Mailing-List: contact accumulo-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: accumulo-dev@incubator.apache.org
Received-SPF: pass (nike.apache.org: domain of wslacum@gmail.com designates
 209.85.214.175 as permitted sender)
MIME-Version: 1.0
In-Reply-To: 
 <87230345.36436.1332253779003.JavaMail.tomcat@hel.zones.apache.org>
References: 
 <87230345.36436.1332253779003.JavaMail.tomcat@hel.zones.apache.org>
Date: Tue, 20 Mar 2012 08:12:25 -0700
Message-ID: 
 <CALs3O6hkQGkVnpUj_JBL8n=MpmOmeqGd84Fhe6Vbfnd-d5fgnA@mail.gmail.com>
Subject: Re: [jira] [Created] (ACCUMULO-474) wikisearch against an unindexed
 field causes tablet servers to run out of memory, or stop-the-world-gc
From: William Slacum <wslacum@gmail.com>
To: accumulo-dev@incubator.apache.org
Content-Type: multipart/alternative; boundary=f46d0447f27439713704bbae1a94

--f46d0447f27439713704bbae1a94
Content-Type: text/plain; charset=ISO-8859-1

Another thing to consider is that the document payloads are contained in
the 'd' column family. During an optimized query using the
BooleanLogicIterator, the 'd' column family is never hit by a scan.
However, on a full scan, the 'd' column family will be thought of as a
single event by the EvaluatingIterator, so it will try to glob together all
document payloads into a single multimap object.

On Tue, Mar 20, 2012 at 7:29 AM, Eric Newton (Created) (JIRA) <
jira@apache.org> wrote:

> wikisearch against an unindexed field causes tablet servers to run out of
> memory, or stop-the-world-gc
>
> ------------------------------------------------------------------------------------------------------
>
>                 Key: ACCUMULO-474
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-474
>             Project: Accumulo
>          Issue Type: Improvement
>          Components: wikisearch
>         Environment: Search on "XTEXT == 'old' and XTEXT == 'man' and
> XTEXT == 'sea'"
>            Reporter: Eric Newton
>            Assignee: Eric Newton
>            Priority: Minor
>             Fix For: 1.4.1
>
>
> Looking at a memory dump of a failing server, it looks like 50% of memory
> comes from ColumnVisibility objects.
>
> --
> This message is automatically generated by JIRA.
> If you think it was sent incorrectly, please contact your JIRA
> administrators:
> https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
> For more information on JIRA, see: http://www.atlassian.com/software/jira
>
>
>

--f46d0447f27439713704bbae1a94--