Mailing-List: contact commits-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cassandra.apache.org
Date: Thu, 20 Mar 2014 23:21:48 +0000 (UTC)
From: "Benedict (JIRA)" <jira@apache.org>
To: commits@cassandra.apache.org
Message-ID: <JIRA.12694636.1392165226027.115121.1395357708890@arcas>
In-Reply-To: <JIRA.12694636.1392165226027@arcas>
References: <JIRA.12694636.1392165226027@arcas>
Subject: [jira] [Comment Edited] (CASSANDRA-6689) Partially Off Heap
 Memtables
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/CASSANDRA-6689?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13942490#comment-13942490 ] 

Benedict edited comment on CASSANDRA-6689 at 3/20/14 11:21 PM:
---------------------------------------------------------------

See my comment explaining [here|https://issues.apache.org/jira/browse/CASSANDRA-6694?focusedCommentId=13904708&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13904708]. It sets up for the off-heap changes, I suppose we could delay the refactor, but it's just more work since they're all targeted for 2.1, and it's not harmful to make it part of this refactor. Note the whole point of #1 was to make these sorts of refactor so the next steps were easier: abstracting key() is also completely unnecessary until we have off-heap, as are a number of other things.

bq. One more nit thing, in DecoratedKey.java there is no need to mark token() and key() explicitly "abstract", also token() is already defined in RingPosition so no need to declare it in DecoratedKey.

The latter is an artefact of merging the getToken() and token() into the same method, and I agree. Although there's no harm in either, happy to change them.


was (Author: benedict):
See my comment explaining [here|https://issues.apache.org/jira/browse/CASSANDRA-6694?focusedCommentId=13904708&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13904708]. It sets up for the off-heap changes, I suppose we could delay the refactor, but it's just more work since they're all targeted for 2.1, and it's not harmful to make it part of this refactor.

bq. One more nit thing, in DecoratedKey.java there is no need to mark token() and key() explicitly "abstract", also token() is already defined in RingPosition so no need to declare it in DecoratedKey.

The latter is an artefact of merging the getToken() and token() into the same method, and I agree. Although there's no harm in either, happy to change them.

> Partially Off Heap Memtables
> ----------------------------
>
>                 Key: CASSANDRA-6689
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-6689
>             Project: Cassandra
>          Issue Type: New Feature
>          Components: Core
>            Reporter: Benedict
>            Assignee: Benedict
>              Labels: performance
>             Fix For: 2.1 beta2
>
>         Attachments: CASSANDRA-6689-small-changes.patch
>
>
> Move the contents of ByteBuffers off-heap for records written to a memtable.
> (See comments for details)


--
This message was sent by Atlassian JIRA
(v6.2#6252)