Mailing-List: contact notifications-help@accumulo.apache.org; run by ezmlm
Precedence: bulk
Reply-To: jira@apache.org
Date: Tue, 15 Oct 2013 13:21:46 +0000 (UTC)
From: "Eric Newton (JIRA)" <jira@apache.org>
To: notifications@accumulo.apache.org
Message-ID: <JIRA.12673272.1381426889958.65958.1381843306773@arcas>
In-Reply-To: <JIRA.12673272.1381426889958@arcas>
References: <JIRA.12673272.1381426889958@arcas>
Subject: [jira] [Commented] (ACCUMULO-1770) out of memory error on very long
 running tablet server
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/ACCUMULO-1770?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13795177#comment-13795177 ] 

Eric Newton commented on ACCUMULO-1770:
---------------------------------------

I ran a tserver with the native map for 3 days.  I've attached the image.  Memory usage grows, but remains under the 4G limit I would expect (3G java heap, 1G native in-memory map).  This is consistent with a growing java heap that remains relatively unused due to fast GC of the New Generation.  Eventually the entire java heap is resident, but does not continue to grow.  I'll let it continue to spin to verify that it does stay under the expected resident memory limits.


> out of memory error on very long running tablet server
> ------------------------------------------------------
>
>                 Key: ACCUMULO-1770
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-1770
>             Project: Accumulo
>          Issue Type: Bug
>          Components: tserver
>            Reporter: Eric Newton
>            Assignee: Eric Newton
>         Attachments: FragmentTest.java, javamap.png, memory-usage.png, nativemap.png, three-day-tserver.png
>
>
> On a large cluster it was noticed that a few of the tablet servers had been pushed into swap.  This didn't effect the performance of the server until it ran out of memory, and the process was killed.  The gc reports in the debug log showed the system had plenty of heap space for the JVM.  The number of threads in the server were not excessive (dozens).  This cluster ingests some large values (megabytes).  The tablet server had been up for a month prior to running out of memory.  MALLOC_ARENA_MAX had already been set to 1.
> * Investigate the effect of fragmentation on memory usage for large value inserts.


--
This message was sent by Atlassian JIRA
(v6.1#6144)