http://git-wip-us.apache.org/repos/asf/hbase-site/blob/d41f56fe/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 b8fea52..afbce13 100644 --- a/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.html +++ b/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/TableMapReduceUtil.html @@ -63,979 +63,1017 @@ <span class="sourceLineNo">055</span>import org.apache.hadoop.hbase.security.token.TokenUtil;<a name="line.55"></a> <span class="sourceLineNo">056</span>import org.apache.hadoop.hbase.util.Base64;<a name="line.56"></a> <span class="sourceLineNo">057</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.57"></a> -<span class="sourceLineNo">058</span>import org.apache.hadoop.hbase.zookeeper.ZKConfig;<a name="line.58"></a> -<span class="sourceLineNo">059</span>import org.apache.hadoop.io.Writable;<a name="line.59"></a> -<span class="sourceLineNo">060</span>import org.apache.hadoop.mapreduce.InputFormat;<a name="line.60"></a> -<span class="sourceLineNo">061</span>import org.apache.hadoop.mapreduce.Job;<a name="line.61"></a> -<span class="sourceLineNo">062</span>import org.apache.hadoop.util.StringUtils;<a name="line.62"></a> -<span class="sourceLineNo">063</span><a name="line.63"></a> -<span class="sourceLineNo">064</span>import com.codahale.metrics.MetricRegistry;<a name="line.64"></a> -<span class="sourceLineNo">065</span><a name="line.65"></a> -<span class="sourceLineNo">066</span>/**<a name="line.66"></a> -<span class="sourceLineNo">067</span> * Utility for {@link TableMapper} and {@link TableReducer}<a name="line.67"></a> -<span class="sourceLineNo">068</span> */<a name="line.68"></a> -<span class="sourceLineNo">069</span>@SuppressWarnings({ "rawtypes", "unchecked" })<a name="line.69"></a> -<span class="sourceLineNo">070</span>@InterfaceAudience.Public<a name="line.70"></a> -<span class="sourceLineNo">071</span>public class TableMapReduceUtil {<a name="line.71"></a> -<span class="sourceLineNo">072</span> private static final Log LOG = LogFactory.getLog(TableMapReduceUtil.class);<a name="line.72"></a> -<span class="sourceLineNo">073</span><a name="line.73"></a> -<span class="sourceLineNo">074</span> /**<a name="line.74"></a> -<span class="sourceLineNo">075</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.75"></a> -<span class="sourceLineNo">076</span> * the job.<a name="line.76"></a> -<span class="sourceLineNo">077</span> *<a name="line.77"></a> -<span class="sourceLineNo">078</span> * @param table The table name to read from.<a name="line.78"></a> -<span class="sourceLineNo">079</span> * @param scan The scan instance with the columns, time range etc.<a name="line.79"></a> -<span class="sourceLineNo">080</span> * @param mapper The mapper class to use.<a name="line.80"></a> -<span class="sourceLineNo">081</span> * @param outputKeyClass The class of the output key.<a name="line.81"></a> -<span class="sourceLineNo">082</span> * @param outputValueClass The class of the output value.<a name="line.82"></a> -<span class="sourceLineNo">083</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.83"></a> -<span class="sourceLineNo">084</span> * carrying all necessary HBase configuration.<a name="line.84"></a> -<span class="sourceLineNo">085</span> * @throws IOException When setting up the details fails.<a name="line.85"></a> -<span class="sourceLineNo">086</span> */<a name="line.86"></a> -<span class="sourceLineNo">087</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.87"></a> -<span class="sourceLineNo">088</span> Class<? extends TableMapper> mapper,<a name="line.88"></a> -<span class="sourceLineNo">089</span> Class<?> outputKeyClass,<a name="line.89"></a> -<span class="sourceLineNo">090</span> Class<?> outputValueClass, Job job)<a name="line.90"></a> -<span class="sourceLineNo">091</span> throws IOException {<a name="line.91"></a> -<span class="sourceLineNo">092</span> initTableMapperJob(table, scan, mapper, outputKeyClass, outputValueClass,<a name="line.92"></a> -<span class="sourceLineNo">093</span> job, true);<a name="line.93"></a> -<span class="sourceLineNo">094</span> }<a name="line.94"></a> -<span class="sourceLineNo">095</span><a name="line.95"></a> +<span class="sourceLineNo">058</span>import org.apache.hadoop.hbase.util.RegionSplitter;<a name="line.58"></a> +<span class="sourceLineNo">059</span>import org.apache.hadoop.hbase.zookeeper.ZKConfig;<a name="line.59"></a> +<span class="sourceLineNo">060</span>import org.apache.hadoop.io.Writable;<a name="line.60"></a> +<span class="sourceLineNo">061</span>import org.apache.hadoop.mapreduce.InputFormat;<a name="line.61"></a> +<span class="sourceLineNo">062</span>import org.apache.hadoop.mapreduce.Job;<a name="line.62"></a> +<span class="sourceLineNo">063</span>import org.apache.hadoop.util.StringUtils;<a name="line.63"></a> +<span class="sourceLineNo">064</span><a name="line.64"></a> +<span class="sourceLineNo">065</span>import com.codahale.metrics.MetricRegistry;<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>public class TableMapReduceUtil {<a name="line.72"></a> +<span class="sourceLineNo">073</span> private static final Log LOG = LogFactory.getLog(TableMapReduceUtil.class);<a name="line.73"></a> +<span class="sourceLineNo">074</span><a name="line.74"></a> +<span class="sourceLineNo">075</span> /**<a name="line.75"></a> +<span class="sourceLineNo">076</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.76"></a> +<span class="sourceLineNo">077</span> * the job.<a name="line.77"></a> +<span class="sourceLineNo">078</span> *<a name="line.78"></a> +<span class="sourceLineNo">079</span> * @param table The table name to read from.<a name="line.79"></a> +<span class="sourceLineNo">080</span> * @param scan The scan instance with the columns, time range etc.<a name="line.80"></a> +<span class="sourceLineNo">081</span> * @param mapper The mapper class to use.<a name="line.81"></a> +<span class="sourceLineNo">082</span> * @param outputKeyClass The class of the output key.<a name="line.82"></a> +<span class="sourceLineNo">083</span> * @param outputValueClass The class of the output value.<a name="line.83"></a> +<span class="sourceLineNo">084</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.84"></a> +<span class="sourceLineNo">085</span> * carrying all necessary HBase configuration.<a name="line.85"></a> +<span class="sourceLineNo">086</span> * @throws IOException When setting up the details fails.<a name="line.86"></a> +<span class="sourceLineNo">087</span> */<a name="line.87"></a> +<span class="sourceLineNo">088</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.88"></a> +<span class="sourceLineNo">089</span> Class<? extends TableMapper> mapper,<a name="line.89"></a> +<span class="sourceLineNo">090</span> Class<?> outputKeyClass,<a name="line.90"></a> +<span class="sourceLineNo">091</span> Class<?> outputValueClass, Job job)<a name="line.91"></a> +<span class="sourceLineNo">092</span> throws IOException {<a name="line.92"></a> +<span class="sourceLineNo">093</span> initTableMapperJob(table, scan, mapper, outputKeyClass, outputValueClass,<a name="line.93"></a> +<span class="sourceLineNo">094</span> job, true);<a name="line.94"></a> +<span class="sourceLineNo">095</span> }<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> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.98"></a> -<span class="sourceLineNo">099</span> * the job.<a name="line.99"></a> -<span class="sourceLineNo">100</span> *<a name="line.100"></a> -<span class="sourceLineNo">101</span> * @param table The table name to read from.<a name="line.101"></a> -<span class="sourceLineNo">102</span> * @param scan The scan instance with the columns, time range etc.<a name="line.102"></a> -<span class="sourceLineNo">103</span> * @param mapper The mapper class to use.<a name="line.103"></a> -<span class="sourceLineNo">104</span> * @param outputKeyClass The class of the output key.<a name="line.104"></a> -<span class="sourceLineNo">105</span> * @param outputValueClass The class of the output value.<a name="line.105"></a> -<span class="sourceLineNo">106</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.106"></a> -<span class="sourceLineNo">107</span> * carrying all necessary HBase configuration.<a name="line.107"></a> -<span class="sourceLineNo">108</span> * @throws IOException When setting up the details fails.<a name="line.108"></a> -<span class="sourceLineNo">109</span> */<a name="line.109"></a> -<span class="sourceLineNo">110</span> public static void initTableMapperJob(TableName table,<a name="line.110"></a> -<span class="sourceLineNo">111</span> Scan scan,<a name="line.111"></a> -<span class="sourceLineNo">112</span> Class<? extends TableMapper> mapper,<a name="line.112"></a> -<span class="sourceLineNo">113</span> Class<?> outputKeyClass,<a name="line.113"></a> -<span class="sourceLineNo">114</span> Class<?> outputValueClass,<a name="line.114"></a> -<span class="sourceLineNo">115</span> Job job) throws IOException {<a name="line.115"></a> -<span class="sourceLineNo">116</span> initTableMapperJob(table.getNameAsString(),<a name="line.116"></a> -<span class="sourceLineNo">117</span> scan,<a name="line.117"></a> -<span class="sourceLineNo">118</span> mapper,<a name="line.118"></a> -<span class="sourceLineNo">119</span> outputKeyClass,<a name="line.119"></a> -<span class="sourceLineNo">120</span> outputValueClass,<a name="line.120"></a> -<span class="sourceLineNo">121</span> job,<a name="line.121"></a> -<span class="sourceLineNo">122</span> true);<a name="line.122"></a> -<span class="sourceLineNo">123</span> }<a name="line.123"></a> -<span class="sourceLineNo">124</span><a name="line.124"></a> -<span class="sourceLineNo">125</span> /**<a name="line.125"></a> -<span class="sourceLineNo">126</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.126"></a> -<span class="sourceLineNo">127</span> * the job.<a name="line.127"></a> -<span class="sourceLineNo">128</span> *<a name="line.128"></a> -<span class="sourceLineNo">129</span> * @param table Binary representation of the table name to read from.<a name="line.129"></a> -<span class="sourceLineNo">130</span> * @param scan The scan instance with the columns, time range etc.<a name="line.130"></a> -<span class="sourceLineNo">131</span> * @param mapper The mapper class to use.<a name="line.131"></a> -<span class="sourceLineNo">132</span> * @param outputKeyClass The class of the output key.<a name="line.132"></a> -<span class="sourceLineNo">133</span> * @param outputValueClass The class of the output value.<a name="line.133"></a> -<span class="sourceLineNo">134</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.134"></a> -<span class="sourceLineNo">135</span> * carrying all necessary HBase configuration.<a name="line.135"></a> -<span class="sourceLineNo">136</span> * @throws IOException When setting up the details fails.<a name="line.136"></a> -<span class="sourceLineNo">137</span> */<a name="line.137"></a> -<span class="sourceLineNo">138</span> public static void initTableMapperJob(byte[] table, Scan scan,<a name="line.138"></a> -<span class="sourceLineNo">139</span> Class<? extends TableMapper> mapper,<a name="line.139"></a> -<span class="sourceLineNo">140</span> Class<?> outputKeyClass,<a name="line.140"></a> -<span class="sourceLineNo">141</span> Class<?> outputValueClass, Job job)<a name="line.141"></a> -<span class="sourceLineNo">142</span> throws IOException {<a name="line.142"></a> -<span class="sourceLineNo">143</span> initTableMapperJob(Bytes.toString(table), scan, mapper, outputKeyClass, outputValueClass,<a name="line.143"></a> -<span class="sourceLineNo">144</span> job, true);<a name="line.144"></a> -<span class="sourceLineNo">145</span> }<a name="line.145"></a> -<span class="sourceLineNo">146</span><a name="line.146"></a> -<span class="sourceLineNo">147</span> /**<a name="line.147"></a> -<span class="sourceLineNo">148</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.148"></a> -<span class="sourceLineNo">149</span> * the job.<a name="line.149"></a> -<span class="sourceLineNo">150</span> *<a name="line.150"></a> -<span class="sourceLineNo">151</span> * @param table The table name to read from.<a name="line.151"></a> -<span class="sourceLineNo">152</span> * @param scan The scan instance with the columns, time range etc.<a name="line.152"></a> -<span class="sourceLineNo">153</span> * @param mapper The mapper class to use.<a name="line.153"></a> -<span class="sourceLineNo">154</span> * @param outputKeyClass The class of the output key.<a name="line.154"></a> -<span class="sourceLineNo">155</span> * @param outputValueClass The class of the output value.<a name="line.155"></a> -<span class="sourceLineNo">156</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.156"></a> -<span class="sourceLineNo">157</span> * carrying all necessary HBase configuration.<a name="line.157"></a> -<span class="sourceLineNo">158</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.158"></a> -<span class="sourceLineNo">159</span> * job classes via the distributed cache (tmpjars).<a name="line.159"></a> -<span class="sourceLineNo">160</span> * @throws IOException When setting up the details fails.<a name="line.160"></a> -<span class="sourceLineNo">161</span> */<a name="line.161"></a> -<span class="sourceLineNo">162</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.162"></a> -<span class="sourceLineNo">163</span> Class<? extends TableMapper> mapper,<a name="line.163"></a> -<span class="sourceLineNo">164</span> Class<?> outputKeyClass,<a name="line.164"></a> -<span class="sourceLineNo">165</span> Class<?> outputValueClass, Job job,<a name="line.165"></a> -<span class="sourceLineNo">166</span> boolean addDependencyJars, Class<? extends InputFormat> inputFormatClass)<a name="line.166"></a> -<span class="sourceLineNo">167</span> throws IOException {<a name="line.167"></a> -<span class="sourceLineNo">168</span> initTableMapperJob(table, scan, mapper, outputKeyClass, outputValueClass, job,<a name="line.168"></a> -<span class="sourceLineNo">169</span> addDependencyJars, true, inputFormatClass);<a name="line.169"></a> -<span class="sourceLineNo">170</span> }<a name="line.170"></a> -<span class="sourceLineNo">171</span><a name="line.171"></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> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.99"></a> +<span class="sourceLineNo">100</span> * the job.<a name="line.100"></a> +<span class="sourceLineNo">101</span> *<a name="line.101"></a> +<span class="sourceLineNo">102</span> * @param table The table name to read from.<a name="line.102"></a> +<span class="sourceLineNo">103</span> * @param scan The scan instance with the columns, time range etc.<a name="line.103"></a> +<span class="sourceLineNo">104</span> * @param mapper The mapper class to use.<a name="line.104"></a> +<span class="sourceLineNo">105</span> * @param outputKeyClass The class of the output key.<a name="line.105"></a> +<span class="sourceLineNo">106</span> * @param outputValueClass The class of the output value.<a name="line.106"></a> +<span class="sourceLineNo">107</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.107"></a> +<span class="sourceLineNo">108</span> * carrying all necessary HBase configuration.<a name="line.108"></a> +<span class="sourceLineNo">109</span> * @throws IOException When setting up the details fails.<a name="line.109"></a> +<span class="sourceLineNo">110</span> */<a name="line.110"></a> +<span class="sourceLineNo">111</span> public static void initTableMapperJob(TableName table,<a name="line.111"></a> +<span class="sourceLineNo">112</span> Scan scan,<a name="line.112"></a> +<span class="sourceLineNo">113</span> Class<? extends TableMapper> mapper,<a name="line.113"></a> +<span class="sourceLineNo">114</span> Class<?> outputKeyClass,<a name="line.114"></a> +<span class="sourceLineNo">115</span> Class<?> outputValueClass,<a name="line.115"></a> +<span class="sourceLineNo">116</span> Job job) throws IOException {<a name="line.116"></a> +<span class="sourceLineNo">117</span> initTableMapperJob(table.getNameAsString(),<a name="line.117"></a> +<span class="sourceLineNo">118</span> scan,<a name="line.118"></a> +<span class="sourceLineNo">119</span> mapper,<a name="line.119"></a> +<span class="sourceLineNo">120</span> outputKeyClass,<a name="line.120"></a> +<span class="sourceLineNo">121</span> outputValueClass,<a name="line.121"></a> +<span class="sourceLineNo">122</span> job,<a name="line.122"></a> +<span class="sourceLineNo">123</span> true);<a name="line.123"></a> +<span class="sourceLineNo">124</span> }<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> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.127"></a> +<span class="sourceLineNo">128</span> * the job.<a name="line.128"></a> +<span class="sourceLineNo">129</span> *<a name="line.129"></a> +<span class="sourceLineNo">130</span> * @param table Binary representation of the table name to read from.<a name="line.130"></a> +<span class="sourceLineNo">131</span> * @param scan The scan instance with the columns, time range etc.<a name="line.131"></a> +<span class="sourceLineNo">132</span> * @param mapper The mapper class to use.<a name="line.132"></a> +<span class="sourceLineNo">133</span> * @param outputKeyClass The class of the output key.<a name="line.133"></a> +<span class="sourceLineNo">134</span> * @param outputValueClass The class of the output value.<a name="line.134"></a> +<span class="sourceLineNo">135</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.135"></a> +<span class="sourceLineNo">136</span> * carrying all necessary HBase configuration.<a name="line.136"></a> +<span class="sourceLineNo">137</span> * @throws IOException When setting up the details fails.<a name="line.137"></a> +<span class="sourceLineNo">138</span> */<a name="line.138"></a> +<span class="sourceLineNo">139</span> public static void initTableMapperJob(byte[] table, Scan scan,<a name="line.139"></a> +<span class="sourceLineNo">140</span> Class<? extends TableMapper> mapper,<a name="line.140"></a> +<span class="sourceLineNo">141</span> Class<?> outputKeyClass,<a name="line.141"></a> +<span class="sourceLineNo">142</span> Class<?> outputValueClass, Job job)<a name="line.142"></a> +<span class="sourceLineNo">143</span> throws IOException {<a name="line.143"></a> +<span class="sourceLineNo">144</span> initTableMapperJob(Bytes.toString(table), scan, mapper, outputKeyClass, outputValueClass,<a name="line.144"></a> +<span class="sourceLineNo">145</span> job, true);<a name="line.145"></a> +<span class="sourceLineNo">146</span> }<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> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.149"></a> +<span class="sourceLineNo">150</span> * the job.<a name="line.150"></a> +<span class="sourceLineNo">151</span> *<a name="line.151"></a> +<span class="sourceLineNo">152</span> * @param table The table name to read from.<a name="line.152"></a> +<span class="sourceLineNo">153</span> * @param scan The scan instance with the columns, time range etc.<a name="line.153"></a> +<span class="sourceLineNo">154</span> * @param mapper The mapper class to use.<a name="line.154"></a> +<span class="sourceLineNo">155</span> * @param outputKeyClass The class of the output key.<a name="line.155"></a> +<span class="sourceLineNo">156</span> * @param outputValueClass The class of the output value.<a name="line.156"></a> +<span class="sourceLineNo">157</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.157"></a> +<span class="sourceLineNo">158</span> * carrying all necessary HBase configuration.<a name="line.158"></a> +<span class="sourceLineNo">159</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.159"></a> +<span class="sourceLineNo">160</span> * job classes via the distributed cache (tmpjars).<a name="line.160"></a> +<span class="sourceLineNo">161</span> * @throws IOException When setting up the details fails.<a name="line.161"></a> +<span class="sourceLineNo">162</span> */<a name="line.162"></a> +<span class="sourceLineNo">163</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.163"></a> +<span class="sourceLineNo">164</span> Class<? extends TableMapper> mapper,<a name="line.164"></a> +<span class="sourceLineNo">165</span> Class<?> outputKeyClass,<a name="line.165"></a> +<span class="sourceLineNo">166</span> Class<?> outputValueClass, Job job,<a name="line.166"></a> +<span class="sourceLineNo">167</span> boolean addDependencyJars, Class<? extends InputFormat> inputFormatClass)<a name="line.167"></a> +<span class="sourceLineNo">168</span> throws IOException {<a name="line.168"></a> +<span class="sourceLineNo">169</span> initTableMapperJob(table, scan, mapper, outputKeyClass, outputValueClass, job,<a name="line.169"></a> +<span class="sourceLineNo">170</span> addDependencyJars, true, inputFormatClass);<a name="line.170"></a> +<span class="sourceLineNo">171</span> }<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> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.174"></a> -<span class="sourceLineNo">175</span> * the job.<a name="line.175"></a> -<span class="sourceLineNo">176</span> *<a name="line.176"></a> -<span class="sourceLineNo">177</span> * @param table The table name to read from.<a name="line.177"></a> -<span class="sourceLineNo">178</span> * @param scan The scan instance with the columns, time range etc.<a name="line.178"></a> -<span class="sourceLineNo">179</span> * @param mapper The mapper class to use.<a name="line.179"></a> -<span class="sourceLineNo">180</span> * @param outputKeyClass The class of the output key.<a name="line.180"></a> -<span class="sourceLineNo">181</span> * @param outputValueClass The class of the output value.<a name="line.181"></a> -<span class="sourceLineNo">182</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.182"></a> -<span class="sourceLineNo">183</span> * carrying all necessary HBase configuration.<a name="line.183"></a> -<span class="sourceLineNo">184</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.184"></a> -<span class="sourceLineNo">185</span> * job classes via the distributed cache (tmpjars).<a name="line.185"></a> -<span class="sourceLineNo">186</span> * @param initCredentials whether to initialize hbase auth credentials for the job<a name="line.186"></a> -<span class="sourceLineNo">187</span> * @param inputFormatClass the input format<a name="line.187"></a> -<span class="sourceLineNo">188</span> * @throws IOException When setting up the details fails.<a name="line.188"></a> -<span class="sourceLineNo">189</span> */<a name="line.189"></a> -<span class="sourceLineNo">190</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.190"></a> -<span class="sourceLineNo">191</span> Class<? extends TableMapper> mapper,<a name="line.191"></a> -<span class="sourceLineNo">192</span> Class<?> outputKeyClass,<a name="line.192"></a> -<span class="sourceLineNo">193</span> Class<?> outputValueClass, Job job,<a name="line.193"></a> -<span class="sourceLineNo">194</span> boolean addDependencyJars, boolean initCredentials,<a name="line.194"></a> -<span class="sourceLineNo">195</span> Class<? extends InputFormat> inputFormatClass)<a name="line.195"></a> -<span class="sourceLineNo">196</span> throws IOException {<a name="line.196"></a> -<span class="sourceLineNo">197</span> job.setInputFormatClass(inputFormatClass);<a name="line.197"></a> -<span class="sourceLineNo">198</span> if (outputValueClass != null) job.setMapOutputValueClass(outputValueClass);<a name="line.198"></a> -<span class="sourceLineNo">199</span> if (outputKeyClass != null) job.setMapOutputKeyClass(outputKeyClass);<a name="line.199"></a> -<span class="sourceLineNo">200</span> job.setMapperClass(mapper);<a name="line.200"></a> -<span class="sourceLineNo">201</span> if (Put.class.equals(outputValueClass)) {<a name="line.201"></a> -<span class="sourceLineNo">202</span> job.setCombinerClass(PutCombiner.class);<a name="line.202"></a> -<span class="sourceLineNo">203</span> }<a name="line.203"></a> -<span class="sourceLineNo">204</span> Configuration conf = job.getConfiguration();<a name="line.204"></a> -<span class="sourceLineNo">205</span> HBaseConfiguration.merge(conf, HBaseConfiguration.create(conf));<a name="line.205"></a> -<span class="sourceLineNo">206</span> conf.set(TableInputFormat.INPUT_TABLE, table);<a name="line.206"></a> -<span class="sourceLineNo">207</span> conf.set(TableInputFormat.SCAN, convertScanToString(scan));<a name="line.207"></a> -<span class="sourceLineNo">208</span> conf.setStrings("io.serializations", conf.get("io.serializations"),<a name="line.208"></a> -<span class="sourceLineNo">209</span> MutationSerialization.class.getName(), ResultSerialization.class.getName(),<a name="line.209"></a> -<span class="sourceLineNo">210</span> KeyValueSerialization.class.getName());<a name="line.210"></a> -<span class="sourceLineNo">211</span> if (addDependencyJars) {<a name="line.211"></a> -<span class="sourceLineNo">212</span> addDependencyJars(job);<a name="line.212"></a> -<span class="sourceLineNo">213</span> }<a name="line.213"></a> -<span class="sourceLineNo">214</span> if (initCredentials) {<a name="line.214"></a> -<span class="sourceLineNo">215</span> initCredentials(job);<a name="line.215"></a> -<span class="sourceLineNo">216</span> }<a name="line.216"></a> -<span class="sourceLineNo">217</span> }<a name="line.217"></a> -<span class="sourceLineNo">218</span><a name="line.218"></a> -<span class="sourceLineNo">219</span> /**<a name="line.219"></a> -<span class="sourceLineNo">220</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.220"></a> -<span class="sourceLineNo">221</span> * the job.<a name="line.221"></a> -<span class="sourceLineNo">222</span> *<a name="line.222"></a> -<span class="sourceLineNo">223</span> * @param table Binary representation of the table name to read from.<a name="line.223"></a> -<span class="sourceLineNo">224</span> * @param scan The scan instance with the columns, time range etc.<a name="line.224"></a> -<span class="sourceLineNo">225</span> * @param mapper The mapper class to use.<a name="line.225"></a> -<span class="sourceLineNo">226</span> * @param outputKeyClass The class of the output key.<a name="line.226"></a> -<span class="sourceLineNo">227</span> * @param outputValueClass The class of the output value.<a name="line.227"></a> -<span class="sourceLineNo">228</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.228"></a> -<span class="sourceLineNo">229</span> * carrying all necessary HBase configuration.<a name="line.229"></a> -<span class="sourceLineNo">230</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.230"></a> -<span class="sourceLineNo">231</span> * job classes via the distributed cache (tmpjars).<a name="line.231"></a> -<span class="sourceLineNo">232</span> * @param inputFormatClass The class of the input format<a name="line.232"></a> -<span class="sourceLineNo">233</span> * @throws IOException When setting up the details fails.<a name="line.233"></a> -<span class="sourceLineNo">234</span> */<a name="line.234"></a> -<span class="sourceLineNo">235</span> public static void initTableMapperJob(byte[] table, Scan scan,<a name="line.235"></a> -<span class="sourceLineNo">236</span> Class<? extends TableMapper> mapper,<a name="line.236"></a> -<span class="sourceLineNo">237</span> Class<?> outputKeyClass,<a name="line.237"></a> -<span class="sourceLineNo">238</span> Class<?> outputValueClass, Job job,<a name="line.238"></a> -<span class="sourceLineNo">239</span> boolean addDependencyJars, Class<? extends InputFormat> inputFormatClass)<a name="line.239"></a> -<span class="sourceLineNo">240</span> throws IOException {<a name="line.240"></a> -<span class="sourceLineNo">241</span> initTableMapperJob(Bytes.toString(table), scan, mapper, outputKeyClass,<a name="line.241"></a> -<span class="sourceLineNo">242</span> outputValueClass, job, addDependencyJars, inputFormatClass);<a name="line.242"></a> -<span class="sourceLineNo">243</span> }<a name="line.243"></a> -<span class="sourceLineNo">244</span><a name="line.244"></a> -<span class="sourceLineNo">245</span> /**<a name="line.245"></a> -<span class="sourceLineNo">246</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.246"></a> -<span class="sourceLineNo">247</span> * the job.<a name="line.247"></a> -<span class="sourceLineNo">248</span> *<a name="line.248"></a> -<span class="sourceLineNo">249</span> * @param table Binary representation of the table name to read from.<a name="line.249"></a> -<span class="sourceLineNo">250</span> * @param scan The scan instance with the columns, time range etc.<a name="line.250"></a> -<span class="sourceLineNo">251</span> * @param mapper The mapper class to use.<a name="line.251"></a> -<span class="sourceLineNo">252</span> * @param outputKeyClass The class of the output key.<a name="line.252"></a> -<span class="sourceLineNo">253</span> * @param outputValueClass The class of the output value.<a name="line.253"></a> -<span class="sourceLineNo">254</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.254"></a> -<span class="sourceLineNo">255</span> * carrying all necessary HBase configuration.<a name="line.255"></a> -<span class="sourceLineNo">256</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.256"></a> -<span class="sourceLineNo">257</span> * job classes via the distributed cache (tmpjars).<a name="line.257"></a> -<span class="sourceLineNo">258</span> * @throws IOException When setting up the details fails.<a name="line.258"></a> -<span class="sourceLineNo">259</span> */<a name="line.259"></a> -<span class="sourceLineNo">260</span> public static void initTableMapperJob(byte[] table, Scan scan,<a name="line.260"></a> -<span class="sourceLineNo">261</span> Class<? extends TableMapper> mapper,<a name="line.261"></a> -<span class="sourceLineNo">262</span> Class<?> outputKeyClass,<a name="line.262"></a> -<span class="sourceLineNo">263</span> Class<?> outputValueClass, Job job,<a name="line.263"></a> -<span class="sourceLineNo">264</span> boolean addDependencyJars)<a name="line.264"></a> -<span class="sourceLineNo">265</span> throws IOException {<a name="line.265"></a> -<span class="sourceLineNo">266</span> initTableMapperJob(Bytes.toString(table), scan, mapper, outputKeyClass,<a name="line.266"></a> -<span class="sourceLineNo">267</span> outputValueClass, job, addDependencyJars, TableInputFormat.class);<a name="line.267"></a> -<span class="sourceLineNo">268</span> }<a name="line.268"></a> -<span class="sourceLineNo">269</span><a name="line.269"></a> -<span class="sourceLineNo">270</span> /**<a name="line.270"></a> -<span class="sourceLineNo">271</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.271"></a> -<span class="sourceLineNo">272</span> * the job.<a name="line.272"></a> -<span class="sourceLineNo">273</span> *<a name="line.273"></a> -<span class="sourceLineNo">274</span> * @param table The table name to read from.<a name="line.274"></a> -<span class="sourceLineNo">275</span> * @param scan The scan instance with the columns, time range etc.<a name="line.275"></a> -<span class="sourceLineNo">276</span> * @param mapper The mapper class to use.<a name="line.276"></a> -<span class="sourceLineNo">277</span> * @param outputKeyClass The class of the output key.<a name="line.277"></a> -<span class="sourceLineNo">278</span> * @param outputValueClass The class of the output value.<a name="line.278"></a> -<span class="sourceLineNo">279</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.279"></a> -<span class="sourceLineNo">280</span> * carrying all necessary HBase configuration.<a name="line.280"></a> -<span class="sourceLineNo">281</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.281"></a> -<span class="sourceLineNo">282</span> * job classes via the distributed cache (tmpjars).<a name="line.282"></a> -<span class="sourceLineNo">283</span> * @throws IOException When setting up the details fails.<a name="line.283"></a> -<span class="sourceLineNo">284</span> */<a name="line.284"></a> -<span class="sourceLineNo">285</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.285"></a> -<span class="sourceLineNo">286</span> Class<? extends TableMapper> mapper,<a name="line.286"></a> -<span class="sourceLineNo">287</span> Class<?> outputKeyClass,<a name="line.287"></a> -<span class="sourceLineNo">288</span> Class<?> outputValueClass, Job job,<a name="line.288"></a> -<span class="sourceLineNo">289</span> boolean addDependencyJars)<a name="line.289"></a> -<span class="sourceLineNo">290</span> throws IOException {<a name="line.290"></a> -<span class="sourceLineNo">291</span> initTableMapperJob(table, scan, mapper, outputKeyClass,<a name="line.291"></a> -<span class="sourceLineNo">292</span> outputValueClass, job, addDependencyJars, TableInputFormat.class);<a name="line.292"></a> -<span class="sourceLineNo">293</span> }<a name="line.293"></a> -<span class="sourceLineNo">294</span><a name="line.294"></a> -<span class="sourceLineNo">295</span> /**<a name="line.295"></a> -<span class="sourceLineNo">296</span> * Enable a basic on-heap cache for these jobs. Any BlockCache implementation based on<a name="line.296"></a> -<span class="sourceLineNo">297</span> * direct memory will likely cause the map tasks to OOM when opening the region. This<a name="line.297"></a> -<span class="sourceLineNo">298</span> * is done here instead of in TableSnapshotRegionRecordReader in case an advanced user<a name="line.298"></a> -<span class="sourceLineNo">299</span> * wants to override this behavior in their job.<a name="line.299"></a> -<span class="sourceLineNo">300</span> */<a name="line.300"></a> -<span class="sourceLineNo">301</span> public static void resetCacheConfig(Configuration conf) {<a name="line.301"></a> -<span class="sourceLineNo">302</span> conf.setFloat(<a name="line.302"></a> -<span class="sourceLineNo">303</span> HConstants.HFILE_BLOCK_CACHE_SIZE_KEY, HConstants.HFILE_BLOCK_CACHE_SIZE_DEFAULT);<a name="line.303"></a> -<span class="sourceLineNo">304</span> conf.setFloat(HConstants.BUCKET_CACHE_SIZE_KEY, 0f);<a name="line.304"></a> -<span class="sourceLineNo">305</span> conf.unset(HConstants.BUCKET_CACHE_IOENGINE_KEY);<a name="line.305"></a> -<span class="sourceLineNo">306</span> }<a name="line.306"></a> -<span class="sourceLineNo">307</span><a name="line.307"></a> -<span class="sourceLineNo">308</span> /**<a name="line.308"></a> -<span class="sourceLineNo">309</span> * Sets up the job for reading from one or more table snapshots, with one or more scans<a name="line.309"></a> -<span class="sourceLineNo">310</span> * per snapshot.<a name="line.310"></a> -<span class="sourceLineNo">311</span> * It bypasses hbase servers and read directly from snapshot files.<a name="line.311"></a> -<span class="sourceLineNo">312</span> *<a name="line.312"></a> -<span class="sourceLineNo">313</span> * @param snapshotScans map of snapshot name to scans on that snapshot.<a name="line.313"></a> -<span class="sourceLineNo">314</span> * @param mapper The mapper class to use.<a name="line.314"></a> -<span class="sourceLineNo">315</span> * @param outputKeyClass The class of the output key.<a name="line.315"></a> -<span class="sourceLineNo">316</span> * @param outputValueClass The class of the output value.<a name="line.316"></a> -<span class="sourceLineNo">317</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.317"></a> -<span class="sourceLineNo">318</span> * carrying all necessary HBase configuration.<a name="line.318"></a> -<span class="sourceLineNo">319</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.319"></a> -<span class="sourceLineNo">320</span> * job classes via the distributed cache (tmpjars).<a name="line.320"></a> -<span class="sourceLineNo">321</span> */<a name="line.321"></a> -<span class="sourceLineNo">322</span> public static void initMultiTableSnapshotMapperJob(Map<String, Collection<Scan>> snapshotScans,<a name="line.322"></a> -<span class="sourceLineNo">323</span> Class<? extends TableMapper> mapper, Class<?> outputKeyClass, Class<?> outputValueClass,<a name="line.323"></a> -<span class="sourceLineNo">324</span> Job job, boolean addDependencyJars, Path tmpRestoreDir) throws IOException {<a name="line.324"></a> -<span class="sourceLineNo">325</span> MultiTableSnapshotInputFormat.setInput(job.getConfiguration(), snapshotScans, tmpRestoreDir);<a name="line.325"></a> -<span class="sourceLineNo">326</span><a name="line.326"></a> -<span class="sourceLineNo">327</span> job.setInputFormatClass(MultiTableSnapshotInputFormat.class);<a name="line.327"></a> -<span class="sourceLineNo">328</span> if (outputValueClass != null) {<a name="line.328"></a> -<span class="sourceLineNo">329</span> job.setMapOutputValueClass(outputValueClass);<a name="line.329"></a> -<span class="sourceLineNo">330</span> }<a name="line.330"></a> -<span class="sourceLineNo">331</span> if (outputKeyClass != null) {<a name="line.331"></a> -<span class="sourceLineNo">332</span> job.setMapOutputKeyClass(outputKeyClass);<a name="line.332"></a> -<span class="sourceLineNo">333</span> }<a name="line.333"></a> -<span class="sourceLineNo">334</span> job.setMapperClass(mapper);<a name="line.334"></a> -<span class="sourceLineNo">335</span> Configuration conf = job.getConfiguration();<a name="line.335"></a> -<span class="sourceLineNo">336</span> HBaseConfiguration.merge(conf, HBaseConfiguration.create(conf));<a name="line.336"></a> -<span class="sourceLineNo">337</span><a name="line.337"></a> -<span class="sourceLineNo">338</span> if (addDependencyJars) {<a name="line.338"></a> -<span class="sourceLineNo">339</span> addDependencyJars(job);<a name="line.339"></a> -<span class="sourceLineNo">340</span> addDependencyJarsForClasses(job.getConfiguration(), MetricRegistry.class);<a name="line.340"></a> -<span class="sourceLineNo">341</span> }<a name="line.341"></a> -<span class="sourceLineNo">342</span><a name="line.342"></a> -<span class="sourceLineNo">343</span> resetCacheConfig(job.getConfiguration());<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> /**<a name="line.346"></a> -<span class="sourceLineNo">347</span> * Sets up the job for reading from a table snapshot. It bypasses hbase servers<a name="line.347"></a> -<span class="sourceLineNo">348</span> * and read directly from snapshot files.<a name="line.348"></a> -<span class="sourceLineNo">349</span> *<a name="line.349"></a> -<span class="sourceLineNo">350</span> * @param snapshotName The name of the snapshot (of a table) to read from.<a name="line.350"></a> -<span class="sourceLineNo">351</span> * @param scan The scan instance with the columns, time range etc.<a name="line.351"></a> -<span class="sourceLineNo">352</span> * @param mapper The mapper class to use.<a name="line.352"></a> -<span class="sourceLineNo">353</span> * @param outputKeyClass The class of the output key.<a name="line.353"></a> -<span class="sourceLineNo">354</span> * @param outputValueClass The class of the output value.<a name="line.354"></a> -<span class="sourceLineNo">355</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.355"></a> -<span class="sourceLineNo">356</span> * carrying all necessary HBase configuration.<a name="line.356"></a> -<span class="sourceLineNo">357</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.357"></a> -<span class="sourceLineNo">358</span> * job classes via the distributed cache (tmpjars).<a name="line.358"></a> -<span class="sourceLineNo">359</span> *<a name="line.359"></a> -<span class="sourceLineNo">360</span> * @param tmpRestoreDir a temporary directory to copy the snapshot files into. Current user should<a name="line.360"></a> -<span class="sourceLineNo">361</span> * have write permissions to this directory, and this should not be a subdirectory of rootdir.<a name="line.361"></a> -<span class="sourceLineNo">362</span> * After the job is finished, restore directory can be deleted.<a name="line.362"></a> -<span class="sourceLineNo">363</span> * @throws IOException When setting up the details fails.<a name="line.363"></a> -<span class="sourceLineNo">364</span> * @see TableSnapshotInputFormat<a name="line.364"></a> -<span class="sourceLineNo">365</span> */<a name="line.365"></a> -<span class="sourceLineNo">366</span> public static void initTableSnapshotMapperJob(String snapshotName, Scan scan,<a name="line.366"></a> -<span class="sourceLineNo">367</span> Class<? extends TableMapper> mapper,<a name="line.367"></a> -<span class="sourceLineNo">368</span> Class<?> outputKeyClass,<a name="line.368"></a> -<span class="sourceLineNo">369</span> Class<?> outputValueClass, Job job,<a name="line.369"></a> -<span class="sourceLineNo">370</span> boolean addDependencyJars, Path tmpRestoreDir)<a name="line.370"></a> -<span class="sourceLineNo">371</span> throws IOException {<a name="line.371"></a> -<span class="sourceLineNo">372</span> TableSnapshotInputFormat.setInput(job, snapshotName, tmpRestoreDir);<a name="line.372"></a> -<span class="sourceLineNo">373</span> initTableMapperJob(snapshotName, scan, mapper, outputKeyClass,<a name="line.373"></a> -<span class="sourceLineNo">374</span> outputValueClass, job, addDependencyJars, false, TableSnapshotInputFormat.class);<a name="line.374"></a> -<span class="sourceLineNo">375</span> resetCacheConfig(job.getConfiguration());<a name="line.375"></a> -<span class="sourceLineNo">376</span> }<a name="line.376"></a> -<span class="sourceLineNo">377</span><a name="line.377"></a> -<span class="sourceLineNo">378</span> /**<a name="line.378"></a> -<span class="sourceLineNo">379</span> * Use this before submitting a Multi TableMap job. It will appropriately set<a name="line.379"></a> -<span class="sourceLineNo">380</span> * up the job.<a name="line.380"></a> -<span class="sourceLineNo">381</span> *<a name="line.381"></a> -<span class="sourceLineNo">382</span> * @param scans The list of {@link Scan} objects to read from.<a name="line.382"></a> -<span class="sourceLineNo">383</span> * @param mapper The mapper class to use.<a name="line.383"></a> -<span class="sourceLineNo">384</span> * @param outputKeyClass The class of the output key.<a name="line.384"></a> -<span class="sourceLineNo">385</span> * @param outputValueClass The class of the output value.<a name="line.385"></a> -<span class="sourceLineNo">386</span> * @param job The current job to adjust. Make sure the passed job is carrying<a name="line.386"></a> -<span class="sourceLineNo">387</span> * all necessary HBase configuration.<a name="line.387"></a> -<span class="sourceLineNo">388</span> * @throws IOException When setting up the details fails.<a name="line.388"></a> -<span class="sourceLineNo">389</span> */<a name="line.389"></a> -<span class="sourceLineNo">390</span> public static void initTableMapperJob(List<Scan> scans,<a name="line.390"></a> -<span class="sourceLineNo">391</span> Class<? extends TableMapper> mapper,<a name="line.391"></a> -<span class="sourceLineNo">392</span> Class<?> outputKeyClass,<a name="line.392"></a> -<span class="sourceLineNo">393</span> Class<?> outputValueClass, Job job) throws IOException {<a name="line.393"></a> -<span class="sourceLineNo">394</span> initTableMapperJob(scans, mapper, outputKeyClass, outputValueClass, job,<a name="line.394"></a> -<span class="sourceLineNo">395</span> true);<a name="line.395"></a> -<span class="sourceLineNo">396</span> }<a name="line.396"></a> -<span class="sourceLineNo">397</span><a name="line.397"></a> -<span class="sourceLineNo">398</span> /**<a name="line.398"></a> -<span class="sourceLineNo">399</span> * Use this before submitting a Multi TableMap job. It will appropriately set<a name="line.399"></a> -<span class="sourceLineNo">400</span> * up the job.<a name="line.400"></a> -<span class="sourceLineNo">401</span> *<a name="line.401"></a> -<span class="sourceLineNo">402</span> * @param scans The list of {@link Scan} objects to read from.<a name="line.402"></a> -<span class="sourceLineNo">403</span> * @param mapper The mapper class to use.<a name="line.403"></a> -<span class="sourceLineNo">404</span> * @param outputKeyClass The class of the output key.<a name="line.404"></a> -<span class="sourceLineNo">405</span> * @param outputValueClass The class of the output value.<a name="line.405"></a> -<span class="sourceLineNo">406</span> * @param job The current job to adjust. Make sure the passed job is carrying<a name="line.406"></a> -<span class="sourceLineNo">407</span> * all necessary HBase configuration.<a name="line.407"></a> -<span class="sourceLineNo">408</span> * @param addDependencyJars upload HBase jars and jars for any of the<a name="line.408"></a> -<span class="sourceLineNo">409</span> * configured job classes via the distributed cache (tmpjars).<a name="line.409"></a> -<span class="sourceLineNo">410</span> * @throws IOException When setting up the details fails.<a name="line.410"></a> -<span class="sourceLineNo">411</span> */<a name="line.411"></a> -<span class="sourceLineNo">412</span> public static void initTableMapperJob(List<Scan> scans,<a name="line.412"></a> -<span class="sourceLineNo">413</span> Class<? extends TableMapper> mapper,<a name="line.413"></a> -<span class="sourceLineNo">414</span> Class<?> outputKeyClass,<a name="line.414"></a> -<span class="sourceLineNo">415</span> Class<?> outputValueClass, Job job,<a name="line.415"></a> -<span class="sourceLineNo">416</span> boolean addDependencyJars) throws IOException {<a name="line.416"></a> -<span class="sourceLineNo">417</span> initTableMapperJob(scans, mapper, outputKeyClass, outputValueClass, job,<a name="line.417"></a> -<span class="sourceLineNo">418</span> addDependencyJars, true);<a name="line.418"></a> -<span class="sourceLineNo">419</span> }<a name="line.419"></a> -<span class="sourceLineNo">420</span><a name="line.420"></a> -<span class="sourceLineNo">421</span> /**<a name="line.421"></a> -<span class="sourceLineNo">422</span> * Use this before submitting a Multi TableMap job. It will appropriately set<a name="line.422"></a> -<span class="sourceLineNo">423</span> * up the job.<a name="line.423"></a> -<span class="sourceLineNo">424</span> *<a name="line.424"></a> -<span class="sourceLineNo">425</span> * @param scans The list of {@link Scan} objects to read from.<a name="line.425"></a> -<span class="sourceLineNo">426</span> * @param mapper The mapper class to use.<a name="line.426"></a> -<span class="sourceLineNo">427</span> * @param outputKeyClass The class of the output key.<a name="line.427"></a> -<span class="sourceLineNo">428</span> * @param outputValueClass The class of the output value.<a name="line.428"></a> -<span class="sourceLineNo">429</span> * @param job The current job to adjust. Make sure the passed job is carrying<a name="line.429"></a> -<span class="sourceLineNo">430</span> * all necessary HBase configuration.<a name="line.430"></a> -<span class="sourceLineNo">431</span> * @param addDependencyJars upload HBase jars and jars for any of the<a name="line.431"></a> -<span class="sourceLineNo">432</span> * configured job classes via the distributed cache (tmpjars).<a name="line.432"></a> -<span class="sourceLineNo">433</span> * @param initCredentials whether to initialize hbase auth credentials for the job<a name="line.433"></a> -<span class="sourceLineNo">434</span> * @throws IOException When setting up the details fails.<a name="line.434"></a> -<span class="sourceLineNo">435</span> */<a name="line.435"></a> -<span class="sourceLineNo">436</span> public static void initTableMapperJob(List<Scan> scans,<a name="line.436"></a> -<span class="sourceLineNo">437</span> Class<? extends TableMapper> mapper,<a name="line.437"></a> -<span class="sourceLineNo">438</span> Class<?> outputKeyClass,<a name="line.438"></a> -<span class="sourceLineNo">439</span> Class<?> outputValueClass, Job job,<a name="line.439"></a> -<span class="sourceLineNo">440</span> boolean addDependencyJars,<a name="line.440"></a> -<span class="sourceLineNo">441</span> boolean initCredentials) throws IOException {<a name="line.441"></a> -<span class="sourceLineNo">442</span> job.setInputFormatClass(MultiTableInputFormat.class);<a name="line.442"></a> -<span class="sourceLineNo">443</span> if (outputValueClass != null) {<a name="line.443"></a> -<span class="sourceLineNo">444</span> job.setMapOutputValueClass(outputValueClass);<a name="line.444"></a> -<span class="sourceLineNo">445</span> }<a name="line.445"></a> -<span class="sourceLineNo">446</span> if (outputKeyClass != null) {<a name="line.446"></a> -<span class="sourceLineNo">447</span> job.setMapOutputKeyClass(outputKeyClass);<a name="line.447"></a> -<span class="sourceLineNo">448</span> }<a name="line.448"></a> -<span class="sourceLineNo">449</span> job.setMapperClass(mapper);<a name="line.449"></a> -<span class="sourceLineNo">450</span> Configuration conf = job.getConfiguration();<a name="line.450"></a> -<span class="sourceLineNo">451</span> HBaseConfiguration.merge(conf, HBaseConfiguration.create(conf));<a name="line.451"></a> -<span class="sourceLineNo">452</span> List<String> scanStrings = new ArrayList<>();<a name="line.452"></a> -<span class="sourceLineNo">453</span><a name="line.453"></a> -<span class="sourceLineNo">454</span> for (Scan scan : scans) {<a name="line.454"></a> -<span class="sourceLineNo">455</span> scanStrings.add(convertScanToString(scan));<a name="line.455"></a> -<span class="sourceLineNo">456</span> }<a name="line.456"></a> -<span class="sourceLineNo">457</span> job.getConfiguration().setStrings(MultiTableInputFormat.SCANS,<a name="line.457"></a> -<span class="sourceLineNo">458</span> scanStrings.toArray(new String[scanStrings.size()]));<a name="line.458"></a> -<span class="sourceLineNo">459</span><a name="line.459"></a> -<span class="sourceLineNo">460</span> if (addDependencyJars) {<a name="line.460"></a> -<span class="sourceLineNo">461</span> addDependencyJars(job);<a name="line.461"></a> -<span class="sourceLineNo">462</span> }<a name="line.462"></a> -<span class="sourceLineNo">463</span><a name="line.463"></a> -<span class="sourceLineNo">464</span> if (initCredentials) {<a name="line.464"></a> -<span class="sourceLineNo">465</span> initCredentials(job);<a name="line.465"></a> -<span class="sourceLineNo">466</span> }<a name="line.466"></a> -<span class="sourceLineNo">467</span> }<a name="line.467"></a> -<span class="sourceLineNo">468</span><a name="line.468"></a> -<span class="sourceLineNo">469</span> public static void initCredentials(Job job) throws IOException {<a name="line.469"></a> -<span class="sourceLineNo">470</span> UserProvider userProvider = UserProvider.instantiate(job.getConfiguration());<a name="line.470"></a> -<span class="sourceLineNo">471</span> if (userProvider.isHadoopSecurityEnabled()) {<a name="line.471"></a> -<span class="sourceLineNo">472</span> // propagate delegation related props from launcher job to MR job<a name="line.472"></a> -<span class="sourceLineNo">473</span> if (System.getenv("HADOOP_TOKEN_FILE_LOCATION") != null) {<a name="line.473"></a> -<span class="sourceLineNo">474</span> job.getConfiguration().set("mapreduce.job.credentials.binary",<a name="line.474"></a> -<span class="sourceLineNo">475</span> System.getenv("HADOOP_TOKEN_FILE_LOCATION"));<a name="line.475"></a> -<span class="sourceLineNo">476</span> }<a name="line.476"></a> -<span class="sourceLineNo">477</span> }<a name="line.477"></a> -<span class="sourceLineNo">478</span><a name="line.478"></a> -<span class="sourceLineNo">479</span> if (userProvider.isHBaseSecurityEnabled()) {<a name="line.479"></a> -<span class="sourceLineNo">480</span> try {<a name="line.480"></a> -<span class="sourceLineNo">481</span> // init credentials for remote cluster<a name="line.481"></a> -<span class="sourceLineNo">482</span> String quorumAddress = job.getConfiguration().get(TableOutputFormat.QUORUM_ADDRESS);<a name="line.482"></a> -<span class="sourceLineNo">483</span> User user = userProvider.getCurrent();<a name="line.483"></a> -<span class="sourceLineNo">484</span> if (quorumAddress != null) {<a name="line.484"></a> -<span class="sourceLineNo">485</span> Configuration peerConf = HBaseConfiguration.createClusterConf(job.getConfiguration(),<a name="line.485"></a> -<span class="sourceLineNo">486</span> quorumAddress, TableOutputFormat.OUTPUT_CONF_PREFIX);<a name="line.486"></a> -<span class="sourceLineNo">487</span> Connection peerConn = ConnectionFactory.createConnection(peerConf);<a name="line.487"></a> -<span class="sourceLineNo">488</span> try {<a name="line.488"></a> -<span class="sourceLineNo">489</span> TokenUtil.addTokenForJob(peerConn, user, job);<a name="line.489"></a> -<span class="sourceLineNo">490</span> } finally {<a name="line.490"></a> -<span class="sourceLineNo">491</span> peerConn.close();<a name="line.491"></a> -<span class="sourceLineNo">492</span> }<a name="line.492"></a> -<span class="sourceLineNo">493</span> }<a name="line.493"></a> -<span class="sourceLineNo">494</span><a name="line.494"></a> -<span class="sourceLineNo">495</span> Connection conn = ConnectionFactory.createConnection(job.getConfiguration());<a name="line.495"></a> -<span class="sourceLineNo">496</span> try {<a name="line.496"></a> -<span class="sourceLineNo">497</span> TokenUtil.addTokenForJob(conn, user, job);<a name="line.497"></a> -<span class="sourceLineNo">498</span> } finally {<a name="line.498"></a> -<span class="sourceLineNo">499</span> conn.close();<a name="line.499"></a> -<span class="sourceLineNo">500</span> }<a name="line.500"></a> -<span class="sourceLineNo">501</span> } catch (InterruptedException ie) {<a name="line.501"></a> -<span class="sourceLineNo">502</span> LOG.info("Interrupted obtaining user authentication token");<a name="line.502"></a> -<span class="sourceLineNo">503</span> Thread.currentThread().interrupt();<a name="line.503"></a> -<span class="sourceLineNo">504</span> }<a name="line.504"></a> -<span class="sourceLineNo">505</span> }<a name="line.505"></a> -<span class="sourceLineNo">506</span> }<a name="line.506"></a> -<span class="sourceLineNo">507</span><a name="line.507"></a> -<span class="sourceLineNo">508</span> /**<a name="line.508"></a> -<span class="sourceLineNo">509</span> * Obtain an authentication token, for the specified cluster, on behalf of the current user<a name="line.509"></a> -<span class="sourceLineNo">510</span> * and add it to the credentials for the given map reduce job.<a name="line.510"></a> -<span class="sourceLineNo">511</span> *<a name="line.511"></a> -<span class="sourceLineNo">512</span> * The quorumAddress is the key to the ZK ensemble, which contains:<a name="line.512"></a> -<span class="sourceLineNo">513</span> * hbase.zookeeper.quorum, hbase.zookeeper.client.port and<a name="line.513"></a> -<span class="sourceLineNo">514</span> * zookeeper.znode.parent<a name="line.514"></a> -<span class="sourceLineNo">515</span> *<a name="line.515"></a> -<span class="sourceLineNo">516</span> * @param job The job that requires the permission.<a name="line.516"></a> -<span class="sourceLineNo">517</span> * @param quorumAddress string that contains the 3 required configuratins<a name="line.517"></a> -<span class="sourceLineNo">518</span> * @throws IOException When the authentication token cannot be obtained.<a name="line.518"></a> -<span class="sourceLineNo">519</span> * @deprecated Since 1.2.0, use {@link #initCredentialsForCluster(Job, Configuration)} instead.<a name="line.519"></a> -<span class="sourceLineNo">520</span> */<a name="line.520"></a> -<span class="sourceLineNo">521</span> @Deprecated<a name="line.521"></a> -<span class="sourceLineNo">522</span> public static void initCredentialsForCluster(Job job, String quorumAddress)<a name="line.522"></a> -<span class="sourceLineNo">523</span> throws IOException {<a name="line.523"></a> -<span class="sourceLineNo">524</span> Configuration peerConf = HBaseConfiguration.createClusterConf(job.getConfiguration(),<a name="line.524"></a> -<span class="sourceLineNo">525</span> quorumAddress);<a name="line.525"></a> -<span class="sourceLineNo">526</span> initCredentialsForCluster(job, peerConf);<a name="line.526"></a> -<span class="sourceLineNo">527</span> }<a name="line.527"></a> -<span class="sourceLineNo">528</span><a name="line.528"></a> -<span class="sourceLineNo">529</span> /**<a name="line.529"></a> -<span class="sourceLineNo">530</span> * Obtain an authentication token, for the specified cluster, on behalf of the current user<a name="line.530"></a> -<span class="sourceLineNo">531</span> * and add it to the credentials for the given map reduce job.<a name="line.531"></a> -<span class="sourceLineNo">532</span> *<a name="line.532"></a> -<span class="sourceLineNo">533</span> * @param job The job that requires the permission.<a name="line.533"></a> -<span class="sourceLineNo">534</span> * @param conf The configuration to use in connecting to the peer cluster<a name="line.534"></a> -<span class="sourceLineNo">535</span> * @throws IOException When the authentication token cannot be obtained.<a name="line.535"></a> -<span class="sourceLineNo">536</span> */<a name="line.536"></a> -<span class="sourceLineNo">537</span> public static void initCredentialsForCluster(Job job, Configuration conf)<a name="line.537"></a> -<span class="sourceLineNo">538</span> throws IOException {<a name="line.538"></a> -<span class="sourceLineNo">539</span> UserProvider userProvider = UserProvider.instantiate(job.getConfiguration());<a name="line.539"></a> -<span class="sourceLineNo">540</span> if (userProvider.isHBaseSecurityEnabled()) {<a name="line.540"></a> -<span class="sourceLineNo">541</span> try {<a name="line.541"></a> -<span class="sourceLineNo">542</span> Connection peerConn = ConnectionFactory.createConnection(conf);<a name="line.542"></a> -<span class="sourceLineNo">543</span> try {<a name="line.543"></a> -<span class="sourceLineNo">544</span> TokenUtil.addTokenForJob(peerConn, userProvider.getCurrent(), job);<a name="line.544"></a> -<span class="sourceLineNo">545</span> } finally {<a name="line.545"></a> -<span class="sourceLineNo">546</span> peerConn.close();<a name="line.546"></a> -<span class="sourceLineNo">547</span> }<a name="line.547"></a> -<span class="sourceLineNo">548</span> } catch (InterruptedException e) {<a name="line.548"></a> -<span class="sourceLineNo">549</span> LOG.info("Interrupted obtaining user authentication token");<a name="line.549"></a> -<span class="sourceLineNo">550</span> Thread.interrupted();<a name="line.550"></a> -<span class="sourceLineNo">551</span> }<a name="line.551"></a> -<span class="sourceLineNo">552</span> }<a name="line.552"></a> -<span class="sourceLineNo">553</span> }<a name="line.553"></a> -<span class="sourceLineNo">554</span><a name="line.554"></a> -<span class="sourceLineNo">555</span> /**<a name="line.555"></a> -<span class="sourceLineNo">556</span> * Writes the given scan into a Base64 encoded string.<a name="line.556"></a> -<span class="sourceLineNo">557</span> *<a name="line.557"></a> -<span class="sourceLineNo">558</span> * @param scan The scan to write out.<a name="line.558"></a> -<span class="sourceLineNo">559</span> * @return The scan saved in a Base64 encoded string.<a name="line.559"></a> -<span class="sourceLineNo">560</span> * @throws IOException When writing the scan fails.<a name="line.560"></a> -<span class="sourceLineNo">561</span> */<a name="line.561"></a> -<span class="sourceLineNo">562</span> public static String convertScanToString(Scan scan) throws IOException {<a name="line.562"></a> -<span class="sourceLineNo">563</span> ClientProtos.Scan proto = ProtobufUtil.toScan(scan);<a name="line.563"></a> -<span class="sourceLineNo">564</span> return Base64.encodeBytes(proto.toByteArray());<a name="line.564"></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> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.175"></a> +<span class="sourceLineNo">176</span> * the job.<a name="line.176"></a> +<span class="sourceLineNo">177</span> *<a name="line.177"></a> +<span class="sourceLineNo">178</span> * @param table The table name to read from.<a name="line.178"></a> +<span class="sourceLineNo">179</span> * @param scan The scan instance with the columns, time range etc.<a name="line.179"></a> +<span class="sourceLineNo">180</span> * @param mapper The mapper class to use.<a name="line.180"></a> +<span class="sourceLineNo">181</span> * @param outputKeyClass The class of the output key.<a name="line.181"></a> +<span class="sourceLineNo">182</span> * @param outputValueClass The class of the output value.<a name="line.182"></a> +<span class="sourceLineNo">183</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.183"></a> +<span class="sourceLineNo">184</span> * carrying all necessary HBase configuration.<a name="line.184"></a> +<span class="sourceLineNo">185</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.185"></a> +<span class="sourceLineNo">186</span> * job classes via the distributed cache (tmpjars).<a name="line.186"></a> +<span class="sourceLineNo">187</span> * @param initCredentials whether to initialize hbase auth credentials for the job<a name="line.187"></a> +<span class="sourceLineNo">188</span> * @param inputFormatClass the input format<a name="line.188"></a> +<span class="sourceLineNo">189</span> * @throws IOException When setting up the details fails.<a name="line.189"></a> +<span class="sourceLineNo">190</span> */<a name="line.190"></a> +<span class="sourceLineNo">191</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.191"></a> +<span class="sourceLineNo">192</span> Class<? extends TableMapper> mapper,<a name="line.192"></a> +<span class="sourceLineNo">193</span> Class<?> outputKeyClass,<a name="line.193"></a> +<span class="sourceLineNo">194</span> Class<?> outputValueClass, Job job,<a name="line.194"></a> +<span class="sourceLineNo">195</span> boolean addDependencyJars, boolean initCredentials,<a name="line.195"></a> +<span class="sourceLineNo">196</span> Class<? extends InputFormat> inputFormatClass)<a name="line.196"></a> +<span class="sourceLineNo">197</span> throws IOException {<a name="line.197"></a> +<span class="sourceLineNo">198</span> job.setInputFormatClass(inputFormatClass);<a name="line.198"></a> +<span class="sourceLineNo">199</span> if (outputValueClass != null) job.setMapOutputValueClass(outputValueClass);<a name="line.199"></a> +<span class="sourceLineNo">200</span> if (outputKeyClass != null) job.setMapOutputKeyClass(outputKeyClass);<a name="line.200"></a> +<span class="sourceLineNo">201</span> job.setMapperClass(mapper);<a name="line.201"></a> +<span class="sourceLineNo">202</span> if (Put.class.equals(outputValueClass)) {<a name="line.202"></a> +<span class="sourceLineNo">203</span> job.setCombinerClass(PutCombiner.class);<a name="line.203"></a> +<span class="sourceLineNo">204</span> }<a name="line.204"></a> +<span class="sourceLineNo">205</span> Configuration conf = job.getConfiguration();<a name="line.205"></a> +<span class="sourceLineNo">206</span> HBaseConfiguration.merge(conf, HBaseConfiguration.create(conf));<a name="line.206"></a> +<span class="sourceLineNo">207</span> conf.set(TableInputFormat.INPUT_TABLE, table);<a name="line.207"></a> +<span class="sourceLineNo">208</span> conf.set(TableInputFormat.SCAN, convertScanToString(scan));<a name="line.208"></a> +<span class="sourceLineNo">209</span> conf.setStrings("io.serializations", conf.get("io.serializations"),<a name="line.209"></a> +<span class="sourceLineNo">210</span> MutationSerialization.class.getName(), ResultSerialization.class.getName(),<a name="line.210"></a> +<span class="sourceLineNo">211</span> KeyValueSerialization.class.getName());<a name="line.211"></a> +<span class="sourceLineNo">212</span> if (addDependencyJars) {<a name="line.212"></a> +<span class="sourceLineNo">213</span> addDependencyJars(job);<a name="line.213"></a> +<span class="sourceLineNo">214</span> }<a name="line.214"></a> +<span class="sourceLineNo">215</span> if (initCredentials) {<a name="line.215"></a> +<span class="sourceLineNo">216</span> initCredentials(job);<a name="line.216"></a> +<span class="sourceLineNo">217</span> }<a name="line.217"></a> +<span class="sourceLineNo">218</span> }<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> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.221"></a> +<span class="sourceLineNo">222</span> * the job.<a name="line.222"></a> +<span class="sourceLineNo">223</span> *<a name="line.223"></a> +<span class="sourceLineNo">224</span> * @param table Binary representation of the table name to read from.<a name="line.224"></a> +<span class="sourceLineNo">225</span> * @param scan The scan instance with the columns, time range etc.<a name="line.225"></a> +<span class="sourceLineNo">226</span> * @param mapper The mapper class to use.<a name="line.226"></a> +<span class="sourceLineNo">227</span> * @param outputKeyClass The class of the output key.<a name="line.227"></a> +<span class="sourceLineNo">228</span> * @param outputValueClass The class of the output value.<a name="line.228"></a> +<span class="sourceLineNo">229</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.229"></a> +<span class="sourceLineNo">230</span> * carrying all necessary HBase configuration.<a name="line.230"></a> +<span class="sourceLineNo">231</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.231"></a> +<span class="sourceLineNo">232</span> * job classes via the distributed cache (tmpjars).<a name="line.232"></a> +<span class="sourceLineNo">233</span> * @param inputFormatClass The class of the input format<a name="line.233"></a> +<span class="sourceLineNo">234</span> * @throws IOException When setting up the details fails.<a name="line.234"></a> +<span class="sourceLineNo">235</span> */<a name="line.235"></a> +<span class="sourceLineNo">236</span> public static void initTableMapperJob(byte[] table, Scan scan,<a name="line.236"></a> +<span class="sourceLineNo">237</span> Class<? extends TableMapper> mapper,<a name="line.237"></a> +<span class="sourceLineNo">238</span> Class<?> outputKeyClass,<a name="line.238"></a> +<span class="sourceLineNo">239</span> Class<?> outputValueClass, Job job,<a name="line.239"></a> +<span class="sourceLineNo">240</span> boolean addDependencyJars, Class<? extends InputFormat> inputFormatClass)<a name="line.240"></a> +<span class="sourceLineNo">241</span> throws IOException {<a name="line.241"></a> +<span class="sourceLineNo">242</span> initTableMapperJob(Bytes.toString(table), scan, mapper, outputKeyClass,<a name="line.242"></a> +<span class="sourceLineNo">243</span> outputValueClass, job, addDependencyJars, inputFormatClass);<a name="line.243"></a> +<span class="sourceLineNo">244</span> }<a name="line.244"></a> +<span class="sourceLineNo">245</span><a name="line.245"></a> +<span class="sourceLineNo">246</span> /**<a name="line.246"></a> +<span class="sourceLineNo">247</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.247"></a> +<span class="sourceLineNo">248</span> * the job.<a name="line.248"></a> +<span class="sourceLineNo">249</span> *<a name="line.249"></a> +<span class="sourceLineNo">250</span> * @param table Binary representation of the table name to read from.<a name="line.250"></a> +<span class="sourceLineNo">251</span> * @param scan The scan instance with the columns, time range etc.<a name="line.251"></a> +<span class="sourceLineNo">252</span> * @param mapper The mapper class to use.<a name="line.252"></a> +<span class="sourceLineNo">253</span> * @param outputKeyClass The class of the output key.<a name="line.253"></a> +<span class="sourceLineNo">254</span> * @param outputValueClass The class of the output value.<a name="line.254"></a> +<span class="sourceLineNo">255</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.255"></a> +<span class="sourceLineNo">256</span> * carrying all necessary HBase configuration.<a name="line.256"></a> +<span class="sourceLineNo">257</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.257"></a> +<span class="sourceLineNo">258</span> * job classes via the distributed cache (tmpjars).<a name="line.258"></a> +<span class="sourceLineNo">259</span> * @throws IOException When setting up the details fails.<a name="line.259"></a> +<span class="sourceLineNo">260</span> */<a name="line.260"></a> +<span class="sourceLineNo">261</span> public static void initTableMapperJob(byte[] table, Scan scan,<a name="line.261"></a> +<span class="sourceLineNo">262</span> Class<? extends TableMapper> mapper,<a name="line.262"></a> +<span class="sourceLineNo">263</span> Class<?> outputKeyClass,<a name="line.263"></a> +<span class="sourceLineNo">264</span> Class<?> outputValueClass, Job job,<a name="line.264"></a> +<span class="sourceLineNo">265</span> boolean addDependencyJars)<a name="line.265"></a> +<span class="sourceLineNo">266</span> throws IOException {<a name="line.266"></a> +<span class="sourceLineNo">267</span> initTableMapperJob(Bytes.toString(table), scan, mapper, outputKeyClass,<a name="line.267"></a> +<span class="sourceLineNo">268</span> outputValueClass, job, addDependencyJars, TableInputFormat.class);<a name="line.268"></a> +<span class="sourceLineNo">269</span> }<a name="line.269"></a> +<span class="sourceLineNo">270</span><a name="line.270"></a> +<span class="sourceLineNo">271</span> /**<a name="line.271"></a> +<span class="sourceLineNo">272</span> * Use this before submitting a TableMap job. It will appropriately set up<a name="line.272"></a> +<span class="sourceLineNo">273</span> * the job.<a name="line.273"></a> +<span class="sourceLineNo">274</span> *<a name="line.274"></a> +<span class="sourceLineNo">275</span> * @param table The table name to read from.<a name="line.275"></a> +<span class="sourceLineNo">276</span> * @param scan The scan instance with the columns, time range etc.<a name="line.276"></a> +<span class="sourceLineNo">277</span> * @param mapper The mapper class to use.<a name="line.277"></a> +<span class="sourceLineNo">278</span> * @param outputKeyClass The class of the output key.<a name="line.278"></a> +<span class="sourceLineNo">279</span> * @param outputValueClass The class of the output value.<a name="line.279"></a> +<span class="sourceLineNo">280</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.280"></a> +<span class="sourceLineNo">281</span> * carrying all necessary HBase configuration.<a name="line.281"></a> +<span class="sourceLineNo">282</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.282"></a> +<span class="sourceLineNo">283</span> * job classes via the distributed cache (tmpjars).<a name="line.283"></a> +<span class="sourceLineNo">284</span> * @throws IOException When setting up the details fails.<a name="line.284"></a> +<span class="sourceLineNo">285</span> */<a name="line.285"></a> +<span class="sourceLineNo">286</span> public static void initTableMapperJob(String table, Scan scan,<a name="line.286"></a> +<span class="sourceLineNo">287</span> Class<? extends TableMapper> mapper,<a name="line.287"></a> +<span class="sourceLineNo">288</span> Class<?> outputKeyClass,<a name="line.288"></a> +<span class="sourceLineNo">289</span> Class<?> outputValueClass, Job job,<a name="line.289"></a> +<span class="sourceLineNo">290</span> boolean addDependencyJars)<a name="line.290"></a> +<span class="sourceLineNo">291</span> throws IOException {<a name="line.291"></a> +<span class="sourceLineNo">292</span> initTableMapperJob(table, scan, mapper, outputKeyClass,<a name="line.292"></a> +<span class="sourceLineNo">293</span> outputValueClass, job, addDependencyJars, TableInputFormat.class);<a name="line.293"></a> +<span class="sourceLineNo">294</span> }<a name="line.294"></a> +<span class="sourceLineNo">295</span><a name="line.295"></a> +<span class="sourceLineNo">296</span> /**<a name="line.296"></a> +<span class="sourceLineNo">297</span> * Enable a basic on-heap cache for these jobs. Any BlockCache implementation based on<a name="line.297"></a> +<span class="sourceLineNo">298</span> * direct memory will likely cause the map tasks to OOM when opening the region. This<a name="line.298"></a> +<span class="sourceLineNo">299</span> * is done here instead of in TableSnapshotRegionRecordReader in case an advanced user<a name="line.299"></a> +<span class="sourceLineNo">300</span> * wants to override this behavior in their job.<a name="line.300"></a> +<span class="sourceLineNo">301</span> */<a name="line.301"></a> +<span class="sourceLineNo">302</span> public static void resetCacheConfig(Configuration conf) {<a name="line.302"></a> +<span class="sourceLineNo">303</span> conf.setFloat(<a name="line.303"></a> +<span class="sourceLineNo">304</span> HConstants.HFILE_BLOCK_CACHE_SIZE_KEY, HConstants.HFILE_BLOCK_CACHE_SIZE_DEFAULT);<a name="line.304"></a> +<span class="sourceLineNo">305</span> conf.setFloat(HConstants.BUCKET_CACHE_SIZE_KEY, 0f);<a name="line.305"></a> +<span class="sourceLineNo">306</span> conf.unset(HConstants.BUCKET_CACHE_IOENGINE_KEY);<a name="line.306"></a> +<span class="sourceLineNo">307</span> }<a name="line.307"></a> +<span class="sourceLineNo">308</span><a name="line.308"></a> +<span class="sourceLineNo">309</span> /**<a name="line.309"></a> +<span class="sourceLineNo">310</span> * Sets up the job for reading from one or more table snapshots, with one or more scans<a name="line.310"></a> +<span class="sourceLineNo">311</span> * per snapshot.<a name="line.311"></a> +<span class="sourceLineNo">312</span> * It bypasses hbase servers and read directly from snapshot files.<a name="line.312"></a> +<span class="sourceLineNo">313</span> *<a name="line.313"></a> +<span class="sourceLineNo">314</span> * @param snapshotScans map of snapshot name to scans on that snapshot.<a name="line.314"></a> +<span class="sourceLineNo">315</span> * @param mapper The mapper class to use.<a name="line.315"></a> +<span class="sourceLineNo">316</span> * @param outputKeyClass The class of the output key.<a name="line.316"></a> +<span class="sourceLineNo">317</span> * @param outputValueClass The class of the output value.<a name="line.317"></a> +<span class="sourceLineNo">318</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.318"></a> +<span class="sourceLineNo">319</span> * carrying all necessary HBase configuration.<a name="line.319"></a> +<span class="sourceLineNo">320</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.320"></a> +<span class="sourceLineNo">321</span> * job classes via the distributed cache (tmpjars).<a name="line.321"></a> +<span class="sourceLineNo">322</span> */<a name="line.322"></a> +<span class="sourceLineNo">323</span> public static void initMultiTableSnapshotMapperJob(Map<String, Collection<Scan>> snapshotScans,<a name="line.323"></a> +<span class="sourceLineNo">324</span> Class<? extends TableMapper> mapper, Class<?> outputKeyClass, Class<?> outputValueClass,<a name="line.324"></a> +<span class="sourceLineNo">325</span> Job job, boolean addDependencyJars, Path tmpRestoreDir) throws IOException {<a name="line.325"></a> +<span class="sourceLineNo">326</span> MultiTableSnapshotInputFormat.setInput(job.getConfiguration(), snapshotScans, tmpRestoreDir);<a name="line.326"></a> +<span class="sourceLineNo">327</span><a name="line.327"></a> +<span class="sourceLineNo">328</span> job.setInputFormatClass(MultiTableSnapshotInputFormat.class);<a name="line.328"></a> +<span class="sourceLineNo">329</span> if (outputValueClass != null) {<a name="line.329"></a> +<span class="sourceLineNo">330</span> job.setMapOutputValueClass(outputValueClass);<a name="line.330"></a> +<span class="sourceLineNo">331</span> }<a name="line.331"></a> +<span class="sourceLineNo">332</span> if (outputKeyClass != null) {<a name="line.332"></a> +<span class="sourceLineNo">333</span> job.setMapOutputKeyClass(outputKeyClass);<a name="line.333"></a> +<span class="sourceLineNo">334</span> }<a name="line.334"></a> +<span class="sourceLineNo">335</span> job.setMapperClass(mapper);<a name="line.335"></a> +<span class="sourceLineNo">336</span> Configuration conf = job.getConfiguration();<a name="line.336"></a> +<span class="sourceLineNo">337</span> HBaseConfiguration.merge(conf, HBaseConfiguration.create(conf));<a name="line.337"></a> +<span class="sourceLineNo">338</span><a name="line.338"></a> +<span class="sourceLineNo">339</span> if (addDependencyJars) {<a name="line.339"></a> +<span class="sourceLineNo">340</span> addDependencyJars(job);<a name="line.340"></a> +<span class="sourceLineNo">341</span> addDependencyJarsForClasses(job.getConfiguration(), MetricRegistry.class);<a name="line.341"></a> +<span class="sourceLineNo">342</span> }<a name="line.342"></a> +<span class="sourceLineNo">343</span><a name="line.343"></a> +<span class="sourceLineNo">344</span> resetCacheConfig(job.getConfiguration());<a name="line.344"></a> +<span class="sourceLineNo">345</span> }<a name="line.345"></a> +<span class="sourceLineNo">346</span><a name="line.346"></a> +<span class="sourceLineNo">347</span> /**<a name="line.347"></a> +<span class="sourceLineNo">348</span> * Sets up the job for reading from a table snapshot. It bypasses hbase servers<a name="line.348"></a> +<span class="sourceLineNo">349</span> * and read directly from snapshot files.<a name="line.349"></a> +<span class="sourceLineNo">350</span> *<a name="line.350"></a> +<span class="sourceLineNo">351</span> * @param snapshotName The name of the snapshot (of a table) to read from.<a name="line.351"></a> +<span class="sourceLineNo">352</span> * @param scan The scan instance with the columns, time range etc.<a name="line.352"></a> +<span class="sourceLineNo">353</span> * @param mapper The mapper class to use.<a name="line.353"></a> +<span class="sourceLineNo">354</span> * @param outputKeyClass The class of the output key.<a name="line.354"></a> +<span class="sourceLineNo">355</span> * @param outputValueClass The class of the output value.<a name="line.355"></a> +<span class="sourceLineNo">356</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.356"></a> +<span class="sourceLineNo">357</span> * carrying all necessary HBase configuration.<a name="line.357"></a> +<span class="sourceLineNo">358</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.358"></a> +<span class="sourceLineNo">359</span> * job classes via the distributed cache (tmpjars).<a name="line.359"></a> +<span class="sourceLineNo">360</span> *<a name="line.360"></a> +<span class="sourceLineNo">361</span> * @param tmpRestoreDir a temporary directory to copy the snapshot files into. Current user should<a name="line.361"></a> +<span class="sourceLineNo">362</span> * have write permissions to this directory, and this should not be a subdirectory of rootdir.<a name="line.362"></a> +<span class="sourceLineNo">363</span> * After the job is finished, restore directory can be deleted.<a name="line.363"></a> +<span class="sourceLineNo">364</span> * @throws IOException When setting up the details fails.<a name="line.364"></a> +<span class="sourceLineNo">365</span> * @see TableSnapshotInputFormat<a name="line.365"></a> +<span class="sourceLineNo">366</span> */<a name="line.366"></a> +<span class="sourceLineNo">367</span> public static void initTableSnapshotMapperJob(String snapshotName, Scan scan,<a name="line.367"></a> +<span class="sourceLineNo">368</span> Class<? extends TableMapper> mapper,<a name="line.368"></a> +<span class="sourceLineNo">369</span> Class<?> outputKeyClass,<a name="line.369"></a> +<span class="sourceLineNo">370</span> Class<?> outputValueClass, Job job,<a name="line.370"></a> +<span class="sourceLineNo">371</span> boolean addDependencyJars, Path tmpRestoreDir)<a name="line.371"></a> +<span class="sourceLineNo">372</span> throws IOException {<a name="line.372"></a> +<span class="sourceLineNo">373</span> TableSnapshotInputFormat.setInput(job, snapshotName, tmpRestoreDir);<a name="line.373"></a> +<span class="sourceLineNo">374</span> initTableMapperJob(snapshotName, scan, mapper, outputKeyClass,<a name="line.374"></a> +<span class="sourceLineNo">375</span> outputValueClass, job, addDependencyJars, false, TableSnapshotInputFormat.class);<a name="line.375"></a> +<span class="sourceLineNo">376</span> resetCacheConfig(job.getConfiguration());<a name="line.376"></a> +<span class="sourceLineNo">377</span> }<a name="line.377"></a> +<span class="sourceLineNo">378</span><a name="line.378"></a> +<span class="sourceLineNo">379</span> /**<a name="line.379"></a> +<span class="sourceLineNo">380</span> * Sets up the job for reading from a table snapshot. It bypasses hbase servers<a name="line.380"></a> +<span class="sourceLineNo">381</span> * and read directly from snapshot files.<a name="line.381"></a> +<span class="sourceLineNo">382</span> *<a name="line.382"></a> +<span class="sourceLineNo">383</span> * @param snapshotName The name of the snapshot (of a table) to read from.<a name="line.383"></a> +<span class="sourceLineNo">384</span> * @param scan The scan instance with the columns, time range etc.<a name="line.384"></a> +<span class="sourceLineNo">385</span> * @param mapper The mapper class to use.<a name="line.385"></a> +<span class="sourceLineNo">386</span> * @param outputKeyClass The class of the output key.<a name="line.386"></a> +<span class="sourceLineNo">387</span> * @param outputValueClass The class of the output value.<a name="line.387"></a> +<span class="sourceLineNo">388</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.388"></a> +<span class="sourceLineNo">389</span> * carrying all necessary HBase configuration.<a name="line.389"></a> +<span class="sourceLineNo">390</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.390"></a> +<span class="sourceLineNo">391</span> * job classes via the distributed cache (tmpjars).<a name="line.391"></a> +<span class="sourceLineNo">392</span> *<a name="line.392"></a> +<span class="sourceLineNo">393</span> * @param tmpRestoreDir a temporary directory to copy the snapshot files into. Current user should<a name="line.393"></a> +<span class="sourceLineNo">394</span> * have write permissions to this directory, and this should not be a subdirectory of rootdir.<a name="line.394"></a> +<span class="sourceLineNo">395</span> * After the job is finished, restore directory can be deleted.<a name="line.395"></a> +<span class="sourceLineNo">396</span> * @param splitAlgo algorithm to split<a name="line.396"></a> +<span class="sourceLineNo">397</span> * @param numSplitsPerRegion
<TRUNCATED>
