Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Date: Tue, 23 Apr 2013 05:35:17 +0000 (UTC)
From: "stack (JIRA)" <jira@apache.org>
To: issues@hbase.apache.org
Message-ID: <JIRA.12643682.1366406070598.206406.1366695317163@arcas>
In-Reply-To: <JIRA.12643682.1366406070598@arcas>
References: <JIRA.12643682.1366406070598@arcas>
Subject: [jira] [Commented] (HBASE-8384) heap size computation for HStore is
 fishy
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HBASE-8384?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13638804#comment-13638804 ] 

stack commented on HBASE-8384:
------------------------------

Patch looks good.  Did you try locally passing -d32 (run on 32bit jvm) and it still passed for you?

bq. On a tangentially related note, if we can obtain size programmatically (the way we do it in test), and only need to do it once, I wonder if we should just do it and remove all the manually modifiable constants stuff.

Yeah, we could do the above; would just be a bit of research and comparing our values/validating our findings w/ the various attempts at sizeof in java.
                
> heap size computation for HStore is fishy
> -----------------------------------------
>
>                 Key: HBASE-8384
>                 URL: https://issues.apache.org/jira/browse/HBASE-8384
>             Project: HBase
>          Issue Type: Bug
>            Reporter: Sergey Shelukhin
>            Assignee: Sergey Shelukhin
>            Priority: Minor
>         Attachments: HBASE-8384-v0.patch
>
>
> There's something off with heap size computation for HStore. If one adds a long to HStore, and size of long to the computation, TestHeapSize passes both locally (for me) and in jenkins. If one adds an int and size of int, it passes locally but fails in Jenkins. Perhaps the numbers are already off and some sort of packing is taking/not taking place differently.
> On a tangentially related note, if we can obtain size programmatically (the way we do it in test), and only need to do it once, I wonder if we should just do it and remove all the manually modifiable constants stuff.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira