Mailing-List: contact commits-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cassandra.apache.org
Date: Thu, 31 Jul 2014 19:34:40 +0000 (UTC)
From: "Brandon Williams (JIRA)" <jira@apache.org>
To: commits@cassandra.apache.org
Message-ID: <JIRA.12730565.1406659711483.89037.1406835280189@arcas>
In-Reply-To: <JIRA.12730565.1406659711483@arcas>
References: <JIRA.12730565.1406659711483@arcas>
Subject: [jira] [Updated] (CASSANDRA-7638) Revisit GCInspector
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


     [ https://issues.apache.org/jira/browse/CASSANDRA-7638?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Brandon Williams updated CASSANDRA-7638:
----------------------------------------

    Attachment: 7638.txt

Here's a patch that basically reverts CASSANDRA-2868.  I tested this on 1.7u60, again setting the GCI interval to 1ms on both a patched and vanilla machine, and there is no difference in resident memory usage between them after a few hours.  I can see just from startup that this is catching a CMS that we miss with the CASSANDRA-2868 way.

> Revisit GCInspector
> -------------------
>
>                 Key: CASSANDRA-7638
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7638
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Brandon Williams
>            Assignee: Brandon Williams
>            Priority: Minor
>             Fix For: 2.0.10
>
>         Attachments: 7638.txt
>
>
> In CASSANDRA-2868 we had to change the api that GCI uses to avoid the native memory leak, but this caused GCI to be less reliable and more 'best effort' than before where it was 100% reliable.  Let's revisit this and see if the native memory leak is fixed in java7.


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