apache
/
hadoop
mirror de https://github.com/apache/hadoop.git


			
				
					
						
						
							123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936
							<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html>
<head>
<META http-equiv="Content-Type" content="text/html; charset=UTF-8">
<meta content="Apache Forrest" name="Generator">
<meta name="Forrest-version" content="0.8">
<meta name="Forrest-skin-name" content="pelt">
<title>Hadoop Cluster Setup</title>
<link type="text/css" href="skin/basic.css" rel="stylesheet">
<link media="screen" type="text/css" href="skin/screen.css" rel="stylesheet">
<link media="print" type="text/css" href="skin/print.css" rel="stylesheet">
<link type="text/css" href="skin/profile.css" rel="stylesheet">
<script src="skin/getBlank.js" language="javascript" type="text/javascript"></script><script src="skin/getMenu.js" language="javascript" type="text/javascript"></script><script src="skin/fontsize.js" language="javascript" type="text/javascript"></script>
<link rel="shortcut icon" href="images/favicon.ico">
</head>
<body onload="init()">
<script type="text/javascript">ndeSetTextSize();</script>
<div id="top">
<!--+
    |breadtrail
    +-->
<div class="breadtrail">
<a href="http://www.apache.org/">Apache</a> &gt; <a href="http://hadoop.apache.org/">Hadoop</a> &gt; <a href="http://hadoop.apache.org/core/">Core</a><script src="skin/breadcrumbs.js" language="JavaScript" type="text/javascript"></script>
</div>
<!--+
    |header
    +-->
<div class="header">
<!--+
    |start group logo
    +-->
<div class="grouplogo">
<a href="http://hadoop.apache.org/"><img class="logoImage" alt="Hadoop" src="images/hadoop-logo.jpg" title="Apache Hadoop"></a>
</div>
<!--+
    |end group logo
    +-->
<!--+
    |start Project Logo
    +-->
<div class="projectlogo">
<a href="http://hadoop.apache.org/core/"><img class="logoImage" alt="Hadoop" src="images/core-logo.gif" title="Scalable Computing Platform"></a>
</div>
<!--+
    |end Project Logo
    +-->
<!--+
    |start Search
    +-->
<div class="searchbox">
<form action="http://www.google.com/search" method="get" class="roundtopsmall">
<input value="hadoop.apache.org" name="sitesearch" type="hidden"><input onFocus="getBlank (this, 'Search the site with google');" size="25" name="q" id="query" type="text" value="Search the site with google">&nbsp; 
                    <input name="Search" value="Search" type="submit">
</form>
</div>
<!--+
    |end search
    +-->
<!--+
    |start Tabs
    +-->
<ul id="tabs">
<li>
<a class="unselected" href="http://hadoop.apache.org/core/">Project</a>
</li>
<li>
<a class="unselected" href="http://wiki.apache.org/hadoop">Wiki</a>
</li>
<li class="current">
<a class="selected" href="index.html">Hadoop 0.19 Documentation</a>
</li>
</ul>
<!--+
    |end Tabs
    +-->
</div>
</div>
<div id="main">
<div id="publishedStrip">
<!--+
    |start Subtabs
    +-->
<div id="level2tabs"></div>
<!--+
    |end Endtabs
    +-->
<script type="text/javascript"><!--
document.write("Last Published: " + document.lastModified);
//  --></script>
</div>
<!--+
    |breadtrail
    +-->
<div class="breadtrail">

             &nbsp;
           </div>
<!--+
    |start Menu, mainarea
    +-->
<!--+
    |start Menu
    +-->
<div id="menu">
<div onclick="SwitchMenu('menu_selected_1.1', 'skin/')" id="menu_selected_1.1Title" class="menutitle" style="background-image: url('skin/images/chapter_open.gif');">Documentation</div>
<div id="menu_selected_1.1" class="selectedmenuitemgroup" style="display: block;">
<div class="menuitem">
<a href="index.html">Overview</a>
</div>
<div class="menuitem">
<a href="quickstart.html">Hadoop Quick Start</a>
</div>
<div class="menupage">
<div class="menupagetitle">Hadoop Cluster Setup</div>
</div>
<div class="menuitem">
<a href="mapred_tutorial.html">Hadoop Map/Reduce Tutorial</a>
</div>
<div class="menuitem">
<a href="commands_manual.html">Hadoop Command Guide</a>
</div>
<div class="menuitem">
<a href="hdfs_shell.html">Hadoop FS Shell Guide</a>
</div>
<div class="menuitem">
<a href="distcp.html">Hadoop DistCp Guide</a>
</div>
<div class="menuitem">
<a href="native_libraries.html">Hadoop Native Libraries</a>
</div>
<div class="menuitem">
<a href="streaming.html">Hadoop Streaming</a>
</div>
<div class="menuitem">
<a href="hadoop_archives.html">Hadoop Archives</a>
</div>
<div class="menuitem">
<a href="hdfs_user_guide.html">HDFS User Guide</a>
</div>
<div class="menuitem">
<a href="hdfs_design.html">HDFS Architecture</a>
</div>
<div class="menuitem">
<a href="hdfs_permissions_guide.html">HDFS Admin Guide: Permissions</a>
</div>
<div class="menuitem">
<a href="hdfs_quota_admin_guide.html">HDFS Admin Guide: Quotas</a>
</div>
<div class="menuitem">
<a href="SLG_user_guide.html">HDFS Utilities</a>
</div>
<div class="menuitem">
<a href="hod_user_guide.html">HOD User Guide</a>
</div>
<div class="menuitem">
<a href="hod_admin_guide.html">HOD Admin Guide</a>
</div>
<div class="menuitem">
<a href="hod_config_guide.html">HOD Config Guide</a>
</div>
<div class="menuitem">
<a href="capacity_scheduler.html">Capacity Scheduler</a>
</div>
<div class="menuitem">
<a href="api/index.html">API Docs</a>
</div>
<div class="menuitem">
<a href="jdiff/changes.html">API Changes</a>
</div>
<div class="menuitem">
<a href="http://wiki.apache.org/hadoop/">Wiki</a>
</div>
<div class="menuitem">
<a href="http://wiki.apache.org/hadoop/FAQ">FAQ</a>
</div>
<div class="menuitem">
<a href="releasenotes.html">Release Notes</a>
</div>
<div class="menuitem">
<a href="changes.html">Change Log</a>
</div>
</div>
<div id="credit"></div>
<div id="roundbottom">
<img style="display: none" class="corner" height="15" width="15" alt="" src="skin/images/rc-b-l-15-1body-2menu-3menu.png"></div>
<!--+
  |alternative credits
  +-->
<div id="credit2"></div>
</div>
<!--+
    |end Menu
    +-->
<!--+
    |start content
    +-->
<div id="content">
<div title="Portable Document Format" class="pdflink">
<a class="dida" href="cluster_setup.pdf"><img alt="PDF -icon" src="skin/images/pdfdoc.gif" class="skin"><br>
        PDF</a>
</div>
<h1>Hadoop Cluster Setup</h1>
<div id="minitoc-area">
<ul class="minitoc">
<li>
<a href="#Purpose">Purpose</a>
</li>
<li>
<a href="#Pre-requisites">Pre-requisites</a>
</li>
<li>
<a href="#Installation">Installation</a>
</li>
<li>
<a href="#Configuration">Configuration</a>
<ul class="minitoc">
<li>
<a href="#Configuration+Files">Configuration Files</a>
</li>
<li>
<a href="#Site+Configuration">Site Configuration</a>
<ul class="minitoc">
<li>
<a href="#Configuring+the+Environment+of+the+Hadoop+Daemons">Configuring the Environment of the Hadoop Daemons</a>
</li>
<li>
<a href="#Configuring+the+Hadoop+Daemons">Configuring the Hadoop Daemons</a>
</li>
<li>
<a href="#Slaves">Slaves</a>
</li>
<li>
<a href="#Logging">Logging</a>
</li>
</ul>
</li>
</ul>
</li>
<li>
<a href="#Cluster+Restartability">Cluster Restartability</a>
<ul class="minitoc">
<li>
<a href="#Map%2FReduce">Map/Reduce</a>
</li>
</ul>
</li>
<li>
<a href="#Hadoop+Rack+Awareness">Hadoop Rack Awareness</a>
</li>
<li>
<a href="#Hadoop+Startup">Hadoop Startup</a>
</li>
<li>
<a href="#Hadoop+Shutdown">Hadoop Shutdown</a>
</li>
</ul>
</div>
  
    
<a name="N1000D"></a><a name="Purpose"></a>
<h2 class="h3">Purpose</h2>
<div class="section">
<p>This document describes how to install, configure and manage non-trivial
      Hadoop clusters ranging from a few nodes to extremely large clusters with 
      thousands of nodes.</p>
<p>
      To play with Hadoop, you may first want to install Hadoop on a single machine (see <a href="quickstart.html"> Hadoop Quick Start</a>).
      </p>
</div>
    
    
<a name="N1001E"></a><a name="Pre-requisites"></a>
<h2 class="h3">Pre-requisites</h2>
<div class="section">
<ol>
        
<li>
          Make sure all <a href="quickstart.html#PreReqs">requisite</a> software 
          is installed on all nodes in your cluster.
        </li>
        
<li>
          
<a href="quickstart.html#Download">Get</a> the Hadoop software.
        </li>
      
</ol>
</div>
    
    
<a name="N10036"></a><a name="Installation"></a>
<h2 class="h3">Installation</h2>
<div class="section">
<p>Installing a Hadoop cluster typically involves unpacking the software 
      on all the machines in the cluster.</p>
<p>Typically one machine in the cluster is designated as the 
      <span class="codefrag">NameNode</span> and another machine the as <span class="codefrag">JobTracker</span>,
      exclusively. These are the <em>masters</em>. The rest of the machines in 
      the cluster act as both <span class="codefrag">DataNode</span> <em>and</em> 
      <span class="codefrag">TaskTracker</span>. These are the <em>slaves</em>.</p>
<p>The root of the distribution is referred to as 
      <span class="codefrag">HADOOP_HOME</span>. All machines in the cluster usually have the same 
      <span class="codefrag">HADOOP_HOME</span> path.</p>
</div>
    
    
<a name="N10061"></a><a name="Configuration"></a>
<h2 class="h3">Configuration</h2>
<div class="section">
<p>The following sections describe how to configure a Hadoop cluster.</p>
<a name="N1006A"></a><a name="Configuration+Files"></a>
<h3 class="h4">Configuration Files</h3>
<p>Hadoop configuration is driven by two important configuration files
        found in the <span class="codefrag">conf/</span> directory of the distribution:</p>
<ol>
          
<li>
            
<a href="http://hadoop.apache.org/core/docs/current/hadoop-default.html">hadoop-default.xml</a> - Read-only 
            default configuration.
          </li>
          
<li>
            
<em>hadoop-site.xml</em> - Site-specific configuration.
          </li>
        
</ol>
<p>To learn more about how the Hadoop framework is controlled by these 
        configuration files, look 
        <a href="api/org/apache/hadoop/conf/Configuration.html">here</a>.</p>
<p>Additionally, you can control the Hadoop scripts found in the 
        <span class="codefrag">bin/</span> directory of the distribution, by setting site-specific 
        values via the <span class="codefrag">conf/hadoop-env.sh</span>.</p>
<a name="N10097"></a><a name="Site+Configuration"></a>
<h3 class="h4">Site Configuration</h3>
<p>To configure the the Hadoop cluster you will need to configure the
        <em>environment</em> in which the Hadoop daemons execute as well as
        the <em>configuration parameters</em> for the Hadoop daemons.</p>
<p>The Hadoop daemons are <span class="codefrag">NameNode</span>/<span class="codefrag">DataNode</span> 
        and <span class="codefrag">JobTracker</span>/<span class="codefrag">TaskTracker</span>.</p>
<a name="N100B5"></a><a name="Configuring+the+Environment+of+the+Hadoop+Daemons"></a>
<h4>Configuring the Environment of the Hadoop Daemons</h4>
<p>Administrators should use the <span class="codefrag">conf/hadoop-env.sh</span> script
          to do site-specific customization of the Hadoop daemons' process 
          environment.</p>
<p>At the very least you should specify the
          <span class="codefrag">JAVA_HOME</span> so that it is correctly defined on each
          remote node.</p>
<p>Administrators can configure individual daemons using the
          configuration options <span class="codefrag">HADOOP_*_OPTS</span>. Various options 
          available are shown below in the table. </p>
<table class="ForrestTable" cellspacing="1" cellpadding="4">
          
<tr>
<th colspan="1" rowspan="1">Daemon</th><th colspan="1" rowspan="1">Configure Options</th>
</tr>
          
<tr>
<td colspan="1" rowspan="1">NameNode</td><td colspan="1" rowspan="1">HADOOP_NAMENODE_OPTS</td>
</tr>
          
<tr>
<td colspan="1" rowspan="1">DataNode</td><td colspan="1" rowspan="1">HADOOP_DATANODE_OPTS</td>
</tr>
          
<tr>
<td colspan="1" rowspan="1">SecondaryNamenode</td>
              <td colspan="1" rowspan="1">HADOOP_SECONDARYNAMENODE_OPTS</td>
</tr>
          
<tr>
<td colspan="1" rowspan="1">JobTracker</td><td colspan="1" rowspan="1">HADOOP_JOBTRACKER_OPTS</td>
</tr>
          
<tr>
<td colspan="1" rowspan="1">TaskTracker</td><td colspan="1" rowspan="1">HADOOP_TASKTRACKER_OPTS</td>
</tr>
          
</table>
<p> For example, To configure Namenode to use parallelGC, the
          following statement should be added in <span class="codefrag">hadoop-env.sh</span> :
          <br>
<span class="codefrag">
          export HADOOP_NAMENODE_OPTS="-XX:+UseParallelGC ${HADOOP_NAMENODE_OPTS}"
          </span>
<br>
</p>
<p>Other useful configuration parameters that you can customize 
          include:</p>
<ul>
            
<li>
              
<span class="codefrag">HADOOP_LOG_DIR</span> - The directory where the daemons'
              log files are stored. They are automatically created if they don't
              exist.
            </li>
            
<li>
              
<span class="codefrag">HADOOP_HEAPSIZE</span> - The maximum amount of heapsize 
              to use, in MB e.g. <span class="codefrag">1000MB</span>. This is used to 
              configure the heap size for the hadoop daemon. By default,
              the value is <span class="codefrag">1000MB</span>.
            </li>
          
</ul>
<a name="N10130"></a><a name="Configuring+the+Hadoop+Daemons"></a>
<h4>Configuring the Hadoop Daemons</h4>
<p>This section deals with important parameters to be specified in the
          <span class="codefrag">conf/hadoop-site.xml</span> for the Hadoop cluster.</p>
<table class="ForrestTable" cellspacing="1" cellpadding="4">
  		    
<tr>
		      
<th colspan="1" rowspan="1">Parameter</th>
		      <th colspan="1" rowspan="1">Value</th> 
		      <th colspan="1" rowspan="1">Notes</th>
		    
</tr>
  		    
<tr>
		      
<td colspan="1" rowspan="1">fs.default.name</td>
  		      <td colspan="1" rowspan="1">URI of <span class="codefrag">NameNode</span>.</td>
		      <td colspan="1" rowspan="1"><em>hdfs://hostname/</em></td>
		    
</tr>
		    
<tr>
		      
<td colspan="1" rowspan="1">mapred.job.tracker</td>
		      <td colspan="1" rowspan="1">Host or IP and port of <span class="codefrag">JobTracker</span>.</td>
		      <td colspan="1" rowspan="1"><em>host:port</em> pair.</td>
		    
</tr>
		    
<tr>
		      
<td colspan="1" rowspan="1">dfs.name.dir</td>
		      <td colspan="1" rowspan="1">
		        Path on the local filesystem where the <span class="codefrag">NameNode</span> 
		        stores the namespace and transactions logs persistently.</td>
		      <td colspan="1" rowspan="1">
		        If this is a comma-delimited list of directories then the name 
		        table is replicated in all of the directories, for redundancy.
		      </td>
		    
</tr>
		    
<tr>
		      
<td colspan="1" rowspan="1">dfs.data.dir</td>
		      <td colspan="1" rowspan="1">
		        Comma separated list of paths on the local filesystem of a 
		        <span class="codefrag">DataNode</span> where it should store its blocks.
		      </td>
		      <td colspan="1" rowspan="1">
		        If this is a comma-delimited list of directories, then data will 
		        be stored in all named directories, typically on different 
		        devices.
		      </td>
		    
</tr>
		    
<tr>
		      
<td colspan="1" rowspan="1">mapred.system.dir</td>
		      <td colspan="1" rowspan="1">
		        Path on the HDFS where where the Map/Reduce framework stores 
		        system files e.g. <span class="codefrag">/hadoop/mapred/system/</span>.
		      </td>
		      <td colspan="1" rowspan="1">
		        This is in the default filesystem (HDFS) and must be accessible 
		        from both the server and client machines.
		      </td>
		    
</tr>
		    
<tr>
		      
<td colspan="1" rowspan="1">mapred.local.dir</td>
		      <td colspan="1" rowspan="1">
		        Comma-separated list of paths on the local filesystem where 
		        temporary Map/Reduce data is written.
		      </td>
		      <td colspan="1" rowspan="1">Multiple paths help spread disk i/o.</td>
		    
</tr>
		    
<tr>
		      
<td colspan="1" rowspan="1">mapred.tasktracker.{map|reduce}.tasks.maximum</td>
		      <td colspan="1" rowspan="1">
		        The maximum number of Map/Reduce tasks, which are run 
		        simultaneously on a given <span class="codefrag">TaskTracker</span>, individually.
		      </td>
		      <td colspan="1" rowspan="1">
		        Defaults to 2 (2 maps and 2 reduces), but vary it depending on 
		        your hardware.
		      </td>
		    
</tr>
		    
<tr>
		      
<td colspan="1" rowspan="1">dfs.hosts/dfs.hosts.exclude</td>
		      <td colspan="1" rowspan="1">List of permitted/excluded DataNodes.</td>
		      <td colspan="1" rowspan="1">
		        If necessary, use these files to control the list of allowable 
		        datanodes.
		      </td>
		    
</tr>
		    
<tr>
		      
<td colspan="1" rowspan="1">mapred.hosts/mapred.hosts.exclude</td>
		      <td colspan="1" rowspan="1">List of permitted/excluded TaskTrackers.</td>
		      <td colspan="1" rowspan="1">
		        If necessary, use these files to control the list of allowable 
		        TaskTrackers.
		      </td>
  		    
</tr>
        
<tr>
          
<td colspan="1" rowspan="1">mapred.queue.names</td>
          <td colspan="1" rowspan="1">Comma separated list of queues to which jobs can be submitted.</td>
          <td colspan="1" rowspan="1">
            The Map/Reduce system always supports atleast one queue
            with the name as <em>default</em>. Hence, this parameter's
            value should always contain the string <em>default</em>.
            Some job schedulers supported in Hadoop, like the 
            <a href="capacity_scheduler.html">Capacity 
            Scheduler</a>, support multiple queues. If such a scheduler is
            being used, the list of configured queue names must be
            specified here. Once queues are defined, users can submit
            jobs to a queue using the property name 
            <em>mapred.job.queue.name</em> in the job configuration.
            There could be a separate 
            configuration file for configuring properties of these 
            queues that is managed by the scheduler. 
            Refer to the documentation of the scheduler for information on 
            the same.
          </td>
        
</tr>
        
<tr>
          
<td colspan="1" rowspan="1">mapred.acls.enabled</td>
          <td colspan="1" rowspan="1">Specifies whether ACLs are supported for controlling job
              submission and administration</td>
          <td colspan="1" rowspan="1">
            If <em>true</em>, ACLs would be checked while submitting
            and administering jobs. ACLs can be specified using the
            configuration parameters of the form
            <em>mapred.queue.queue-name.acl-name</em>, defined below.
          </td>
        
</tr>
        
<tr>
          
<td colspan="1" rowspan="1">mapred.queue.<em>queue-name</em>.acl-submit-job</td>
          <td colspan="1" rowspan="1">List of users and groups that can submit jobs to the
              specified <em>queue-name</em>.</td>
          <td colspan="1" rowspan="1">
            The list of users and groups are both comma separated
            list of names. The two lists are separated by a blank.
            Example: <em>user1,user2 group1,group2</em>.
            If you wish to define only a list of groups, provide
            a blank at the beginning of the value.
          </td>
        
</tr>
        
<tr>
          
<td colspan="1" rowspan="1">mapred.queue.<em>queue-name</em>.acl-administer-job</td>
          <td colspan="1" rowspan="1">List of users and groups that can change the priority
              or kill jobs that have been submitted to the
              specified <em>queue-name</em>.</td>
          <td colspan="1" rowspan="1">
            The list of users and groups are both comma separated
            list of names. The two lists are separated by a blank.
            Example: <em>user1,user2 group1,group2</em>.
            If you wish to define only a list of groups, provide
            a blank at the beginning of the value. Note that an
            owner of a job can always change the priority or kill
            his/her own job, irrespective of the ACLs.
          </td>
        
</tr>
		  
</table>
<p>Typically all the above parameters are marked as 
          <a href="api/org/apache/hadoop/conf/Configuration.html#FinalParams">
          final</a> to ensure that they cannot be overriden by user-applications.
          </p>
<a name="N1027C"></a><a name="Real-World+Cluster+Configurations"></a>
<h5>Real-World Cluster Configurations</h5>
<p>This section lists some non-default configuration parameters which 
            have been used to run the <em>sort</em> benchmark on very large 
            clusters.</p>
<ul>
              
<li>
                
<p>Some non-default configuration values used to run sort900,
                that is 9TB of data sorted on a cluster with 900 nodes:</p>
                
<table class="ForrestTable" cellspacing="1" cellpadding="4">
  		          
<tr>
		            
<th colspan="1" rowspan="1">Parameter</th>
		            <th colspan="1" rowspan="1">Value</th> 
		            <th colspan="1" rowspan="1">Notes</th>
		          
</tr>
                  
<tr>
                    
<td colspan="1" rowspan="1">dfs.block.size</td>
                    <td colspan="1" rowspan="1">134217728</td>
                    <td colspan="1" rowspan="1">HDFS blocksize of 128MB for large file-systems.</td>
                  
</tr>
                  
<tr>
                    
<td colspan="1" rowspan="1">dfs.namenode.handler.count</td>
                    <td colspan="1" rowspan="1">40</td>
                    <td colspan="1" rowspan="1">
                      More NameNode server threads to handle RPCs from large 
                      number of DataNodes.
                    </td>
                  
</tr>
                  
<tr>
                    
<td colspan="1" rowspan="1">mapred.reduce.parallel.copies</td>
                    <td colspan="1" rowspan="1">20</td>
                    <td colspan="1" rowspan="1">
                      Higher number of parallel copies run by reduces to fetch
                      outputs from very large number of maps.
                    </td>
                  
</tr>
                  
<tr>
                    
<td colspan="1" rowspan="1">mapred.child.java.opts</td>
                    <td colspan="1" rowspan="1">-Xmx512M</td>
                    <td colspan="1" rowspan="1">
                      Larger heap-size for child jvms of maps/reduces. 
                    </td>
                  
</tr>
                  
<tr>
                    
<td colspan="1" rowspan="1">fs.inmemory.size.mb</td>
                    <td colspan="1" rowspan="1">200</td>
                    <td colspan="1" rowspan="1">
                      Larger amount of memory allocated for the in-memory 
                      file-system used to merge map-outputs at the reduces.
                    </td>
                  
</tr>
                  
<tr>
                    
<td colspan="1" rowspan="1">io.sort.factor</td>
                    <td colspan="1" rowspan="1">100</td>
                    <td colspan="1" rowspan="1">More streams merged at once while sorting files.</td>
                  
</tr>
                  
<tr>
                    
<td colspan="1" rowspan="1">io.sort.mb</td>
                    <td colspan="1" rowspan="1">200</td>
                    <td colspan="1" rowspan="1">Higher memory-limit while sorting data.</td>
                  
</tr>
                  
<tr>
                    
<td colspan="1" rowspan="1">io.file.buffer.size</td>
                    <td colspan="1" rowspan="1">131072</td>
                    <td colspan="1" rowspan="1">Size of read/write buffer used in SequenceFiles.</td>
                  
</tr>
                
</table>
              
</li>
              
<li>
                
<p>Updates to some configuration values to run sort1400 and 
                sort2000, that is 14TB of data sorted on 1400 nodes and 20TB of
                data sorted on 2000 nodes:</p>
                
<table class="ForrestTable" cellspacing="1" cellpadding="4">
  		          
<tr>
		            
<th colspan="1" rowspan="1">Parameter</th>
		            <th colspan="1" rowspan="1">Value</th> 
		            <th colspan="1" rowspan="1">Notes</th>
		          
</tr>
                  
<tr>
                    
<td colspan="1" rowspan="1">mapred.job.tracker.handler.count</td>
                    <td colspan="1" rowspan="1">60</td>
                    <td colspan="1" rowspan="1">
                      More JobTracker server threads to handle RPCs from large 
                      number of TaskTrackers.
                    </td>
                  
</tr>
                  
<tr>
                    
<td colspan="1" rowspan="1">mapred.reduce.parallel.copies</td>
                    <td colspan="1" rowspan="1">50</td>
                    <td colspan="1" rowspan="1"></td>
                  
</tr>
                  
<tr>
                    
<td colspan="1" rowspan="1">tasktracker.http.threads</td>
                    <td colspan="1" rowspan="1">50</td>
                    <td colspan="1" rowspan="1">
                      More worker threads for the TaskTracker's http server. The
                      http server is used by reduces to fetch intermediate 
                      map-outputs.
                    </td>
                  
</tr>
                  
<tr>
                    
<td colspan="1" rowspan="1">mapred.child.java.opts</td>
                    <td colspan="1" rowspan="1">-Xmx1024M</td>
                    <td colspan="1" rowspan="1">Larger heap-size for child jvms of maps/reduces.</td>
                  
</tr>
                
</table>
              
</li>
            
</ul>
<a name="N1039A"></a><a name="Slaves"></a>
<h4>Slaves</h4>
<p>Typically you choose one machine in the cluster to act as the 
          <span class="codefrag">NameNode</span> and one machine as to act as the 
          <span class="codefrag">JobTracker</span>, exclusively. The rest of the machines act as 
          both a <span class="codefrag">DataNode</span> and <span class="codefrag">TaskTracker</span> and are 
          referred to as <em>slaves</em>.</p>
<p>List all slave hostnames or IP addresses in your 
          <span class="codefrag">conf/slaves</span> file, one per line.</p>
<a name="N103B9"></a><a name="Logging"></a>
<h4>Logging</h4>
<p>Hadoop uses the <a href="http://logging.apache.org/log4j/">Apache 
          log4j</a> via the <a href="http://commons.apache.org/logging/">Apache 
          Commons Logging</a> framework for logging. Edit the 
          <span class="codefrag">conf/log4j.properties</span> file to customize the Hadoop 
          daemons' logging configuration (log-formats and so on).</p>
<a name="N103CD"></a><a name="History+Logging"></a>
<h5>History Logging</h5>
<p> The job history files are stored in central location 
            <span class="codefrag"> hadoop.job.history.location </span> which can be on DFS also,
            whose default value is <span class="codefrag">${HADOOP_LOG_DIR}/history</span>. 
            The history web UI is accessible from job tracker web UI.</p>
<p> The history files are also logged to user specified directory
            <span class="codefrag">hadoop.job.history.user.location</span> 
            which defaults to job output directory. The files are stored in
            "_logs/history/" in the specified directory. Hence, by default 
            they will be in "mapred.output.dir/_logs/history/". User can stop
            logging by giving the value <span class="codefrag">none</span> for 
            <span class="codefrag">hadoop.job.history.user.location</span> 
</p>
<p> User can view the history logs summary in specified directory 
            using the following command <br>
            
<span class="codefrag">$ bin/hadoop job -history output-dir</span>
<br> 
            This command will print job details, failed and killed tip
            details. <br>
            More details about the job such as successful tasks and 
            task attempts made for each task can be viewed using the  
            following command <br>
            
<span class="codefrag">$ bin/hadoop job -history all output-dir</span>
<br>
</p>
<p>Once all the necessary configuration is complete, distribute the files
      to the <span class="codefrag">HADOOP_CONF_DIR</span> directory on all the machines, 
      typically <span class="codefrag">${HADOOP_HOME}/conf</span>.</p>
</div>
    
<a name="N10405"></a><a name="Cluster+Restartability"></a>
<h2 class="h3">Cluster Restartability</h2>
<div class="section">
<a name="N1040B"></a><a name="Map%2FReduce"></a>
<h3 class="h4">Map/Reduce</h3>
<p>The job tracker restart can recover running jobs if 
        <span class="codefrag">mapred.jobtracker.restart.recover</span> is set true and 
        <a href="#Logging">JobHistory logging</a> is enabled. Also 
        <span class="codefrag">mapred.jobtracker.job.history.block.size</span> value should be 
        set to an optimal value to dump job history to disk as soon as 
        possible, the typical value is 3145728(3MB).</p>
</div>
    
    
<a name="N10420"></a><a name="Hadoop+Rack+Awareness"></a>
<h2 class="h3">Hadoop Rack Awareness</h2>
<div class="section">
<p>The HDFS and the Map/Reduce components are rack-aware.</p>
<p>The <span class="codefrag">NameNode</span> and the <span class="codefrag">JobTracker</span> obtains the
      <span class="codefrag">rack id</span> of the slaves in the cluster by invoking an API 
      <a href="api/org/apache/hadoop/net/DNSToSwitchMapping.html#resolve(java.util.List)">resolve</a> in an administrator configured
      module. The API resolves the slave's DNS name (also IP address) to a 
      rack id. What module to use can be configured using the configuration
      item <span class="codefrag">topology.node.switch.mapping.impl</span>. The default 
      implementation of the same runs a script/command configured using 
      <span class="codefrag">topology.script.file.name</span>. If topology.script.file.name is
      not set, the rack id <span class="codefrag">/default-rack</span> is returned for any 
      passed IP address. The additional configuration in the Map/Reduce
      part is <span class="codefrag">mapred.cache.task.levels</span> which determines the number
      of levels (in the network topology) of caches. So, for example, if it is
      the default value of 2, two levels of caches will be constructed - 
      one for hosts (host -&gt; task mapping) and another for racks 
      (rack -&gt; task mapping).
      </p>
</div>
    
    
<a name="N10446"></a><a name="Hadoop+Startup"></a>
<h2 class="h3">Hadoop Startup</h2>
<div class="section">
<p>To start a Hadoop cluster you will need to start both the HDFS and 
      Map/Reduce cluster.</p>
<p>
        Format a new distributed filesystem:<br>
        
<span class="codefrag">$ bin/hadoop namenode -format</span>
      
</p>
<p>
        Start the HDFS with the following command, run on the designated
        <span class="codefrag">NameNode</span>:<br>
        
<span class="codefrag">$ bin/start-dfs.sh</span>
      
</p>
<p>The <span class="codefrag">bin/start-dfs.sh</span> script also consults the 
      <span class="codefrag">${HADOOP_CONF_DIR}/slaves</span> file on the <span class="codefrag">NameNode</span> 
      and starts the <span class="codefrag">DataNode</span> daemon on all the listed slaves.</p>
<p>
        Start Map-Reduce with the following command, run on the designated
        <span class="codefrag">JobTracker</span>:<br>
        
<span class="codefrag">$ bin/start-mapred.sh</span>
      
</p>
<p>The <span class="codefrag">bin/start-mapred.sh</span> script also consults the 
      <span class="codefrag">${HADOOP_CONF_DIR}/slaves</span> file on the <span class="codefrag">JobTracker</span> 
      and starts the <span class="codefrag">TaskTracker</span> daemon on all the listed slaves.
      </p>
</div>
    
    
<a name="N1048C"></a><a name="Hadoop+Shutdown"></a>
<h2 class="h3">Hadoop Shutdown</h2>
<div class="section">
<p>
        Stop HDFS with the following command, run on the designated 
        <span class="codefrag">NameNode</span>:<br>
        
<span class="codefrag">$ bin/stop-dfs.sh</span>
      
</p>
<p>The <span class="codefrag">bin/stop-dfs.sh</span> script also consults the 
      <span class="codefrag">${HADOOP_CONF_DIR}/slaves</span> file on the <span class="codefrag">NameNode</span> 
      and stops the <span class="codefrag">DataNode</span> daemon on all the listed slaves.</p>
<p>
        Stop Map/Reduce with the following command, run on the designated
        the designated <span class="codefrag">JobTracker</span>:<br>
        
<span class="codefrag">$ bin/stop-mapred.sh</span>
<br>
      
</p>
<p>The <span class="codefrag">bin/stop-mapred.sh</span> script also consults the 
      <span class="codefrag">${HADOOP_CONF_DIR}/slaves</span> file on the <span class="codefrag">JobTracker</span> 
      and stops the <span class="codefrag">TaskTracker</span> daemon on all the listed slaves.</p>
</div>
  
</div>
<!--+
    |end content
    +-->
<div class="clearboth">&nbsp;</div>
</div>
<div id="footer">
<!--+
    |start bottomstrip
    +-->
<div class="lastmodified">
<script type="text/javascript"><!--
document.write("Last Published: " + document.lastModified);
//  --></script>
</div>
<div class="copyright">
        Copyright &copy;
         2008 <a href="http://www.apache.org/licenses/">The Apache Software Foundation.</a>
</div>
<!--+
    |end bottomstrip
    +-->
</div>
</body>
</html>