Return-Path: 
 <user-return-22433-apmail-cassandra-user-archive=cassandra.apache.org@cassandra.apache.org>
X-Original-To: apmail-cassandra-user-archive@www.apache.org
Delivered-To: apmail-cassandra-user-archive@www.apache.org
Received: from mail.apache.org (hermes.apache.org [140.211.11.3])
	by minotaur.apache.org (Postfix) with SMTP id 28D479C6B
	for <apmail-cassandra-user-archive@www.apache.org>;
 Fri, 18 Nov 2011 04:42:53 +0000 (UTC)
Received: (qmail 70682 invoked by uid 500); 18 Nov 2011 04:42:51 -0000
Delivered-To: apmail-cassandra-user-archive@cassandra.apache.org
Received: (qmail 70651 invoked by uid 500); 18 Nov 2011 04:42:50 -0000
Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
List-Help: <mailto:user-help@cassandra.apache.org>
List-Unsubscribe: <mailto:user-unsubscribe@cassandra.apache.org>
List-Post: <mailto:user@cassandra.apache.org>
List-Id: <user.cassandra.apache.org>
Reply-To: user@cassandra.apache.org
Delivered-To: mailing list user@cassandra.apache.org
Received: (qmail 70638 invoked by uid 99); 18 Nov 2011 04:42:50 -0000
Received: from nike.apache.org (HELO nike.apache.org) (192.87.106.230)
    by apache.org (qpsmtpd/0.29) with ESMTP; Fri, 18 Nov 2011 04:42:50 +0000
X-ASF-Spam-Status: No, hits=1.5 required=5.0
	tests=HS_INDEX_PARAM,HTML_MESSAGE,RCVD_IN_DNSWL_LOW,SPF_PASS
X-Spam-Check-By: apache.org
Received-SPF: pass (nike.apache.org: domain of tyler@datastax.com designates
 209.85.161.44 as permitted sender)
Received: from [209.85.161.44] (HELO mail-fx0-f44.google.com) (209.85.161.44)
    by apache.org (qpsmtpd/0.29) with ESMTP; Fri, 18 Nov 2011 04:42:44 +0000
Received: by faap16 with SMTP id p16so5791386faa.31
        for <user@cassandra.apache.org>; Thu, 17 Nov 2011 20:42:24 -0800 (PST)
MIME-Version: 1.0
Received: by 10.152.104.130 with SMTP id ge2mr871181lab.43.1321591343869; Thu,
 17 Nov 2011 20:42:23 -0800 (PST)
Received: by 10.152.19.102 with HTTP; Thu, 17 Nov 2011 20:42:23 -0800 (PST)
In-Reply-To: <80C374D0BB0E4F94ACEE1DD08B956EC8@gmail.com>
References: <80C374D0BB0E4F94ACEE1DD08B956EC8@gmail.com>
Date: Thu, 17 Nov 2011 22:42:23 -0600
Message-ID: 
 <CAAam9stJ6A1qjhvTjr=cHFZXp21urf6xi-4tuonrZ2AVV-rWUw@mail.gmail.com>
Subject: Re: mmap I/O and shared memory
From: Tyler Hobbs <tyler@datastax.com>
To: user@cassandra.apache.org
Content-Type: multipart/alternative; boundary=f46d0408d3c196efb404b1faf626
X-Virus-Checked: Checked by ClamAV on apache.org

--f46d0408d3c196efb404b1faf626
Content-Type: text/plain; charset=ISO-8859-1

This FAQ entry and the linked document provide a pretty good explanation:
http://wiki.apache.org/cassandra/FAQ#mmap

By the way, you should almost always turn off swap.

On Thu, Nov 17, 2011 at 1:16 AM, Jaesung Lee <ljsking@gmail.com> wrote:

>  I am running 7 nodes cassandra(v1.0.2) cluster.
> I am putting 20K rows per sec to the cluster.
> This cluster has 1 KS, 3CFs.
> Each CF has 4-5 secondary indices.
>
> After I'v run for 1 week, nodes use swap memory.
> I changed disk-access-mode to index_only or standard.
> I got strange memory results.
>
> using mmap:
> VIRT: 566g  RES: 36g  SHR:12g
> standard disk access mode
> VIRT:24.7g  RES: 24g  SHR:68m
>
>
> I allocated 24g memory for JVM heap.
>
> I have some questions about mmap.
> It is easy to analyze standard disk access mode's memory result.
>
> I know cassandra use huge virtual memory for mmap I/O and each mmaped
> addresses are mapped to indexed file not swap memory.
>
> But, I don't understand why cassandra use shared memory, if using mmap I/O.
>
> Are there some documents that explain this situation?
>
> --
> Jaesung Lee
> Sent with Sparrow <http://www.sparrowmailapp.com/?sig>
>
>


-- 
Tyler Hobbs
DataStax <http://datastax.com/>

--f46d0408d3c196efb404b1faf626
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

This FAQ entry and the linked document provide a pretty good explanation: <=
a href=3D"http://wiki.apache.org/cassandra/FAQ#mmap">http://wiki.apache.org=
/cassandra/FAQ#mmap</a><br><br>By the way, you should almost always turn of=
f swap.<br>
<br><div class=3D"gmail_quote">On Thu, Nov 17, 2011 at 1:16 AM, Jaesung Lee=
 <span dir=3D"ltr">&lt;<a href=3D"mailto:ljsking@gmail.com">ljsking@gmail.c=
om</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_quote" style=3D"marg=
in:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">

                <div>
                    I am running 7 nodes cassandra(v1.0.2) cluster.<div>I a=
m putting 20K rows per sec to the cluster.</div><div>This cluster has 1 KS,=
 3CFs.</div><div>Each CF has 4-5 secondary indices.</div><div><br></div>
<div>After I&#39;v run for 1 week, nodes use swap memory.</div><div>I chang=
ed disk-access-mode to index_only or standard.</div><div>I got strange memo=
ry results.</div><div><br></div><blockquote type=3D"cite" style=3D"border-l=
eft-style:solid;border-left-color:rgb(0, 33, 98);border-top-width:1px;borde=
r-right-width:1px;border-bottom-width:1px;border-left-width:1px;color:rgb(0=
, 33, 98);margin-left:0px;padding-left:10px;padding-right:0px;margin-right:=
0px">
<div>using mmap:<br><span style=3D"font-family:sans-serif">VIRT: 566g =A0RE=
S: 36g =A0SHR:12g</span><span style=3D"font-family:sans-serif"><br></span><=
span style=3D"font-family:sans-serif">standard disk access mode<br></span><=
span style=3D"font-family:sans-serif"><span style=3D"line-height:20px"><spa=
n style=3D"font-family:sans-serif;line-height:normal">VIRT:</span></span><f=
ont face=3D"sans-serif"><span style=3D"line-height:normal">24.7g =A0</span>=
</font><span style=3D"font-family:sans-serif;line-height:normal">RES:=A0</s=
pan><span style=3D"font-family:sans-serif;line-height:normal">24g =A0</span=
><span style=3D"font-family:sans-serif;line-height:normal">SHR:</span><span=
 style=3D"font-family:sans-serif;line-height:normal">68m</span></span></div=
>
</blockquote><div><span style=3D"font-family:sans-serif"><span style=3D"lin=
e-height:20px"><span style=3D"font-family:sans-serif;line-height:normal"><b=
r></span></span></span></div><div><font face=3D"sans-serif">I allocated 24g=
 memory for JVM heap.</font></div>
<div><br></div><div>I have some questions about mmap.</div><div>It=A0is eas=
y to=A0analyze standard disk access mode&#39;s memory result.</div><div><br=
></div><div>I know cassandra use huge virtual memory for mmap I/O and each =
mmaped addresses are mapped to indexed file not swap memory.</div>
<div><br></div><div>But, I don&#39;t understand why cassandra use shared me=
mory, if using mmap I/O.</div><div><br></div><div>Are there some documents =
that=A0explain this situation?</div><span class=3D"HOEnZb"><font color=3D"#=
888888">
                </font></span></div><span class=3D"HOEnZb"><font color=3D"#=
888888">
                <div><div><br></div>--=A0<br>Jaesung Lee<br>Sent with <a hr=
ef=3D"http://www.sparrowmailapp.com/?sig" target=3D"_blank">Sparrow</a><div=
><br></div></div>
            </font></span></blockquote></div><br><br clear=3D"all"><br>-- <=
br><font color=3D"#888888">Tyler Hobbs<span></span><br>
<a href=3D"http://datastax.com/" target=3D"_blank">DataStax</a><br></font><=
br>

--f46d0408d3c196efb404b1faf626--