Mailing-List: contact user-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@hadoop.apache.org
Received-SPF: pass (nike.apache.org: local policy includes SPF record at
 spf.trusted-forwarder.org)
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
  s=s1024; d=ymail.com;
  h=X-YMail-OSG:Received:X-Rocket-MIMEInfo:X-Mailer:References:Message-ID:Date:From:Reply-To:Subject:To:In-Reply-To:MIME-Version:Content-Type;
  b=d6NG5m4eiVHHTUQ4Uzmu3OyULtz3eAkAGT6cwZiosdPaGzEaPik11pEuJZpETXhqsuVSV7XTNzSxgomyHEOJhKRFWd4SS5D4CyVYKwOhT3m8ByeK9FONy5BbS4mZth+B/ikcKx5blKltc4T4gXFDrt6UaRQUz4BOA6PYzxN7NW8=;
References: <99B9D7A61BCB4F0999FC9EEEA1A15087@gmail.com>
Message-ID: <1382637850.14709.YahooMailNeo@web141203.mail.bf1.yahoo.com>
Date: Thu, 24 Oct 2013 11:04:10 -0700 (PDT)
From: Ravi Prakash <ravihoo@ymail.com>
Reply-To: Ravi Prakash <ravihoo@ymail.com>
Subject: Re: dynamically resizing the Hadoop cluster?
To: "user@hadoop.apache.org" <user@hadoop.apache.org>
In-Reply-To: <99B9D7A61BCB4F0999FC9EEEA1A15087@gmail.com>
MIME-Version: 1.0
Content-Type: multipart/alternative;
 boundary="1123101620-1229506293-1382637850=:14709"

--1123101620-1229506293-1382637850=:14709
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

Hi Nan!=0A=0AUsually nodes are decommissioned slowly over some period of ti=
me so as not to disrupt the running jobs. When a node is decommissioned, th=
e NameNode must re-replicate all under-replicated blocks. Rather than sudde=
nly remove half the nodes, you might want to take a few nodes offline at a =
time. Hadoop should be able to handle rescheduling tasks on nodes no longer=
 available (even without speculative execution. Speculative execution is fo=
r something else). =0A=0A=0AHTH=0ARavi=0A=0A=0A=0A=0AOn Wednesday, October =
23, 2013 10:26 PM, Nan Zhu <zhunansjtu@gmail.com> wrote:=0A =0AHi, all=0A=
=0AI=E2=80=99m running a Hadoop cluster on AWS EC2,=C2=A0=0A=0AI would like=
 to dynamically resizing the cluster so as to reduce the cost, is there any=
 solution to achieve this?=C2=A0=0A=0AE.g. I would like to cut the cluster =
size with a half, is it safe to just shutdown the instances (if some tasks =
are just running on them, can I rely on the speculative execution to re-run=
 them on the other nodes?)=0A=0AI cannot use EMR, since I=E2=80=99m running=
 a customized version of Hadoop=C2=A0=0A=0ABest,=0A=0A--=C2=A0=0ANan Zhu=0A=
School of Computer Science,=0AMcGill University
--1123101620-1229506293-1382637850=:14709
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable

<html><body><div style=3D"color:#000; background-color:#fff; font-family:He=
lveticaNeue, Helvetica Neue, Helvetica, Arial, Lucida Grande, sans-serif;fo=
nt-size:10pt"><div><span>Hi Nan!</span></div><div style=3D"color: rgb(0, 0,=
 0); font-size: 13.3333px; font-family: HelveticaNeue,Helvetica Neue,Helvet=
ica,Arial,Lucida Grande,sans-serif; background-color: transparent; font-sty=
le: normal;"><br></div><div style=3D"color: rgb(0, 0, 0); font-size: 13.333=
3px; font-family: HelveticaNeue,Helvetica Neue,Helvetica,Arial,Lucida Grand=
e,sans-serif; background-color: transparent; font-style: normal;">Usually n=
odes are decommissioned slowly over some period of time so as not to disrup=
t the running jobs. When a node is decommissioned, the NameNode must re-rep=
licate all under-replicated blocks. Rather than suddenly remove half the no=
des, you might want to take a few nodes offline at a time. Hadoop should be=
 able to handle rescheduling tasks on nodes no longer available (even
 without speculative execution. Speculative execution is for something else=
). <br></div><div style=3D"color: rgb(0, 0, 0); font-size: 13.3333px; font-=
family: HelveticaNeue,Helvetica Neue,Helvetica,Arial,Lucida Grande,sans-ser=
if; background-color: transparent; font-style: normal;"><br></div><div styl=
e=3D"color: rgb(0, 0, 0); font-size: 13.3333px; font-family: HelveticaNeue,=
Helvetica Neue,Helvetica,Arial,Lucida Grande,sans-serif; background-color: =
transparent; font-style: normal;">HTH</div><div style=3D"color: rgb(0, 0, 0=
); font-size: 13.3333px; font-family: HelveticaNeue,Helvetica Neue,Helvetic=
a,Arial,Lucida Grande,sans-serif; background-color: transparent; font-style=
: normal;">Ravi<br><span></span></div><div style=3D"display: block;" class=
=3D"yahoo_quoted"> <br> <br> <div style=3D"font-family: HelveticaNeue, Helv=
etica Neue, Helvetica, Arial, Lucida Grande, sans-serif; font-size: 10pt;">=
 <div style=3D"font-family: HelveticaNeue, Helvetica Neue, Helvetica, Arial=
, Lucida
 Grande, sans-serif; font-size: 12pt;"> <div dir=3D"ltr"> <font face=3D"Ari=
al" size=3D"2"> On Wednesday, October 23, 2013 10:26 PM, Nan Zhu &lt;zhunan=
sjtu@gmail.com&gt; wrote:<br> </font> </div>  <div class=3D"y_msg_container=
"><div id=3D"yiv4937801102">=0A                <div><div style=3D"font-size=
:12.800000190734863px;">Hi, all</div><div style=3D"font-size:12.80000019073=
4863px;"><br></div><div style=3D"font-size:12.800000190734863px;">I=E2=80=
=99m running a Hadoop cluster on AWS EC2,&nbsp;</div><div style=3D"font-siz=
e:12.800000190734863px;"><br></div><div style=3D"font-size:12.8000001907348=
63px;">I would like to dynamically resizing the cluster so as to reduce the=
 cost, is there any solution to achieve this?&nbsp;</div><div style=3D"font=
-size:12.800000190734863px;"><br></div><div style=3D"font-size:12.800000190=
734863px;">E.g. I would like to cut the cluster size with a half, is it saf=
e to just shutdown the instances (if some tasks are just running on them, c=
an I rely on the speculative execution to re-run them on the other nodes?)<=
/div><div style=3D"font-size:12.800000190734863px;"><br></div><div style=3D=
"font-size:12.800000190734863px;">I cannot use EMR, since I=E2=80=99m runni=
ng a customized version of Hadoop&nbsp;</div><div
 style=3D"font-size:12.800000190734863px;"><br></div><div style=3D"font-siz=
e:12.800000190734863px;">Best,</div></div><div style=3D"font-size:12.800000=
190734863px;"><br></div><div><div>--&nbsp;</div><div>Nan Zhu</div><div>Scho=
ol of Computer Science,</div><div>McGill University</div><div><br></div><di=
v><br></div></div>=0A</div><br><br></div>  </div> </div>  </div> </div></bo=
dy></html>
--1123101620-1229506293-1382637850=:14709--