionserver consults the _region split policy_ to determine if the region has grown too large or should be split for another policy-specific reason. A region split request is enqueued if the policy recommends it.

Above
bq. Since all data files in HBase are immutable,
consider:

Logically the process of splitting a region is simple. We find a suitable point in the keyspace of the region where we should divide the region in half, then split the region's data into two new regions at that point. The details of the process however are not simple. 

Finally, I suppose it's fine but I wonder if there is a way to explicitly set numbers in a numbered list. Otherwise an edit to the text might accidentally make the list numbering go out of sync with the diagram. 

> Document the phases of the split transaction
> --------------------------------------------
>
>                 Key: HBASE-12425
>                 URL: https://issues.apache.org/jira/browse/HBASE-12425
>             Project: HBase
>          Issue Type: Sub-task
>          Components: documentation
>            Reporter: Andrew Purtell
>            Assignee: Misty Stanley-Jones
>             Fix For: 2.0.0
>
>         Attachments: HBASE-12425.patch, region_split_process.png
>
>
> See PDF document attached to parent issue


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)