You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@hbase.apache.org by mi...@apache.org on 2015/12/10 18:13:03 UTC
[41/51] [partial] hbase-site git commit: Published site at
26ac60b03f80c9215103a02db783341e67037753.
http://git-wip-us.apache.org/repos/asf/hbase-site/blob/f68717a0/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.html
----------------------------------------------------------------------
diff --git a/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.html b/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.html
index 88589f1..d6cf161 100644
--- a/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.html
+++ b/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.html
@@ -51,943 +51,953 @@
<span class="sourceLineNo">043</span>import org.apache.hadoop.hbase.security.token.TokenUtil;<a name="line.43"></a>
<span class="sourceLineNo">044</span>import org.apache.hadoop.hbase.util.Base64;<a name="line.44"></a>
<span class="sourceLineNo">045</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.45"></a>
-<span class="sourceLineNo">046</span>import org.apache.hadoop.hbase.zookeeper.ZKUtil;<a name="line.46"></a>
+<span class="sourceLineNo">046</span>import org.apache.hadoop.hbase.zookeeper.ZKConfig;<a name="line.46"></a>
<span class="sourceLineNo">047</span>import org.apache.hadoop.io.Writable;<a name="line.47"></a>
<span class="sourceLineNo">048</span>import org.apache.hadoop.mapreduce.InputFormat;<a name="line.48"></a>
<span class="sourceLineNo">049</span>import org.apache.hadoop.mapreduce.Job;<a name="line.49"></a>
<span class="sourceLineNo">050</span>import org.apache.hadoop.util.StringUtils;<a name="line.50"></a>
-<span class="sourceLineNo">051</span>import com.google.protobuf.InvalidProtocolBufferException;<a name="line.51"></a>
-<span class="sourceLineNo">052</span><a name="line.52"></a>
-<span class="sourceLineNo">053</span>import java.io.File;<a name="line.53"></a>
-<span class="sourceLineNo">054</span>import java.io.IOException;<a name="line.54"></a>
-<span class="sourceLineNo">055</span>import java.net.URL;<a name="line.55"></a>
-<span class="sourceLineNo">056</span>import java.net.URLDecoder;<a name="line.56"></a>
-<span class="sourceLineNo">057</span>import java.util.ArrayList;<a name="line.57"></a>
-<span class="sourceLineNo">058</span>import java.util.Collection;<a name="line.58"></a>
-<span class="sourceLineNo">059</span>import java.util.Enumeration;<a name="line.59"></a>
-<span class="sourceLineNo">060</span>import java.util.HashMap;<a name="line.60"></a>
-<span class="sourceLineNo">061</span>import java.util.HashSet;<a name="line.61"></a>
-<span class="sourceLineNo">062</span>import java.util.List;<a name="line.62"></a>
-<span class="sourceLineNo">063</span>import java.util.Map;<a name="line.63"></a>
-<span class="sourceLineNo">064</span>import java.util.Set;<a name="line.64"></a>
-<span class="sourceLineNo">065</span>import java.util.zip.ZipEntry;<a name="line.65"></a>
-<span class="sourceLineNo">066</span>import java.util.zip.ZipFile;<a name="line.66"></a>
-<span class="sourceLineNo">067</span><a name="line.67"></a>
-<span class="sourceLineNo">068</span>/**<a name="line.68"></a>
-<span class="sourceLineNo">069</span> * Utility for {@link TableMapper} and {@link TableReducer}<a name="line.69"></a>
-<span class="sourceLineNo">070</span> */<a name="line.70"></a>
-<span class="sourceLineNo">071</span>@SuppressWarnings({ "rawtypes", "unchecked" })<a name="line.71"></a>
-<span class="sourceLineNo">072</span>@InterfaceAudience.Public<a name="line.72"></a>
-<span class="sourceLineNo">073</span>@InterfaceStability.Stable<a name="line.73"></a>
-<span class="sourceLineNo">074</span>public class TableMapReduceUtil {<a name="line.74"></a>
-<span class="sourceLineNo">075</span> private static final Log LOG = LogFactory.getLog(TableMapReduceUtil.class);<a name="line.75"></a>
-<span class="sourceLineNo">076</span><a name="line.76"></a>
-<span class="sourceLineNo">077</span> /**<a name="line.77"></a>
-<span class="sourceLineNo">078</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.78"></a>
-<span class="sourceLineNo">079</span> * the job.<a name="line.79"></a>
-<span class="sourceLineNo">080</span> *<a name="line.80"></a>
-<span class="sourceLineNo">081</span> * @param table The table name to read from.<a name="line.81"></a>
-<span class="sourceLineNo">082</span> * @param scan The scan instance with the columns, time range etc.<a name="line.82"></a>
-<span class="sourceLineNo">083</span> * @param mapper The mapper class to use.<a name="line.83"></a>
-<span class="sourceLineNo">084</span> * @param outputKeyClass The class of the output key.<a name="line.84"></a>
-<span class="sourceLineNo">085</span> * @param outputValueClass The class of the output value.<a name="line.85"></a>
-<span class="sourceLineNo">086</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.86"></a>
-<span class="sourceLineNo">087</span> * carrying all necessary HBase configuration.<a name="line.87"></a>
-<span class="sourceLineNo">088</span> * @throws IOException When setting up the details fails.<a name="line.88"></a>
-<span class="sourceLineNo">089</span> */<a name="line.89"></a>
-<span class="sourceLineNo">090</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.90"></a>
-<span class="sourceLineNo">091</span> Class<? extends TableMapper> mapper,<a name="line.91"></a>
-<span class="sourceLineNo">092</span> Class<?> outputKeyClass,<a name="line.92"></a>
-<span class="sourceLineNo">093</span> Class<?> outputValueClass, Job job)<a name="line.93"></a>
-<span class="sourceLineNo">094</span> throws IOException {<a name="line.94"></a>
-<span class="sourceLineNo">095</span> initTableMapperJob(table, scan, mapper, outputKeyClass, outputValueClass,<a name="line.95"></a>
-<span class="sourceLineNo">096</span> job, true);<a name="line.96"></a>
-<span class="sourceLineNo">097</span> }<a name="line.97"></a>
+<span class="sourceLineNo">051</span><a name="line.51"></a>
+<span class="sourceLineNo">052</span>import java.io.File;<a name="line.52"></a>
+<span class="sourceLineNo">053</span>import java.io.IOException;<a name="line.53"></a>
+<span class="sourceLineNo">054</span>import java.net.URL;<a name="line.54"></a>
+<span class="sourceLineNo">055</span>import java.net.URLDecoder;<a name="line.55"></a>
+<span class="sourceLineNo">056</span>import java.util.ArrayList;<a name="line.56"></a>
+<span class="sourceLineNo">057</span>import java.util.Collection;<a name="line.57"></a>
+<span class="sourceLineNo">058</span>import java.util.Enumeration;<a name="line.58"></a>
+<span class="sourceLineNo">059</span>import java.util.HashMap;<a name="line.59"></a>
+<span class="sourceLineNo">060</span>import java.util.HashSet;<a name="line.60"></a>
+<span class="sourceLineNo">061</span>import java.util.List;<a name="line.61"></a>
+<span class="sourceLineNo">062</span>import java.util.Map;<a name="line.62"></a>
+<span class="sourceLineNo">063</span>import java.util.Set;<a name="line.63"></a>
+<span class="sourceLineNo">064</span>import java.util.zip.ZipEntry;<a name="line.64"></a>
+<span class="sourceLineNo">065</span>import java.util.zip.ZipFile;<a name="line.65"></a>
+<span class="sourceLineNo">066</span><a name="line.66"></a>
+<span class="sourceLineNo">067</span>/**<a name="line.67"></a>
+<span class="sourceLineNo">068</span> * Utility for {@link TableMapper} and {@link TableReducer}<a name="line.68"></a>
+<span class="sourceLineNo">069</span> */<a name="line.69"></a>
+<span class="sourceLineNo">070</span>@SuppressWarnings({ "rawtypes", "unchecked" })<a name="line.70"></a>
+<span class="sourceLineNo">071</span>@InterfaceAudience.Public<a name="line.71"></a>
+<span class="sourceLineNo">072</span>@InterfaceStability.Stable<a name="line.72"></a>
+<span class="sourceLineNo">073</span>public class TableMapReduceUtil {<a name="line.73"></a>
+<span class="sourceLineNo">074</span> private static final Log LOG = LogFactory.getLog(TableMapReduceUtil.class);<a name="line.74"></a>
+<span class="sourceLineNo">075</span><a name="line.75"></a>
+<span class="sourceLineNo">076</span> /**<a name="line.76"></a>
+<span class="sourceLineNo">077</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.77"></a>
+<span class="sourceLineNo">078</span> * the job.<a name="line.78"></a>
+<span class="sourceLineNo">079</span> *<a name="line.79"></a>
+<span class="sourceLineNo">080</span> * @param table The table name to read from.<a name="line.80"></a>
+<span class="sourceLineNo">081</span> * @param scan The scan instance with the columns, time range etc.<a name="line.81"></a>
+<span class="sourceLineNo">082</span> * @param mapper The mapper class to use.<a name="line.82"></a>
+<span class="sourceLineNo">083</span> * @param outputKeyClass The class of the output key.<a name="line.83"></a>
+<span class="sourceLineNo">084</span> * @param outputValueClass The class of the output value.<a name="line.84"></a>
+<span class="sourceLineNo">085</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.85"></a>
+<span class="sourceLineNo">086</span> * carrying all necessary HBase configuration.<a name="line.86"></a>
+<span class="sourceLineNo">087</span> * @throws IOException When setting up the details fails.<a name="line.87"></a>
+<span class="sourceLineNo">088</span> */<a name="line.88"></a>
+<span class="sourceLineNo">089</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.89"></a>
+<span class="sourceLineNo">090</span> Class<? extends TableMapper> mapper,<a name="line.90"></a>
+<span class="sourceLineNo">091</span> Class<?> outputKeyClass,<a name="line.91"></a>
+<span class="sourceLineNo">092</span> Class<?> outputValueClass, Job job)<a name="line.92"></a>
+<span class="sourceLineNo">093</span> throws IOException {<a name="line.93"></a>
+<span class="sourceLineNo">094</span> initTableMapperJob(table, scan, mapper, outputKeyClass, outputValueClass,<a name="line.94"></a>
+<span class="sourceLineNo">095</span> job, true);<a name="line.95"></a>
+<span class="sourceLineNo">096</span> }<a name="line.96"></a>
+<span class="sourceLineNo">097</span><a name="line.97"></a>
<span class="sourceLineNo">098</span><a name="line.98"></a>
-<span class="sourceLineNo">099</span><a name="line.99"></a>
-<span class="sourceLineNo">100</span> /**<a name="line.100"></a>
-<span class="sourceLineNo">101</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.101"></a>
-<span class="sourceLineNo">102</span> * the job.<a name="line.102"></a>
-<span class="sourceLineNo">103</span> *<a name="line.103"></a>
-<span class="sourceLineNo">104</span> * @param table The table name to read from.<a name="line.104"></a>
-<span class="sourceLineNo">105</span> * @param scan The scan instance with the columns, time range etc.<a name="line.105"></a>
-<span class="sourceLineNo">106</span> * @param mapper The mapper class to use.<a name="line.106"></a>
-<span class="sourceLineNo">107</span> * @param outputKeyClass The class of the output key.<a name="line.107"></a>
-<span class="sourceLineNo">108</span> * @param outputValueClass The class of the output value.<a name="line.108"></a>
-<span class="sourceLineNo">109</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.109"></a>
-<span class="sourceLineNo">110</span> * carrying all necessary HBase configuration.<a name="line.110"></a>
-<span class="sourceLineNo">111</span> * @throws IOException When setting up the details fails.<a name="line.111"></a>
-<span class="sourceLineNo">112</span> */<a name="line.112"></a>
-<span class="sourceLineNo">113</span> public static void initTableMapperJob(TableName table,<a name="line.113"></a>
-<span class="sourceLineNo">114</span> Scan scan,<a name="line.114"></a>
-<span class="sourceLineNo">115</span> Class<? extends TableMapper> mapper,<a name="line.115"></a>
-<span class="sourceLineNo">116</span> Class<?> outputKeyClass,<a name="line.116"></a>
-<span class="sourceLineNo">117</span> Class<?> outputValueClass,<a name="line.117"></a>
-<span class="sourceLineNo">118</span> Job job) throws IOException {<a name="line.118"></a>
-<span class="sourceLineNo">119</span> initTableMapperJob(table.getNameAsString(),<a name="line.119"></a>
-<span class="sourceLineNo">120</span> scan,<a name="line.120"></a>
-<span class="sourceLineNo">121</span> mapper,<a name="line.121"></a>
-<span class="sourceLineNo">122</span> outputKeyClass,<a name="line.122"></a>
-<span class="sourceLineNo">123</span> outputValueClass,<a name="line.123"></a>
-<span class="sourceLineNo">124</span> job,<a name="line.124"></a>
-<span class="sourceLineNo">125</span> true);<a name="line.125"></a>
-<span class="sourceLineNo">126</span> }<a name="line.126"></a>
-<span class="sourceLineNo">127</span><a name="line.127"></a>
-<span class="sourceLineNo">128</span> /**<a name="line.128"></a>
-<span class="sourceLineNo">129</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.129"></a>
-<span class="sourceLineNo">130</span> * the job.<a name="line.130"></a>
-<span class="sourceLineNo">131</span> *<a name="line.131"></a>
-<span class="sourceLineNo">132</span> * @param table Binary representation of the table name to read from.<a name="line.132"></a>
-<span class="sourceLineNo">133</span> * @param scan The scan instance with the columns, time range etc.<a name="line.133"></a>
-<span class="sourceLineNo">134</span> * @param mapper The mapper class to use.<a name="line.134"></a>
-<span class="sourceLineNo">135</span> * @param outputKeyClass The class of the output key.<a name="line.135"></a>
-<span class="sourceLineNo">136</span> * @param outputValueClass The class of the output value.<a name="line.136"></a>
-<span class="sourceLineNo">137</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.137"></a>
-<span class="sourceLineNo">138</span> * carrying all necessary HBase configuration.<a name="line.138"></a>
-<span class="sourceLineNo">139</span> * @throws IOException When setting up the details fails.<a name="line.139"></a>
-<span class="sourceLineNo">140</span> */<a name="line.140"></a>
-<span class="sourceLineNo">141</span> public static void initTableMapperJob(byte[] table, Scan scan,<a name="line.141"></a>
-<span class="sourceLineNo">142</span> Class<? extends TableMapper> mapper,<a name="line.142"></a>
-<span class="sourceLineNo">143</span> Class<?> outputKeyClass,<a name="line.143"></a>
-<span class="sourceLineNo">144</span> Class<?> outputValueClass, Job job)<a name="line.144"></a>
-<span class="sourceLineNo">145</span> throws IOException {<a name="line.145"></a>
-<span class="sourceLineNo">146</span> initTableMapperJob(Bytes.toString(table), scan, mapper, outputKeyClass, outputValueClass,<a name="line.146"></a>
-<span class="sourceLineNo">147</span> job, true);<a name="line.147"></a>
-<span class="sourceLineNo">148</span> }<a name="line.148"></a>
-<span class="sourceLineNo">149</span><a name="line.149"></a>
-<span class="sourceLineNo">150</span> /**<a name="line.150"></a>
-<span class="sourceLineNo">151</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.151"></a>
-<span class="sourceLineNo">152</span> * the job.<a name="line.152"></a>
-<span class="sourceLineNo">153</span> *<a name="line.153"></a>
-<span class="sourceLineNo">154</span> * @param table The table name to read from.<a name="line.154"></a>
-<span class="sourceLineNo">155</span> * @param scan The scan instance with the columns, time range etc.<a name="line.155"></a>
-<span class="sourceLineNo">156</span> * @param mapper The mapper class to use.<a name="line.156"></a>
-<span class="sourceLineNo">157</span> * @param outputKeyClass The class of the output key.<a name="line.157"></a>
-<span class="sourceLineNo">158</span> * @param outputValueClass The class of the output value.<a name="line.158"></a>
-<span class="sourceLineNo">159</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.159"></a>
-<span class="sourceLineNo">160</span> * carrying all necessary HBase configuration.<a name="line.160"></a>
-<span class="sourceLineNo">161</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.161"></a>
-<span class="sourceLineNo">162</span> * job classes via the distributed cache (tmpjars).<a name="line.162"></a>
-<span class="sourceLineNo">163</span> * @throws IOException When setting up the details fails.<a name="line.163"></a>
-<span class="sourceLineNo">164</span> */<a name="line.164"></a>
-<span class="sourceLineNo">165</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.165"></a>
-<span class="sourceLineNo">166</span> Class<? extends TableMapper> mapper,<a name="line.166"></a>
-<span class="sourceLineNo">167</span> Class<?> outputKeyClass,<a name="line.167"></a>
-<span class="sourceLineNo">168</span> Class<?> outputValueClass, Job job,<a name="line.168"></a>
-<span class="sourceLineNo">169</span> boolean addDependencyJars, Class<? extends InputFormat> inputFormatClass)<a name="line.169"></a>
-<span class="sourceLineNo">170</span> throws IOException {<a name="line.170"></a>
-<span class="sourceLineNo">171</span> initTableMapperJob(table, scan, mapper, outputKeyClass, outputValueClass, job,<a name="line.171"></a>
-<span class="sourceLineNo">172</span> addDependencyJars, true, inputFormatClass);<a name="line.172"></a>
-<span class="sourceLineNo">173</span> }<a name="line.173"></a>
+<span class="sourceLineNo">099</span> /**<a name="line.99"></a>
+<span class="sourceLineNo">100</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.100"></a>
+<span class="sourceLineNo">101</span> * the job.<a name="line.101"></a>
+<span class="sourceLineNo">102</span> *<a name="line.102"></a>
+<span class="sourceLineNo">103</span> * @param table The table name to read from.<a name="line.103"></a>
+<span class="sourceLineNo">104</span> * @param scan The scan instance with the columns, time range etc.<a name="line.104"></a>
+<span class="sourceLineNo">105</span> * @param mapper The mapper class to use.<a name="line.105"></a>
+<span class="sourceLineNo">106</span> * @param outputKeyClass The class of the output key.<a name="line.106"></a>
+<span class="sourceLineNo">107</span> * @param outputValueClass The class of the output value.<a name="line.107"></a>
+<span class="sourceLineNo">108</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.108"></a>
+<span class="sourceLineNo">109</span> * carrying all necessary HBase configuration.<a name="line.109"></a>
+<span class="sourceLineNo">110</span> * @throws IOException When setting up the details fails.<a name="line.110"></a>
+<span class="sourceLineNo">111</span> */<a name="line.111"></a>
+<span class="sourceLineNo">112</span> public static void initTableMapperJob(TableName table,<a name="line.112"></a>
+<span class="sourceLineNo">113</span> Scan scan,<a name="line.113"></a>
+<span class="sourceLineNo">114</span> Class<? extends TableMapper> mapper,<a name="line.114"></a>
+<span class="sourceLineNo">115</span> Class<?> outputKeyClass,<a name="line.115"></a>
+<span class="sourceLineNo">116</span> Class<?> outputValueClass,<a name="line.116"></a>
+<span class="sourceLineNo">117</span> Job job) throws IOException {<a name="line.117"></a>
+<span class="sourceLineNo">118</span> initTableMapperJob(table.getNameAsString(),<a name="line.118"></a>
+<span class="sourceLineNo">119</span> scan,<a name="line.119"></a>
+<span class="sourceLineNo">120</span> mapper,<a name="line.120"></a>
+<span class="sourceLineNo">121</span> outputKeyClass,<a name="line.121"></a>
+<span class="sourceLineNo">122</span> outputValueClass,<a name="line.122"></a>
+<span class="sourceLineNo">123</span> job,<a name="line.123"></a>
+<span class="sourceLineNo">124</span> true);<a name="line.124"></a>
+<span class="sourceLineNo">125</span> }<a name="line.125"></a>
+<span class="sourceLineNo">126</span><a name="line.126"></a>
+<span class="sourceLineNo">127</span> /**<a name="line.127"></a>
+<span class="sourceLineNo">128</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.128"></a>
+<span class="sourceLineNo">129</span> * the job.<a name="line.129"></a>
+<span class="sourceLineNo">130</span> *<a name="line.130"></a>
+<span class="sourceLineNo">131</span> * @param table Binary representation of the table name to read from.<a name="line.131"></a>
+<span class="sourceLineNo">132</span> * @param scan The scan instance with the columns, time range etc.<a name="line.132"></a>
+<span class="sourceLineNo">133</span> * @param mapper The mapper class to use.<a name="line.133"></a>
+<span class="sourceLineNo">134</span> * @param outputKeyClass The class of the output key.<a name="line.134"></a>
+<span class="sourceLineNo">135</span> * @param outputValueClass The class of the output value.<a name="line.135"></a>
+<span class="sourceLineNo">136</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.136"></a>
+<span class="sourceLineNo">137</span> * carrying all necessary HBase configuration.<a name="line.137"></a>
+<span class="sourceLineNo">138</span> * @throws IOException When setting up the details fails.<a name="line.138"></a>
+<span class="sourceLineNo">139</span> */<a name="line.139"></a>
+<span class="sourceLineNo">140</span> public static void initTableMapperJob(byte[] table, Scan scan,<a name="line.140"></a>
+<span class="sourceLineNo">141</span> Class<? extends TableMapper> mapper,<a name="line.141"></a>
+<span class="sourceLineNo">142</span> Class<?> outputKeyClass,<a name="line.142"></a>
+<span class="sourceLineNo">143</span> Class<?> outputValueClass, Job job)<a name="line.143"></a>
+<span class="sourceLineNo">144</span> throws IOException {<a name="line.144"></a>
+<span class="sourceLineNo">145</span> initTableMapperJob(Bytes.toString(table), scan, mapper, outputKeyClass, outputValueClass,<a name="line.145"></a>
+<span class="sourceLineNo">146</span> job, true);<a name="line.146"></a>
+<span class="sourceLineNo">147</span> }<a name="line.147"></a>
+<span class="sourceLineNo">148</span><a name="line.148"></a>
+<span class="sourceLineNo">149</span> /**<a name="line.149"></a>
+<span class="sourceLineNo">150</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.150"></a>
+<span class="sourceLineNo">151</span> * the job.<a name="line.151"></a>
+<span class="sourceLineNo">152</span> *<a name="line.152"></a>
+<span class="sourceLineNo">153</span> * @param table The table name to read from.<a name="line.153"></a>
+<span class="sourceLineNo">154</span> * @param scan The scan instance with the columns, time range etc.<a name="line.154"></a>
+<span class="sourceLineNo">155</span> * @param mapper The mapper class to use.<a name="line.155"></a>
+<span class="sourceLineNo">156</span> * @param outputKeyClass The class of the output key.<a name="line.156"></a>
+<span class="sourceLineNo">157</span> * @param outputValueClass The class of the output value.<a name="line.157"></a>
+<span class="sourceLineNo">158</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.158"></a>
+<span class="sourceLineNo">159</span> * carrying all necessary HBase configuration.<a name="line.159"></a>
+<span class="sourceLineNo">160</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.160"></a>
+<span class="sourceLineNo">161</span> * job classes via the distributed cache (tmpjars).<a name="line.161"></a>
+<span class="sourceLineNo">162</span> * @throws IOException When setting up the details fails.<a name="line.162"></a>
+<span class="sourceLineNo">163</span> */<a name="line.163"></a>
+<span class="sourceLineNo">164</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.164"></a>
+<span class="sourceLineNo">165</span> Class<? extends TableMapper> mapper,<a name="line.165"></a>
+<span class="sourceLineNo">166</span> Class<?> outputKeyClass,<a name="line.166"></a>
+<span class="sourceLineNo">167</span> Class<?> outputValueClass, Job job,<a name="line.167"></a>
+<span class="sourceLineNo">168</span> boolean addDependencyJars, Class<? extends InputFormat> inputFormatClass)<a name="line.168"></a>
+<span class="sourceLineNo">169</span> throws IOException {<a name="line.169"></a>
+<span class="sourceLineNo">170</span> initTableMapperJob(table, scan, mapper, outputKeyClass, outputValueClass, job,<a name="line.170"></a>
+<span class="sourceLineNo">171</span> addDependencyJars, true, inputFormatClass);<a name="line.171"></a>
+<span class="sourceLineNo">172</span> }<a name="line.172"></a>
+<span class="sourceLineNo">173</span><a name="line.173"></a>
<span class="sourceLineNo">174</span><a name="line.174"></a>
-<span class="sourceLineNo">175</span><a name="line.175"></a>
-<span class="sourceLineNo">176</span> /**<a name="line.176"></a>
-<span class="sourceLineNo">177</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.177"></a>
-<span class="sourceLineNo">178</span> * the job.<a name="line.178"></a>
-<span class="sourceLineNo">179</span> *<a name="line.179"></a>
-<span class="sourceLineNo">180</span> * @param table The table name to read from.<a name="line.180"></a>
-<span class="sourceLineNo">181</span> * @param scan The scan instance with the columns, time range etc.<a name="line.181"></a>
-<span class="sourceLineNo">182</span> * @param mapper The mapper class to use.<a name="line.182"></a>
-<span class="sourceLineNo">183</span> * @param outputKeyClass The class of the output key.<a name="line.183"></a>
-<span class="sourceLineNo">184</span> * @param outputValueClass The class of the output value.<a name="line.184"></a>
-<span class="sourceLineNo">185</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.185"></a>
-<span class="sourceLineNo">186</span> * carrying all necessary HBase configuration.<a name="line.186"></a>
-<span class="sourceLineNo">187</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.187"></a>
-<span class="sourceLineNo">188</span> * job classes via the distributed cache (tmpjars).<a name="line.188"></a>
-<span class="sourceLineNo">189</span> * @param initCredentials whether to initialize hbase auth credentials for the job<a name="line.189"></a>
-<span class="sourceLineNo">190</span> * @param inputFormatClass the input format<a name="line.190"></a>
-<span class="sourceLineNo">191</span> * @throws IOException When setting up the details fails.<a name="line.191"></a>
-<span class="sourceLineNo">192</span> */<a name="line.192"></a>
-<span class="sourceLineNo">193</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.193"></a>
-<span class="sourceLineNo">194</span> Class<? extends TableMapper> mapper,<a name="line.194"></a>
-<span class="sourceLineNo">195</span> Class<?> outputKeyClass,<a name="line.195"></a>
-<span class="sourceLineNo">196</span> Class<?> outputValueClass, Job job,<a name="line.196"></a>
-<span class="sourceLineNo">197</span> boolean addDependencyJars, boolean initCredentials,<a name="line.197"></a>
-<span class="sourceLineNo">198</span> Class<? extends InputFormat> inputFormatClass)<a name="line.198"></a>
-<span class="sourceLineNo">199</span> throws IOException {<a name="line.199"></a>
-<span class="sourceLineNo">200</span> job.setInputFormatClass(inputFormatClass);<a name="line.200"></a>
-<span class="sourceLineNo">201</span> if (outputValueClass != null) job.setMapOutputValueClass(outputValueClass);<a name="line.201"></a>
-<span class="sourceLineNo">202</span> if (outputKeyClass != null) job.setMapOutputKeyClass(outputKeyClass);<a name="line.202"></a>
-<span class="sourceLineNo">203</span> job.setMapperClass(mapper);<a name="line.203"></a>
-<span class="sourceLineNo">204</span> if (Put.class.equals(outputValueClass)) {<a name="line.204"></a>
-<span class="sourceLineNo">205</span> job.setCombinerClass(PutCombiner.class);<a name="line.205"></a>
-<span class="sourceLineNo">206</span> }<a name="line.206"></a>
-<span class="sourceLineNo">207</span> Configuration conf = job.getConfiguration();<a name="line.207"></a>
-<span class="sourceLineNo">208</span> HBaseConfiguration.merge(conf, HBaseConfiguration.create(conf));<a name="line.208"></a>
-<span class="sourceLineNo">209</span> conf.set(TableInputFormat.INPUT_TABLE, table);<a name="line.209"></a>
-<span class="sourceLineNo">210</span> conf.set(TableInputFormat.SCAN, convertScanToString(scan));<a name="line.210"></a>
-<span class="sourceLineNo">211</span> conf.setStrings("io.serializations", conf.get("io.serializations"),<a name="line.211"></a>
-<span class="sourceLineNo">212</span> MutationSerialization.class.getName(), ResultSerialization.class.getName(),<a name="line.212"></a>
-<span class="sourceLineNo">213</span> KeyValueSerialization.class.getName());<a name="line.213"></a>
-<span class="sourceLineNo">214</span> if (addDependencyJars) {<a name="line.214"></a>
-<span class="sourceLineNo">215</span> addDependencyJars(job);<a name="line.215"></a>
-<span class="sourceLineNo">216</span> }<a name="line.216"></a>
-<span class="sourceLineNo">217</span> if (initCredentials) {<a name="line.217"></a>
-<span class="sourceLineNo">218</span> initCredentials(job);<a name="line.218"></a>
-<span class="sourceLineNo">219</span> }<a name="line.219"></a>
-<span class="sourceLineNo">220</span> }<a name="line.220"></a>
-<span class="sourceLineNo">221</span><a name="line.221"></a>
-<span class="sourceLineNo">222</span> /**<a name="line.222"></a>
-<span class="sourceLineNo">223</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.223"></a>
-<span class="sourceLineNo">224</span> * the job.<a name="line.224"></a>
-<span class="sourceLineNo">225</span> *<a name="line.225"></a>
-<span class="sourceLineNo">226</span> * @param table Binary representation of the table name to read from.<a name="line.226"></a>
-<span class="sourceLineNo">227</span> * @param scan The scan instance with the columns, time range etc.<a name="line.227"></a>
-<span class="sourceLineNo">228</span> * @param mapper The mapper class to use.<a name="line.228"></a>
-<span class="sourceLineNo">229</span> * @param outputKeyClass The class of the output key.<a name="line.229"></a>
-<span class="sourceLineNo">230</span> * @param outputValueClass The class of the output value.<a name="line.230"></a>
-<span class="sourceLineNo">231</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.231"></a>
-<span class="sourceLineNo">232</span> * carrying all necessary HBase configuration.<a name="line.232"></a>
-<span class="sourceLineNo">233</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.233"></a>
-<span class="sourceLineNo">234</span> * job classes via the distributed cache (tmpjars).<a name="line.234"></a>
-<span class="sourceLineNo">235</span> * @param inputFormatClass The class of the input format<a name="line.235"></a>
-<span class="sourceLineNo">236</span> * @throws IOException When setting up the details fails.<a name="line.236"></a>
-<span class="sourceLineNo">237</span> */<a name="line.237"></a>
-<span class="sourceLineNo">238</span> public static void initTableMapperJob(byte[] table, Scan scan,<a name="line.238"></a>
-<span class="sourceLineNo">239</span> Class<? extends TableMapper> mapper,<a name="line.239"></a>
-<span class="sourceLineNo">240</span> Class<?> outputKeyClass,<a name="line.240"></a>
-<span class="sourceLineNo">241</span> Class<?> outputValueClass, Job job,<a name="line.241"></a>
-<span class="sourceLineNo">242</span> boolean addDependencyJars, Class<? extends InputFormat> inputFormatClass)<a name="line.242"></a>
-<span class="sourceLineNo">243</span> throws IOException {<a name="line.243"></a>
-<span class="sourceLineNo">244</span> initTableMapperJob(Bytes.toString(table), scan, mapper, outputKeyClass,<a name="line.244"></a>
-<span class="sourceLineNo">245</span> outputValueClass, job, addDependencyJars, inputFormatClass);<a name="line.245"></a>
-<span class="sourceLineNo">246</span> }<a name="line.246"></a>
-<span class="sourceLineNo">247</span><a name="line.247"></a>
-<span class="sourceLineNo">248</span> /**<a name="line.248"></a>
-<span class="sourceLineNo">249</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.249"></a>
-<span class="sourceLineNo">250</span> * the job.<a name="line.250"></a>
-<span class="sourceLineNo">251</span> *<a name="line.251"></a>
-<span class="sourceLineNo">252</span> * @param table Binary representation of the table name to read from.<a name="line.252"></a>
-<span class="sourceLineNo">253</span> * @param scan The scan instance with the columns, time range etc.<a name="line.253"></a>
-<span class="sourceLineNo">254</span> * @param mapper The mapper class to use.<a name="line.254"></a>
-<span class="sourceLineNo">255</span> * @param outputKeyClass The class of the output key.<a name="line.255"></a>
-<span class="sourceLineNo">256</span> * @param outputValueClass The class of the output value.<a name="line.256"></a>
-<span class="sourceLineNo">257</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.257"></a>
-<span class="sourceLineNo">258</span> * carrying all necessary HBase configuration.<a name="line.258"></a>
-<span class="sourceLineNo">259</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.259"></a>
-<span class="sourceLineNo">260</span> * job classes via the distributed cache (tmpjars).<a name="line.260"></a>
-<span class="sourceLineNo">261</span> * @throws IOException When setting up the details fails.<a name="line.261"></a>
-<span class="sourceLineNo">262</span> */<a name="line.262"></a>
-<span class="sourceLineNo">263</span> public static void initTableMapperJob(byte[] table, Scan scan,<a name="line.263"></a>
-<span class="sourceLineNo">264</span> Class<? extends TableMapper> mapper,<a name="line.264"></a>
-<span class="sourceLineNo">265</span> Class<?> outputKeyClass,<a name="line.265"></a>
-<span class="sourceLineNo">266</span> Class<?> outputValueClass, Job job,<a name="line.266"></a>
-<span class="sourceLineNo">267</span> boolean addDependencyJars)<a name="line.267"></a>
-<span class="sourceLineNo">268</span> throws IOException {<a name="line.268"></a>
-<span class="sourceLineNo">269</span> initTableMapperJob(Bytes.toString(table), scan, mapper, outputKeyClass,<a name="line.269"></a>
-<span class="sourceLineNo">270</span> outputValueClass, job, addDependencyJars, TableInputFormat.class);<a name="line.270"></a>
-<span class="sourceLineNo">271</span> }<a name="line.271"></a>
-<span class="sourceLineNo">272</span><a name="line.272"></a>
-<span class="sourceLineNo">273</span> /**<a name="line.273"></a>
-<span class="sourceLineNo">274</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.274"></a>
-<span class="sourceLineNo">275</span> * the job.<a name="line.275"></a>
-<span class="sourceLineNo">276</span> *<a name="line.276"></a>
-<span class="sourceLineNo">277</span> * @param table The table name to read from.<a name="line.277"></a>
-<span class="sourceLineNo">278</span> * @param scan The scan instance with the columns, time range etc.<a name="line.278"></a>
-<span class="sourceLineNo">279</span> * @param mapper The mapper class to use.<a name="line.279"></a>
-<span class="sourceLineNo">280</span> * @param outputKeyClass The class of the output key.<a name="line.280"></a>
-<span class="sourceLineNo">281</span> * @param outputValueClass The class of the output value.<a name="line.281"></a>
-<span class="sourceLineNo">282</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.282"></a>
-<span class="sourceLineNo">283</span> * carrying all necessary HBase configuration.<a name="line.283"></a>
-<span class="sourceLineNo">284</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.284"></a>
-<span class="sourceLineNo">285</span> * job classes via the distributed cache (tmpjars).<a name="line.285"></a>
-<span class="sourceLineNo">286</span> * @throws IOException When setting up the details fails.<a name="line.286"></a>
-<span class="sourceLineNo">287</span> */<a name="line.287"></a>
-<span class="sourceLineNo">288</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.288"></a>
-<span class="sourceLineNo">289</span> Class<? extends TableMapper> mapper,<a name="line.289"></a>
-<span class="sourceLineNo">290</span> Class<?> outputKeyClass,<a name="line.290"></a>
-<span class="sourceLineNo">291</span> Class<?> outputValueClass, Job job,<a name="line.291"></a>
-<span class="sourceLineNo">292</span> boolean addDependencyJars)<a name="line.292"></a>
-<span class="sourceLineNo">293</span> throws IOException {<a name="line.293"></a>
-<span class="sourceLineNo">294</span> initTableMapperJob(table, scan, mapper, outputKeyClass,<a name="line.294"></a>
-<span class="sourceLineNo">295</span> outputValueClass, job, addDependencyJars, TableInputFormat.class);<a name="line.295"></a>
-<span class="sourceLineNo">296</span> }<a name="line.296"></a>
-<span class="sourceLineNo">297</span><a name="line.297"></a>
-<span class="sourceLineNo">298</span> /**<a name="line.298"></a>
-<span class="sourceLineNo">299</span> * Enable a basic on-heap cache for these jobs. Any BlockCache implementation based on<a name="line.299"></a>
-<span class="sourceLineNo">300</span> * direct memory will likely cause the map tasks to OOM when opening the region. This<a name="line.300"></a>
-<span class="sourceLineNo">301</span> * is done here instead of in TableSnapshotRegionRecordReader in case an advanced user<a name="line.301"></a>
-<span class="sourceLineNo">302</span> * wants to override this behavior in their job.<a name="line.302"></a>
-<span class="sourceLineNo">303</span> */<a name="line.303"></a>
-<span class="sourceLineNo">304</span> public static void resetCacheConfig(Configuration conf) {<a name="line.304"></a>
-<span class="sourceLineNo">305</span> conf.setFloat(<a name="line.305"></a>
-<span class="sourceLineNo">306</span> HConstants.HFILE_BLOCK_CACHE_SIZE_KEY, HConstants.HFILE_BLOCK_CACHE_SIZE_DEFAULT);<a name="line.306"></a>
-<span class="sourceLineNo">307</span> conf.setFloat(HConstants.BUCKET_CACHE_SIZE_KEY, 0f);<a name="line.307"></a>
-<span class="sourceLineNo">308</span> conf.unset(HConstants.BUCKET_CACHE_IOENGINE_KEY);<a name="line.308"></a>
-<span class="sourceLineNo">309</span> }<a name="line.309"></a>
-<span class="sourceLineNo">310</span><a name="line.310"></a>
-<span class="sourceLineNo">311</span> /**<a name="line.311"></a>
-<span class="sourceLineNo">312</span> * Sets up the job for reading from one or more table snapshots, with one or more scans<a name="line.312"></a>
-<span class="sourceLineNo">313</span> * per snapshot.<a name="line.313"></a>
-<span class="sourceLineNo">314</span> * It bypasses hbase servers and read directly from snapshot files.<a name="line.314"></a>
-<span class="sourceLineNo">315</span> *<a name="line.315"></a>
-<span class="sourceLineNo">316</span> * @param snapshotScans map of snapshot name to scans on that snapshot.<a name="line.316"></a>
-<span class="sourceLineNo">317</span> * @param mapper The mapper class to use.<a name="line.317"></a>
-<span class="sourceLineNo">318</span> * @param outputKeyClass The class of the output key.<a name="line.318"></a>
-<span class="sourceLineNo">319</span> * @param outputValueClass The class of the output value.<a name="line.319"></a>
-<span class="sourceLineNo">320</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.320"></a>
-<span class="sourceLineNo">321</span> * carrying all necessary HBase configuration.<a name="line.321"></a>
-<span class="sourceLineNo">322</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.322"></a>
-<span class="sourceLineNo">323</span> * job classes via the distributed cache (tmpjars).<a name="line.323"></a>
-<span class="sourceLineNo">324</span> */<a name="line.324"></a>
-<span class="sourceLineNo">325</span> public static void initMultiTableSnapshotMapperJob(Map<String, Collection<Scan>> snapshotScans,<a name="line.325"></a>
-<span class="sourceLineNo">326</span> Class<? extends TableMapper> mapper, Class<?> outputKeyClass, Class<?> outputValueClass,<a name="line.326"></a>
-<span class="sourceLineNo">327</span> Job job, boolean addDependencyJars, Path tmpRestoreDir) throws IOException {<a name="line.327"></a>
-<span class="sourceLineNo">328</span> MultiTableSnapshotInputFormat.setInput(job.getConfiguration(), snapshotScans, tmpRestoreDir);<a name="line.328"></a>
-<span class="sourceLineNo">329</span><a name="line.329"></a>
-<span class="sourceLineNo">330</span> job.setInputFormatClass(MultiTableSnapshotInputFormat.class);<a name="line.330"></a>
-<span class="sourceLineNo">331</span> if (outputValueClass != null) {<a name="line.331"></a>
-<span class="sourceLineNo">332</span> job.setMapOutputValueClass(outputValueClass);<a name="line.332"></a>
-<span class="sourceLineNo">333</span> }<a name="line.333"></a>
-<span class="sourceLineNo">334</span> if (outputKeyClass != null) {<a name="line.334"></a>
-<span class="sourceLineNo">335</span> job.setMapOutputKeyClass(outputKeyClass);<a name="line.335"></a>
-<span class="sourceLineNo">336</span> }<a name="line.336"></a>
-<span class="sourceLineNo">337</span> job.setMapperClass(mapper);<a name="line.337"></a>
-<span class="sourceLineNo">338</span> Configuration conf = job.getConfiguration();<a name="line.338"></a>
-<span class="sourceLineNo">339</span> HBaseConfiguration.merge(conf, HBaseConfiguration.create(conf));<a name="line.339"></a>
-<span class="sourceLineNo">340</span><a name="line.340"></a>
-<span class="sourceLineNo">341</span> if (addDependencyJars) {<a name="line.341"></a>
-<span class="sourceLineNo">342</span> addDependencyJars(job);<a name="line.342"></a>
-<span class="sourceLineNo">343</span> addDependencyJars(job.getConfiguration(), MetricsRegistry.class);<a name="line.343"></a>
-<span class="sourceLineNo">344</span> }<a name="line.344"></a>
-<span class="sourceLineNo">345</span><a name="line.345"></a>
-<span class="sourceLineNo">346</span> resetCacheConfig(job.getConfiguration());<a name="line.346"></a>
-<span class="sourceLineNo">347</span> }<a name="line.347"></a>
-<span class="sourceLineNo">348</span><a name="line.348"></a>
-<span class="sourceLineNo">349</span> /**<a name="line.349"></a>
-<span class="sourceLineNo">350</span> * Sets up the job for reading from a table snapshot. It bypasses hbase servers<a name="line.350"></a>
-<span class="sourceLineNo">351</span> * and read directly from snapshot files.<a name="line.351"></a>
-<span class="sourceLineNo">352</span> *<a name="line.352"></a>
-<span class="sourceLineNo">353</span> * @param snapshotName The name of the snapshot (of a table) to read from.<a name="line.353"></a>
-<span class="sourceLineNo">354</span> * @param scan The scan instance with the columns, time range etc.<a name="line.354"></a>
-<span class="sourceLineNo">355</span> * @param mapper The mapper class to use.<a name="line.355"></a>
-<span class="sourceLineNo">356</span> * @param outputKeyClass The class of the output key.<a name="line.356"></a>
-<span class="sourceLineNo">357</span> * @param outputValueClass The class of the output value.<a name="line.357"></a>
-<span class="sourceLineNo">358</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.358"></a>
-<span class="sourceLineNo">359</span> * carrying all necessary HBase configuration.<a name="line.359"></a>
-<span class="sourceLineNo">360</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.360"></a>
-<span class="sourceLineNo">361</span> * job classes via the distributed cache (tmpjars).<a name="line.361"></a>
-<span class="sourceLineNo">362</span> *<a name="line.362"></a>
-<span class="sourceLineNo">363</span> * @param tmpRestoreDir a temporary directory to copy the snapshot files into. Current user should<a name="line.363"></a>
-<span class="sourceLineNo">364</span> * have write permissions to this directory, and this should not be a subdirectory of rootdir.<a name="line.364"></a>
-<span class="sourceLineNo">365</span> * After the job is finished, restore directory can be deleted.<a name="line.365"></a>
-<span class="sourceLineNo">366</span> * @throws IOException When setting up the details fails.<a name="line.366"></a>
-<span class="sourceLineNo">367</span> * @see TableSnapshotInputFormat<a name="line.367"></a>
-<span class="sourceLineNo">368</span> */<a name="line.368"></a>
-<span class="sourceLineNo">369</span> public static void initTableSnapshotMapperJob(String snapshotName, Scan scan,<a name="line.369"></a>
-<span class="sourceLineNo">370</span> Class<? extends TableMapper> mapper,<a name="line.370"></a>
-<span class="sourceLineNo">371</span> Class<?> outputKeyClass,<a name="line.371"></a>
-<span class="sourceLineNo">372</span> Class<?> outputValueClass, Job job,<a name="line.372"></a>
-<span class="sourceLineNo">373</span> boolean addDependencyJars, Path tmpRestoreDir)<a name="line.373"></a>
-<span class="sourceLineNo">374</span> throws IOException {<a name="line.374"></a>
-<span class="sourceLineNo">375</span> TableSnapshotInputFormat.setInput(job, snapshotName, tmpRestoreDir);<a name="line.375"></a>
-<span class="sourceLineNo">376</span> initTableMapperJob(snapshotName, scan, mapper, outputKeyClass,<a name="line.376"></a>
-<span class="sourceLineNo">377</span> outputValueClass, job, addDependencyJars, false, TableSnapshotInputFormat.class);<a name="line.377"></a>
-<span class="sourceLineNo">378</span> resetCacheConfig(job.getConfiguration());<a name="line.378"></a>
-<span class="sourceLineNo">379</span> }<a name="line.379"></a>
-<span class="sourceLineNo">380</span><a name="line.380"></a>
-<span class="sourceLineNo">381</span> /**<a name="line.381"></a>
-<span class="sourceLineNo">382</span> * Use this before submitting a Multi TableMap job. It will appropriately set<a name="line.382"></a>
-<span class="sourceLineNo">383</span> * up the job.<a name="line.383"></a>
-<span class="sourceLineNo">384</span> *<a name="line.384"></a>
-<span class="sourceLineNo">385</span> * @param scans The list of {@link Scan} objects to read from.<a name="line.385"></a>
-<span class="sourceLineNo">386</span> * @param mapper The mapper class to use.<a name="line.386"></a>
-<span class="sourceLineNo">387</span> * @param outputKeyClass The class of the output key.<a name="line.387"></a>
-<span class="sourceLineNo">388</span> * @param outputValueClass The class of the output value.<a name="line.388"></a>
-<span class="sourceLineNo">389</span> * @param job The current job to adjust. Make sure the passed job is carrying<a name="line.389"></a>
-<span class="sourceLineNo">390</span> * all necessary HBase configuration.<a name="line.390"></a>
-<span class="sourceLineNo">391</span> * @throws IOException When setting up the details fails.<a name="line.391"></a>
-<span class="sourceLineNo">392</span> */<a name="line.392"></a>
-<span class="sourceLineNo">393</span> public static void initTableMapperJob(List<Scan> scans,<a name="line.393"></a>
-<span class="sourceLineNo">394</span> Class<? extends TableMapper> mapper,<a name="line.394"></a>
-<span class="sourceLineNo">395</span> Class<?> outputKeyClass,<a name="line.395"></a>
-<span class="sourceLineNo">396</span> Class<?> outputValueClass, Job job) throws IOException {<a name="line.396"></a>
-<span class="sourceLineNo">397</span> initTableMapperJob(scans, mapper, outputKeyClass, outputValueClass, job,<a name="line.397"></a>
-<span class="sourceLineNo">398</span> true);<a name="line.398"></a>
-<span class="sourceLineNo">399</span> }<a name="line.399"></a>
-<span class="sourceLineNo">400</span><a name="line.400"></a>
-<span class="sourceLineNo">401</span> /**<a name="line.401"></a>
-<span class="sourceLineNo">402</span> * Use this before submitting a Multi TableMap job. It will appropriately set<a name="line.402"></a>
-<span class="sourceLineNo">403</span> * up the job.<a name="line.403"></a>
-<span class="sourceLineNo">404</span> *<a name="line.404"></a>
-<span class="sourceLineNo">405</span> * @param scans The list of {@link Scan} objects to read from.<a name="line.405"></a>
-<span class="sourceLineNo">406</span> * @param mapper The mapper class to use.<a name="line.406"></a>
-<span class="sourceLineNo">407</span> * @param outputKeyClass The class of the output key.<a name="line.407"></a>
-<span class="sourceLineNo">408</span> * @param outputValueClass The class of the output value.<a name="line.408"></a>
-<span class="sourceLineNo">409</span> * @param job The current job to adjust. Make sure the passed job is carrying<a name="line.409"></a>
-<span class="sourceLineNo">410</span> * all necessary HBase configuration.<a name="line.410"></a>
-<span class="sourceLineNo">411</span> * @param addDependencyJars upload HBase jars and jars for any of the<a name="line.411"></a>
-<span class="sourceLineNo">412</span> * configured job classes via the distributed cache (tmpjars).<a name="line.412"></a>
-<span class="sourceLineNo">413</span> * @throws IOException When setting up the details fails.<a name="line.413"></a>
-<span class="sourceLineNo">414</span> */<a name="line.414"></a>
-<span class="sourceLineNo">415</span> public static void initTableMapperJob(List<Scan> scans,<a name="line.415"></a>
-<span class="sourceLineNo">416</span> Class<? extends TableMapper> mapper,<a name="line.416"></a>
-<span class="sourceLineNo">417</span> Class<?> outputKeyClass,<a name="line.417"></a>
-<span class="sourceLineNo">418</span> Class<?> outputValueClass, Job job,<a name="line.418"></a>
-<span class="sourceLineNo">419</span> boolean addDependencyJars) throws IOException {<a name="line.419"></a>
-<span class="sourceLineNo">420</span> initTableMapperJob(scans, mapper, outputKeyClass, outputValueClass, job,<a name="line.420"></a>
-<span class="sourceLineNo">421</span> addDependencyJars, true);<a name="line.421"></a>
-<span class="sourceLineNo">422</span> }<a name="line.422"></a>
-<span class="sourceLineNo">423</span><a name="line.423"></a>
-<span class="sourceLineNo">424</span> /**<a name="line.424"></a>
-<span class="sourceLineNo">425</span> * Use this before submitting a Multi TableMap job. It will appropriately set<a name="line.425"></a>
-<span class="sourceLineNo">426</span> * up the job.<a name="line.426"></a>
-<span class="sourceLineNo">427</span> *<a name="line.427"></a>
-<span class="sourceLineNo">428</span> * @param scans The list of {@link Scan} objects to read from.<a name="line.428"></a>
-<span class="sourceLineNo">429</span> * @param mapper The mapper class to use.<a name="line.429"></a>
-<span class="sourceLineNo">430</span> * @param outputKeyClass The class of the output key.<a name="line.430"></a>
-<span class="sourceLineNo">431</span> * @param outputValueClass The class of the output value.<a name="line.431"></a>
-<span class="sourceLineNo">432</span> * @param job The current job to adjust. Make sure the passed job is carrying<a name="line.432"></a>
-<span class="sourceLineNo">433</span> * all necessary HBase configuration.<a name="line.433"></a>
-<span class="sourceLineNo">434</span> * @param addDependencyJars upload HBase jars and jars for any of the<a name="line.434"></a>
-<span class="sourceLineNo">435</span> * configured job classes via the distributed cache (tmpjars).<a name="line.435"></a>
-<span class="sourceLineNo">436</span> * @param initCredentials whether to initialize hbase auth credentials for the job<a name="line.436"></a>
-<span class="sourceLineNo">437</span> * @throws IOException When setting up the details fails.<a name="line.437"></a>
-<span class="sourceLineNo">438</span> */<a name="line.438"></a>
-<span class="sourceLineNo">439</span> public static void initTableMapperJob(List<Scan> scans,<a name="line.439"></a>
-<span class="sourceLineNo">440</span> Class<? extends TableMapper> mapper,<a name="line.440"></a>
-<span class="sourceLineNo">441</span> Class<?> outputKeyClass,<a name="line.441"></a>
-<span class="sourceLineNo">442</span> Class<?> outputValueClass, Job job,<a name="line.442"></a>
-<span class="sourceLineNo">443</span> boolean addDependencyJars,<a name="line.443"></a>
-<span class="sourceLineNo">444</span> boolean initCredentials) throws IOException {<a name="line.444"></a>
-<span class="sourceLineNo">445</span> job.setInputFormatClass(MultiTableInputFormat.class);<a name="line.445"></a>
-<span class="sourceLineNo">446</span> if (outputValueClass != null) {<a name="line.446"></a>
-<span class="sourceLineNo">447</span> job.setMapOutputValueClass(outputValueClass);<a name="line.447"></a>
-<span class="sourceLineNo">448</span> }<a name="line.448"></a>
-<span class="sourceLineNo">449</span> if (outputKeyClass != null) {<a name="line.449"></a>
-<span class="sourceLineNo">450</span> job.setMapOutputKeyClass(outputKeyClass);<a name="line.450"></a>
-<span class="sourceLineNo">451</span> }<a name="line.451"></a>
-<span class="sourceLineNo">452</span> job.setMapperClass(mapper);<a name="line.452"></a>
-<span class="sourceLineNo">453</span> Configuration conf = job.getConfiguration();<a name="line.453"></a>
-<span class="sourceLineNo">454</span> HBaseConfiguration.merge(conf, HBaseConfiguration.create(conf));<a name="line.454"></a>
-<span class="sourceLineNo">455</span> List<String> scanStrings = new ArrayList<String>();<a name="line.455"></a>
-<span class="sourceLineNo">456</span><a name="line.456"></a>
-<span class="sourceLineNo">457</span> for (Scan scan : scans) {<a name="line.457"></a>
-<span class="sourceLineNo">458</span> scanStrings.add(convertScanToString(scan));<a name="line.458"></a>
-<span class="sourceLineNo">459</span> }<a name="line.459"></a>
-<span class="sourceLineNo">460</span> job.getConfiguration().setStrings(MultiTableInputFormat.SCANS,<a name="line.460"></a>
-<span class="sourceLineNo">461</span> scanStrings.toArray(new String[scanStrings.size()]));<a name="line.461"></a>
-<span class="sourceLineNo">462</span><a name="line.462"></a>
-<span class="sourceLineNo">463</span> if (addDependencyJars) {<a name="line.463"></a>
-<span class="sourceLineNo">464</span> addDependencyJars(job);<a name="line.464"></a>
-<span class="sourceLineNo">465</span> }<a name="line.465"></a>
-<span class="sourceLineNo">466</span><a name="line.466"></a>
-<span class="sourceLineNo">467</span> if (initCredentials) {<a name="line.467"></a>
-<span class="sourceLineNo">468</span> initCredentials(job);<a name="line.468"></a>
-<span class="sourceLineNo">469</span> }<a name="line.469"></a>
-<span class="sourceLineNo">470</span> }<a name="line.470"></a>
-<span class="sourceLineNo">471</span><a name="line.471"></a>
-<span class="sourceLineNo">472</span> public static void initCredentials(Job job) throws IOException {<a name="line.472"></a>
-<span class="sourceLineNo">473</span> UserProvider userProvider = UserProvider.instantiate(job.getConfiguration());<a name="line.473"></a>
-<span class="sourceLineNo">474</span> if (userProvider.isHadoopSecurityEnabled()) {<a name="line.474"></a>
-<span class="sourceLineNo">475</span> // propagate delegation related props from launcher job to MR job<a name="line.475"></a>
-<span class="sourceLineNo">476</span> if (System.getenv("HADOOP_TOKEN_FILE_LOCATION") != null) {<a name="line.476"></a>
-<span class="sourceLineNo">477</span> job.getConfiguration().set("mapreduce.job.credentials.binary",<a name="line.477"></a>
-<span class="sourceLineNo">478</span> System.getenv("HADOOP_TOKEN_FILE_LOCATION"));<a name="line.478"></a>
-<span class="sourceLineNo">479</span> }<a name="line.479"></a>
-<span class="sourceLineNo">480</span> }<a name="line.480"></a>
-<span class="sourceLineNo">481</span><a name="line.481"></a>
-<span class="sourceLineNo">482</span> if (userProvider.isHBaseSecurityEnabled()) {<a name="line.482"></a>
-<span class="sourceLineNo">483</span> try {<a name="line.483"></a>
-<span class="sourceLineNo">484</span> // init credentials for remote cluster<a name="line.484"></a>
-<span class="sourceLineNo">485</span> String quorumAddress = job.getConfiguration().get(TableOutputFormat.QUORUM_ADDRESS);<a name="line.485"></a>
-<span class="sourceLineNo">486</span> User user = userProvider.getCurrent();<a name="line.486"></a>
-<span class="sourceLineNo">487</span> if (quorumAddress != null) {<a name="line.487"></a>
-<span class="sourceLineNo">488</span> Configuration peerConf = HBaseConfiguration.create(job.getConfiguration());<a name="line.488"></a>
-<span class="sourceLineNo">489</span> ZKUtil.applyClusterKeyToConf(peerConf, quorumAddress);<a name="line.489"></a>
-<span class="sourceLineNo">490</span> // apply any "hbase.mapred.output." configuration overrides<a name="line.490"></a>
-<span class="sourceLineNo">491</span> Configuration outputOverrides =<a name="line.491"></a>
-<span class="sourceLineNo">492</span> HBaseConfiguration.subset(peerConf, TableOutputFormat.OUTPUT_CONF_PREFIX);<a name="line.492"></a>
-<span class="sourceLineNo">493</span> HBaseConfiguration.merge(peerConf, outputOverrides);<a name="line.493"></a>
-<span class="sourceLineNo">494</span> Connection peerConn = ConnectionFactory.createConnection(peerConf);<a name="line.494"></a>
-<span class="sourceLineNo">495</span> try {<a name="line.495"></a>
-<span class="sourceLineNo">496</span> TokenUtil.addTokenForJob(peerConn, user, job);<a name="line.496"></a>
-<span class="sourceLineNo">497</span> } finally {<a name="line.497"></a>
-<span class="sourceLineNo">498</span> peerConn.close();<a name="line.498"></a>
-<span class="sourceLineNo">499</span> }<a name="line.499"></a>
-<span class="sourceLineNo">500</span> }<a name="line.500"></a>
-<span class="sourceLineNo">501</span><a name="line.501"></a>
-<span class="sourceLineNo">502</span> Connection conn = ConnectionFactory.createConnection(job.getConfiguration());<a name="line.502"></a>
-<span class="sourceLineNo">503</span> try {<a name="line.503"></a>
-<span class="sourceLineNo">504</span> TokenUtil.addTokenForJob(conn, user, job);<a name="line.504"></a>
-<span class="sourceLineNo">505</span> } finally {<a name="line.505"></a>
-<span class="sourceLineNo">506</span> conn.close();<a name="line.506"></a>
-<span class="sourceLineNo">507</span> }<a name="line.507"></a>
-<span class="sourceLineNo">508</span> } catch (InterruptedException ie) {<a name="line.508"></a>
-<span class="sourceLineNo">509</span> LOG.info("Interrupted obtaining user authentication token");<a name="line.509"></a>
-<span class="sourceLineNo">510</span> Thread.currentThread().interrupt();<a name="line.510"></a>
-<span class="sourceLineNo">511</span> }<a name="line.511"></a>
-<span class="sourceLineNo">512</span> }<a name="line.512"></a>
-<span class="sourceLineNo">513</span> }<a name="line.513"></a>
-<span class="sourceLineNo">514</span><a name="line.514"></a>
-<span class="sourceLineNo">515</span> /**<a name="line.515"></a>
-<span class="sourceLineNo">516</span> * Obtain an authentication token, for the specified cluster, on behalf of the current user<a name="line.516"></a>
-<span class="sourceLineNo">517</span> * and add it to the credentials for the given map reduce job.<a name="line.517"></a>
-<span class="sourceLineNo">518</span> *<a name="line.518"></a>
-<span class="sourceLineNo">519</span> * The quorumAddress is the key to the ZK ensemble, which contains:<a name="line.519"></a>
-<span class="sourceLineNo">520</span> * hbase.zookeeper.quorum, hbase.zookeeper.client.port and<a name="line.520"></a>
-<span class="sourceLineNo">521</span> * zookeeper.znode.parent<a name="line.521"></a>
-<span class="sourceLineNo">522</span> *<a name="line.522"></a>
-<span class="sourceLineNo">523</span> * @param job The job that requires the permission.<a name="line.523"></a>
-<span class="sourceLineNo">524</span> * @param quorumAddress string that contains the 3 required configuratins<a name="line.524"></a>
-<span class="sourceLineNo">525</span> * @throws IOException When the authentication token cannot be obtained.<a name="line.525"></a>
-<span class="sourceLineNo">526</span> */<a name="line.526"></a>
-<span class="sourceLineNo">527</span> public static void initCredentialsForCluster(Job job, String quorumAddress)<a name="line.527"></a>
-<span class="sourceLineNo">528</span> throws IOException {<a name="line.528"></a>
-<span class="sourceLineNo">529</span> UserProvider userProvider = UserProvider.instantiate(job.getConfiguration());<a name="line.529"></a>
-<span class="sourceLineNo">530</span> if (userProvider.isHBaseSecurityEnabled()) {<a name="line.530"></a>
-<span class="sourceLineNo">531</span> try {<a name="line.531"></a>
-<span class="sourceLineNo">532</span> Configuration peerConf = HBaseConfiguration.create(job.getConfiguration());<a name="line.532"></a>
-<span class="sourceLineNo">533</span> ZKUtil.applyClusterKeyToConf(peerConf, quorumAddress);<a name="line.533"></a>
-<span class="sourceLineNo">534</span> Connection peerConn = ConnectionFactory.createConnection(peerConf);<a name="line.534"></a>
-<span class="sourceLineNo">535</span> try {<a name="line.535"></a>
-<span class="sourceLineNo">536</span> TokenUtil.addTokenForJob(peerConn, userProvider.getCurrent(), job);<a name="line.536"></a>
-<span class="sourceLineNo">537</span> } finally {<a name="line.537"></a>
-<span class="sourceLineNo">538</span> peerConn.close();<a name="line.538"></a>
-<span class="sourceLineNo">539</span> }<a name="line.539"></a>
-<span class="sourceLineNo">540</span> } catch (InterruptedException e) {<a name="line.540"></a>
-<span class="sourceLineNo">541</span> LOG.info("Interrupted obtaining user authentication token");<a name="line.541"></a>
-<span class="sourceLineNo">542</span> Thread.interrupted();<a name="line.542"></a>
-<span class="sourceLineNo">543</span> }<a name="line.543"></a>
-<span class="sourceLineNo">544</span> }<a name="line.544"></a>
-<span class="sourceLineNo">545</span> }<a name="line.545"></a>
-<span class="sourceLineNo">546</span><a name="line.546"></a>
-<span class="sourceLineNo">547</span> /**<a name="line.547"></a>
-<span class="sourceLineNo">548</span> * Writes the given scan into a Base64 encoded string.<a name="line.548"></a>
-<span class="sourceLineNo">549</span> *<a name="line.549"></a>
-<span class="sourceLineNo">550</span> * @param scan The scan to write out.<a name="line.550"></a>
-<span class="sourceLineNo">551</span> * @return The scan saved in a Base64 encoded string.<a name="line.551"></a>
-<span class="sourceLineNo">552</span> * @throws IOException When writing the scan fails.<a name="line.552"></a>
-<span class="sourceLineNo">553</span> */<a name="line.553"></a>
-<span class="sourceLineNo">554</span> static String convertScanToString(Scan scan) throws IOException {<a name="line.554"></a>
-<span class="sourceLineNo">555</span> ClientProtos.Scan proto = ProtobufUtil.toScan(scan);<a name="line.555"></a>
-<span class="sourceLineNo">556</span> return Base64.encodeBytes(proto.toByteArray());<a name="line.556"></a>
-<span class="sourceLineNo">557</span> }<a name="line.557"></a>
-<span class="sourceLineNo">558</span><a name="line.558"></a>
-<span class="sourceLineNo">559</span> /**<a name="line.559"></a>
-<span class="sourceLineNo">560</span> * Converts the given Base64 string back into a Scan instance.<a name="line.560"></a>
-<span class="sourceLineNo">561</span> *<a name="line.561"></a>
-<span class="sourceLineNo">562</span> * @param base64 The scan details.<a name="line.562"></a>
-<span class="sourceLineNo">563</span> * @return The newly created Scan instance.<a name="line.563"></a>
-<span class="sourceLineNo">564</span> * @throws IOException When reading the scan instance fails.<a name="line.564"></a>
-<span class="sourceLineNo">565</span> */<a name="line.565"></a>
-<span class="sourceLineNo">566</span> static Scan convertStringToScan(String base64) throws IOException {<a name="line.566"></a>
-<span class="sourceLineNo">567</span> byte [] decoded = Base64.decode(base64);<a name="line.567"></a>
-<span class="sourceLineNo">568</span> ClientProtos.Scan scan;<a name="line.568"></a>
-<span class="sourceLineNo">569</span> try {<a name="line.569"></a>
-<span class="sourceLineNo">570</span> scan = ClientProtos.Scan.parseFrom(decoded);<a name="line.570"></a>
-<span class="sourceLineNo">571</span> } catch (InvalidProtocolBufferException ipbe) {<a name="line.571"></a>
-<span class="sourceLineNo">572</span> throw new IOException(ipbe);<a name="line.572"></a>
-<span class="sourceLineNo">573</span> }<a name="line.573"></a>
-<span class="sourceLineNo">574</span><a name="line.574"></a>
-<span class="sourceLineNo">575</span> return ProtobufUtil.toScan(scan);<a name="line.575"></a>
-<span class="sourceLineNo">576</span> }<a name="line.576"></a>
-<span class="sourceLineNo">577</span><a name="line.577"></a>
-<span class="sourceLineNo">578</span> /**<a name="line.578"></a>
-<span class="sourceLineNo">579</span> * Use this before submitting a TableReduce job. It will<a name="line.579"></a>
-<span class="sourceLineNo">580</span> * appropriately set up the JobConf.<a name="line.580"></a>
-<span class="sourceLineNo">581</span> *<a name="line.581"></a>
-<span class="sourceLineNo">582</span> * @param table The output table.<a name="line.582"></a>
-<span class="sourceLineNo">583</span> * @param reducer The reducer class to use.<a name="line.583"></a>
-<span class="sourceLineNo">584</span> * @param job The current job to adjust.<a name="line.584"></a>
-<span class="sourceLineNo">585</span> * @throws IOException When determining the region count fails.<a name="line.585"></a>
-<span class="sourceLineNo">586</span> */<a name="line.586"></a>
-<span class="sourceLineNo">587</span> public static void initTableReducerJob(String table,<a name="line.587"></a>
-<span class="sourceLineNo">588</span> Class<? extends TableReducer> reducer, Job job)<a name="line.588"></a>
-<span class="sourceLineNo">589</span> throws IOException {<a name="line.589"></a>
-<span class="sourceLineNo">590</span> initTableReducerJob(table, reducer, job, null);<a name="line.590"></a>
-<span class="sourceLineNo">591</span> }<a name="line.591"></a>
-<span class="sourceLineNo">592</span><a name="line.592"></a>
-<span class="sourceLineNo">593</span> /**<a name="line.593"></a>
-<span class="sourceLineNo">594</span> * Use this before submitting a TableReduce job. It will<a name="line.594"></a>
-<span class="sourceLineNo">595</span> * appropriately set up the JobConf.<a name="line.595"></a>
-<span class="sourceLineNo">596</span> *<a name="line.596"></a>
-<span class="sourceLineNo">597</span> * @param table The output table.<a name="line.597"></a>
-<span class="sourceLineNo">598</span> * @param reducer The reducer class to use.<a name="line.598"></a>
-<span class="sourceLineNo">599</span> * @param job The current job to adjust.<a name="line.599"></a>
-<span class="sourceLineNo">600</span> * @param partitioner Partitioner to use. Pass <code>null</code> to use<a name="line.600"></a>
-<span class="sourceLineNo">601</span> * default partitioner.<a name="line.601"></a>
-<span class="sourceLineNo">602</span> * @throws IOException When determining the region count fails.<a name="line.602"></a>
-<span class="sourceLineNo">603</span> */<a name="line.603"></a>
-<span class="sourceLineNo">604</span> public static void initTableReducerJob(String table,<a name="line.604"></a>
-<span class="sourceLineNo">605</span> Class<? extends TableReducer> reducer, Job job,<a name="line.605"></a>
-<span class="sourceLineNo">606</span> Class partitioner) throws IOException {<a name="line.606"></a>
-<span class="sourceLineNo">607</span> initTableReducerJob(table, reducer, job, partitioner, null, null, null);<a name="line.607"></a>
-<span class="sourceLineNo">608</span> }<a name="line.608"></a>
-<span class="sourceLineNo">609</span><a name="line.609"></a>
-<span class="sourceLineNo">610</span> /**<a name="line.610"></a>
-<span class="sourceLineNo">611</span> * Use this before submitting a TableReduce job. It will<a name="line.611"></a>
-<span class="sourceLineNo">612</span> * appropriately set up the JobConf.<a name="line.612"></a>
-<span class="sourceLineNo">613</span> *<a name="line.613"></a>
-<span class="sourceLineNo">614</span> * @param table The output table.<a name="line.614"></a>
-<span class="sourceLineNo">615</span> * @param reducer The reducer class to use.<a name="line.615"></a>
-<span class="sourceLineNo">616</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.616"></a>
-<span class="sourceLineNo">617</span> * carrying all necessary HBase configuration.<a name="line.617"></a>
-<span class="sourceLineNo">618</span> * @param partitioner Partitioner to use. Pass <code>null</code> to use<a name="line.618"></a>
-<span class="sourceLineNo">619</span> * default partitioner.<a name="line.619"></a>
-<span class="sourceLineNo">620</span> * @param quorumAddress Distant cluster to write to; default is null for<a name="line.620"></a>
-<span class="sourceLineNo">621</span> * output to the cluster that is designated in <code>hbase-site.xml</code>.<a name="line.621"></a>
-<span class="sourceLineNo">622</span> * Set this String to the zookeeper ensemble of an alternate remote cluster<a name="line.622"></a>
-<span class="sourceLineNo">623</span> * when you would have the reduce write a cluster that is other than the<a name="line.623"></a>
-<span class="sourceLineNo">624</span> * default; e.g. copying tables between clusters, the source would be<a name="line.624"></a>
-<span class="sourceLineNo">625</span> * designated by <code>hbase-site.xml</code> and this param would have the<a name="line.625"></a>
-<span class="sourceLineNo">626</span> * ensemble address of the remote cluster. The format to pass is particular.<a name="line.626"></a>
-<span class="sourceLineNo">627</span> * Pass <code> &lt;hbase.zookeeper.quorum&gt;:&lt;<a name="line.627"></a>
-<span class="sourceLineNo">628</span> * hbase.zookeeper.client.port&gt;:&lt;zookeeper.znode.parent&gt;<a name="line.628"></a>
-<span class="sourceLineNo">629</span> * </code> such as <code>server,server2,server3:2181:/hbase</code>.<a name="line.629"></a>
-<span class="sourceLineNo">630</span> * @param serverClass redefined hbase.regionserver.class<a name="line.630"></a>
-<span class="sourceLineNo">631</span> * @param serverImpl redefined hbase.regionserver.impl<a name="line.631"></a>
-<span class="sourceLineNo">632</span> * @throws IOException When determining the region count fails.<a name="line.632"></a>
-<span class="sourceLineNo">633</span> */<a name="line.633"></a>
-<span class="sourceLineNo">634</span> public static void initTableReducerJob(String table,<a name="line.634"></a>
-<span class="sourceLineNo">635</span> Class<? extends TableReducer> reducer, Job job,<a name="line.635"></a>
-<span class="sourceLineNo">636</span> Class partitioner, String quorumAddress, String serverClass,<a name="line.636"></a>
-<span class="sourceLineNo">637</span> String serverImpl) throws IOException {<a name="line.637"></a>
-<span class="sourceLineNo">638</span> initTableReducerJob(table, reducer, job, partitioner, quorumAddress,<a name="line.638"></a>
-<span class="sourceLineNo">639</span> serverClass, serverImpl, true);<a name="line.639"></a>
-<span class="sourceLineNo">640</span> }<a name="line.640"></a>
-<span class="sourceLineNo">641</span><a name="line.641"></a>
-<span class="sourceLineNo">642</span> /**<a name="line.642"></a>
-<span class="sourceLineNo">643</span> * Use this before submitting a TableReduce job. It will<a name="line.643"></a>
-<span class="sourceLineNo">644</span> * appropriately set up the JobConf.<a name="line.644"></a>
-<span class="sourceLineNo">645</span> *<a name="line.645"></a>
-<span class="sourceLineNo">646</span> * @param table The output table.<a name="line.646"></a>
-<span class="sourceLineNo">647</span> * @param reducer The reducer class to use.<a name="line.647"></a>
-<span class="sourceLineNo">648</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.648"></a>
-<span class="sourceLineNo">649</span> * carrying all necessary HBase configuration.<a name="line.649"></a>
-<span class="sourceLineNo">650</span> * @param partitioner Partitioner to use. Pass <code>null</code> to use<a name="line.650"></a>
-<span class="sourceLineNo">651</span> * default partitioner.<a name="line.651"></a>
-<span class="sourceLineNo">652</span> * @param quorumAddress Distant cluster to write to; default is null for<a name="line.652"></a>
-<span class="sourceLineNo">653</span> * output to the cluster that is designated in <code>hbase-site.xml</code>.<a name="line.653"></a>
-<span class="sourceLineNo">654</span> * Set this String to the zookeeper ensemble of an alternate remote cluster<a name="line.654"></a>
-<span class="sourceLineNo">655</span> * when you would have the reduce write a cluster that is other than the<a name="line.655"></a>
-<span class="sourceLineNo">656</span> * default; e.g. copying tables between clusters, the source would be<a name="line.656"></a>
-<span class="sourceLineNo">657</span> * designated by <code>hbase-site.xml</code> and this param would have the<a name="line.657"></a>
-<span class="sourceLineNo">658</span> * ensemble address of the remote cluster. The format to pass is particular.<a name="line.658"></a>
-<span class="sourceLineNo">659</span> * Pass <code> &lt;hbase.zookeeper.quorum&gt;:&lt;<a name="line.659"></a>
-<span class="sourceLineNo">660</span> * hbase.zookeeper.client.port&gt;:&lt;zookeeper.znode.parent&gt;<a name="line.660"></a>
-<span class="sourceLineNo">661</span> * </code> such as <code>server,server2,server3:2181:/hbase</code>.<a name="line.661"></a>
-<span class="sourceLineNo">662</span> * @param serverClass redefined hbase.regionserver.class<a name="line.662"></a>
-<span class="sourceLineNo">663</span> * @param serverImpl redefined hbase.regionserver.impl<a name="line.663"></a>
-<span class="sourceLineNo">664</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.664"></a>
-<span class="sourceLineNo">665</span> * job classes via the distributed cache (tmpjars).<a name="line.665"></a>
-<span class="sourceLineNo">666</span> * @throws IOException When determining the region count fails.<a name="line.666"></a>
-<span class="sourceLineNo">667</span> */<a name="line.667"></a>
-<span class="sourceLineNo">668</span> public static void initTableReducerJob(String table,<a name="line.668"></a>
-<span class="sourceLineNo">669</span> Class<? extends TableReducer> reducer, Job job,<a name="line.669"></a>
-<span class="sourceLineNo">670</span> Class partitioner, String quorumAddress, String serverClass,<a name="line.670"></a>
-<span class="sourceLineNo">671</span> String serverImpl, boolean addDependencyJars) throws IOException {<a name="line.671"></a>
-<span class="sourceLineNo">672</span><a name="line.672"></a>
-<span class="sourceLineNo">673</span> Configuration conf = job.getConfiguration();<a name="line.673"></a>
-<span class="sourceLineNo">674</span> HBaseConfiguration.merge(conf, HBaseConfiguration.create(conf));<a name="line.674"></a>
-<span class="sourceLineNo">675</span> job.setOutputFormatClass(TableOutputFormat.class);<a name="line.675"></a>
-<span class="sourceLineNo">676</span> if (reducer != null) job.setReducerClass(reducer);<a name="line.676"></a>
-<span class="sourceLineNo">677</span> conf.set(TableOutputFormat.OUTPUT_TABLE, table);<a name="line.677"></a>
-<span class="sourceLineNo">678</span> conf.setStrings("io.serializations", conf.get("io.serializations"),<a name="line.678"></a>
-<span class="sourceLineNo">679</span> MutationSerialization.class.getName(), ResultSerialization.class.getName());<a name="line.679"></a>
-<span class="sourceLineNo">680</span> // If passed a quorum/ensemble address, pass it on to TableOutputFormat.<a name="line.680"></a>
-<span class="sourceLineNo">681</span> if (quorumAddress != null) {<a name="line.681"></a>
-<span class="sourceLineNo">682</span> // Calling this will validate the format<a name="line.682"></a>
-<span class="sourceLineNo">683</span> ZKUtil.transformClusterKey(quorumAddress);<a name="line.683"></a>
-<span class="sourceLineNo">684</span> conf.set(TableOutputFormat.QUORUM_ADDRESS,quorumAddress);<a name="line.684"></a>
-<span class="sourceLineNo">685</span> }<a name="line.685"></a>
-<span class="sourceLineNo">686</span> if (serverClass != null && serverImpl != null) {<a name="line.686"></a>
-<span class="sourceLineNo">687</span> conf.set(TableOutputFormat.REGION_SERVER_CLASS, serverClass);<a name="line.687"></a>
-<span class="sourceLineNo">688</span> conf.set(TableOutputFormat.REGION_SERVER_IMPL, serverImpl);<a name="line.688"></a>
-<span class="sourceLineNo">689</span> }<a name="line.689"></a>
-<span class="sourceLineNo">690</span> job.setOutputKeyClass(ImmutableBytesWritable.class);<a name="line.690"></a>
-<span class="sourceLineNo">691</span> job.setOutputValueClass(Writable.class);<a name="line.691"></a>
-<span class="sourceLineNo">692</span> if (partitioner == HRegionPartitioner.class) {<a name="line.692"></a>
-<span class="sourceLineNo">693</span> job.setPartitionerClass(HRegionPartitioner.class);<a name="line.693"></a>
-<span class="sourceLineNo">694</span> int regions = MetaTableAccessor.getRegionCount(conf, TableName.valueOf(table));<a name="line.694"></a>
-<span class="sourceLineNo">695</span> if (job.getNumReduceTasks() > regions) {<a name="line.695"></a>
-<span class="sourceLineNo">696</span> job.setNumReduceTasks(regions);<a name="line.696"></a>
-<span class="sourceLineNo">697</span> }<a name="line.697"></a>
-<span class="sourceLineNo">698</span> } else if (partitioner != null) {<a name="line.698"></a>
-<span class="sourceLineNo">699</span> job.setPartitionerClass(partitioner);<a name="line.699"></a>
-<span class="sourceLineNo">700</span> }<a name="line.700"></a>
-<span class="sourceLineNo">701</span><a name="line.701"></a>
-<span class="sourceLineNo">702</span> if (addDependencyJars) {<a name="line.702"></a>
-<span class="sourceLineNo">703</span> addDependencyJars(job);<a name="line.703"></a>
-<span class="sourceLineNo">704</span> }<a name="line.704"></a>
-<span class="sourceLineNo">705</span><a name="line.705"></a>
-<span class="sourceLineNo">706</span> initCredentials(job);<a name="line.706"></a>
-<span class="sourceLineNo">707</span> }<a name="line.707"></a>
-<span class="sourceLineNo">708</span><a name="line.708"></a>
-<span class="sourceLineNo">709</span> /**<a name="line.709"></a>
-<span class="sourceLineNo">710</span> * Ensures that the given number of reduce tasks for the given job<a name="line.710"></a>
-<span class="sourceLineNo">711</span> * configuration does not exceed the number of regions for the given table.<a name="line.711"></a>
-<span class="sourceLineNo">712</span> *<a name="line.712"></a>
-<span class="sourceLineNo">713</span> * @param table The table to get the region count for.<a name="line.713"></a>
-<span class="sourceLineNo">714</span> * @param job The current job to adjust.<a name="line.714"></a>
-<span class="sourceLineNo">715</span> * @throws IOException When retrieving the table details fails.<a name="line.715"></a>
-<span class="sourceLineNo">716</span> */<a name="line.716"></a>
-<span class="sourceLineNo">717</span> public static void limitNumReduceTasks(String table, Job job)<a name="line.717"></a>
-<span class="sourceLineNo">718</span> throws IOException {<a name="line.718"></a>
-<span class="sourceLineNo">719</span> int regions =<a name="line.719"></a>
-<span class="sourceLineNo">720</span> MetaTableAccessor.getRegionCount(job.getConfiguration(), TableName.valueOf(table));<a name="line.720"></a>
-<span class="sourceLineNo">721</span> if (job.getNumReduceTasks() > regions)<a name="line.721"></a>
-<span class="sourceLineNo">722</span> job.setNumReduceTasks(regions);<a name="line.722"></a>
-<span class="sourceLineNo">723</span> }<a name="line.723"></a>
-<span class="sourceLineNo">724</span><a name="line.724"></a>
-<span class="sourceLineNo">725</span> /**<a name="line.725"></a>
-<span class="sourceLineNo">726</span> * Sets the number of reduce tasks for the given job configuration to the<a name="line.726"></a>
-<span class="sourceLineNo">727</span> * number of regions the given table has.<a name="line.727"></a>
-<span class="sourceLineNo">728</span> *<a name="line.728"></a>
-<span class="sourceLineNo">729</span> * @param table The table to get the region count for.<a name="line.729"></a>
-<span class="sourceLineNo">730</span> * @param job The current job to adjust.<a name="line.730"></a>
-<span class="sourceLineNo">731</span> * @throws IOException When retrieving the table details fails.<a name="line.731"></a>
-<span class="sourceLineNo">732</span> */<a name="line.732"></a>
-<span class="sourceLineNo">733</span> public static void setNumReduceTasks(String table, Job job)<a name="line.733"></a>
-<span class="sourceLineNo">734</span> throws IOException {<a name="line.734"></a>
-<span class="sourceLineNo">735</span> job.setNumReduceTasks(MetaTableAccessor.getRegionCount(job.getConfiguration(),<a name="line.735"></a>
-<span class="sourceLineNo">736</span> TableName.valueOf(table)));<a name="line.736"></a>
-<span class="sourceLineNo">737</span> }<a name="line.737"></a>
-<span class="sourceLineNo">738</span><a name="line.738"></a>
-<span class="sourceLineNo">739</span> /**<a name="line.739"></a>
-<span class="sourceLineNo">740</span> * Sets the number of rows to return and cache with each scanner iteration.<a name="line.740"></a>
-<span class="sourceLineNo">741</span> * Higher caching values will enable faster mapreduce jobs at the expense of<a name="line.741"></a>
-<span class="sourceLineNo">742</span> * requiring more heap to contain the cached rows.<a name="line.742"></a>
-<span class="sourceLineNo">743</span> *<a name="line.743"></a>
-<span class="sourceLineNo">744</span> * @param job The current job to adjust.<a name="line.744"></a>
-<span class="sourceLineNo">745</span> * @param batchSize The number of rows to return in batch with each scanner<a name="line.745"></a>
-<span class="sourceLineNo">746</span> * iteration.<a name="line.746"></a>
-<span class="sourceLineNo">747</span> */<a name="line.747"></a>
-<span class="sourceLineNo">748</span> public static void setScannerCaching(Job job, int batchSize) {<a name="line.748"></a>
-<span class="sourceLineNo">749</span> job.getConfiguration().setInt("hbase.client.scanner.caching", batchSize);<a name="line.749"></a>
-<span class="sourceLineNo">750</span> }<a name="line.750"></a>
-<span class="sourceLineNo">751</span><a name="line.751"></a>
-<span class="sourceLineNo">752</span> /**<a name="line.752"></a>
-<span class="sourceLineNo">753</span> * Add HBase and its dependencies (only) to the job configuration.<a name="line.753"></a>
-<span class="sourceLineNo">754</span> * <p><a name="line.754"></a>
-<span class="sourceLineNo">755</span> * This is intended as a low-level API, facilitating code reuse between this<a name="line.755"></a>
-<span class="sourceLineNo">756</span> * class and its mapred counterpart. It also of use to extenral tools that<a name="line.756"></a>
-<span class="sourceLineNo">757</span> * need to build a MapReduce job that interacts with HBase but want<a name="line.757"></a>
-<span class="sourceLineNo">758</span> * fine-grained control over the jars shipped to the cluster.<a name="line.758"></a>
-<span class="sourceLineNo">759</span> * </p><a name="line.759"></a>
-<span class="sourceLineNo">760</span> * @param conf The Configuration object to extend with dependencies.<a name="line.760"></a>
-<span class="sourceLineNo">761</span> * @see org.apache.hadoop.hbase.mapred.TableMapReduceUtil<a name="line.761"></a>
-<span class="sourceLineNo">762</span> * @see <a href="https://issues.apache.org/jira/browse/PIG-3285">PIG-3285</a><a name="line.762"></a>
-<span class="sourceLineNo">763</span> */<a name="line.763"></a>
-<span class="sourceLineNo">764</span> public static void addHBaseDependencyJars(Configuration conf) throws IOException {<a name="line.764"></a>
-<span class="sourceLineNo">765</span> addDependencyJars(conf,<a name="line.765"></a>
-<span class="sourceLineNo">766</span> // explicitly pull a class from each module<a name="line.766"></a>
-<span class="sourceLineNo">767</span> org.apache.hadoop.hbase.HConstants.class, // hbase-common<a name="line.767"></a>
-<span class="sourceLineNo">768</span> org.apache.hadoop.hbase.protobuf.generated.ClientProtos.class, // hbase-protocol<a name="line.768"></a>
-<span class="sourceLineNo">769</span> org.apache.hadoop.hbase.client.Put.class, // hbase-client<a name="line.769"></a>
-<span class="sourceLineNo">770</span> org.apache.hadoop.hbase.CompatibilityFactory.class, // hbase-hadoop-compat<a name="line.770"></a>
-<span class="sourceLineNo">771</span> org.apache.hadoop.hbase.mapreduce.TableMapper.class, // hbase-server<a name="line.771"></a>
-<span class="sourceLineNo">772</span> // pull necessary dependencies<a name="line.772"></a>
-<span class="sourceLineNo">773</span> org.apache.zookeeper.ZooKeeper.class,<a name="line.773"></a>
-<span class="sourceLineNo">774</span> io.netty.channel.Channel.class,<a name="line.774"></a>
-<span class="sourceLineNo">775</span> com.google.protobuf.Message.class,<a name="line.775"></a>
-<span class="sourceLineNo">776</span> com.google.common.collect.Lists.class,<a name="line.776"></a>
-<span class="sourceLineNo">777</span> org.apache.htrace.Trace.class,<a name="line.777"></a>
-<span class="sourceLineNo">778</s
<TRUNCATED>