| <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd"> |
| <html lang="en"> |
| <head> |
| <title>Source code</title> |
| <link rel="stylesheet" type="text/css" href="../../../../../../stylesheet.css" title="Style"> |
| </head> |
| <body> |
| <div class="sourceContainer"> |
| <pre><span class="sourceLineNo">001</span>/**<a name="line.1"></a> |
| <span class="sourceLineNo">002</span> *<a name="line.2"></a> |
| <span class="sourceLineNo">003</span> * Licensed to the Apache Software Foundation (ASF) under one<a name="line.3"></a> |
| <span class="sourceLineNo">004</span> * or more contributor license agreements. See the NOTICE file<a name="line.4"></a> |
| <span class="sourceLineNo">005</span> * distributed with this work for additional information<a name="line.5"></a> |
| <span class="sourceLineNo">006</span> * regarding copyright ownership. The ASF licenses this file<a name="line.6"></a> |
| <span class="sourceLineNo">007</span> * to you under the Apache License, Version 2.0 (the<a name="line.7"></a> |
| <span class="sourceLineNo">008</span> * "License"); you may not use this file except in compliance<a name="line.8"></a> |
| <span class="sourceLineNo">009</span> * with the License. You may obtain a copy of the License at<a name="line.9"></a> |
| <span class="sourceLineNo">010</span> *<a name="line.10"></a> |
| <span class="sourceLineNo">011</span> * http://www.apache.org/licenses/LICENSE-2.0<a name="line.11"></a> |
| <span class="sourceLineNo">012</span> *<a name="line.12"></a> |
| <span class="sourceLineNo">013</span> * Unless required by applicable law or agreed to in writing, software<a name="line.13"></a> |
| <span class="sourceLineNo">014</span> * distributed under the License is distributed on an "AS IS" BASIS,<a name="line.14"></a> |
| <span class="sourceLineNo">015</span> * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.<a name="line.15"></a> |
| <span class="sourceLineNo">016</span> * See the License for the specific language governing permissions and<a name="line.16"></a> |
| <span class="sourceLineNo">017</span> * limitations under the License.<a name="line.17"></a> |
| <span class="sourceLineNo">018</span> */<a name="line.18"></a> |
| <span class="sourceLineNo">019</span>package org.apache.hadoop.hbase.mapreduce;<a name="line.19"></a> |
| <span class="sourceLineNo">020</span><a name="line.20"></a> |
| <span class="sourceLineNo">021</span>import java.io.File;<a name="line.21"></a> |
| <span class="sourceLineNo">022</span>import java.io.IOException;<a name="line.22"></a> |
| <span class="sourceLineNo">023</span>import java.net.URL;<a name="line.23"></a> |
| <span class="sourceLineNo">024</span>import java.net.URLDecoder;<a name="line.24"></a> |
| <span class="sourceLineNo">025</span>import java.util.ArrayList;<a name="line.25"></a> |
| <span class="sourceLineNo">026</span>import java.util.Base64;<a name="line.26"></a> |
| <span class="sourceLineNo">027</span>import java.util.Collection;<a name="line.27"></a> |
| <span class="sourceLineNo">028</span>import java.util.Enumeration;<a name="line.28"></a> |
| <span class="sourceLineNo">029</span>import java.util.HashMap;<a name="line.29"></a> |
| <span class="sourceLineNo">030</span>import java.util.HashSet;<a name="line.30"></a> |
| <span class="sourceLineNo">031</span>import java.util.List;<a name="line.31"></a> |
| <span class="sourceLineNo">032</span>import java.util.Map;<a name="line.32"></a> |
| <span class="sourceLineNo">033</span>import java.util.Set;<a name="line.33"></a> |
| <span class="sourceLineNo">034</span>import java.util.zip.ZipEntry;<a name="line.34"></a> |
| <span class="sourceLineNo">035</span>import java.util.zip.ZipFile;<a name="line.35"></a> |
| <span class="sourceLineNo">036</span><a name="line.36"></a> |
| <span class="sourceLineNo">037</span>import org.apache.hadoop.conf.Configuration;<a name="line.37"></a> |
| <span class="sourceLineNo">038</span>import org.apache.hadoop.fs.FileSystem;<a name="line.38"></a> |
| <span class="sourceLineNo">039</span>import org.apache.hadoop.fs.Path;<a name="line.39"></a> |
| <span class="sourceLineNo">040</span>import org.apache.hadoop.hbase.HBaseConfiguration;<a name="line.40"></a> |
| <span class="sourceLineNo">041</span>import org.apache.hadoop.hbase.HConstants;<a name="line.41"></a> |
| <span class="sourceLineNo">042</span>import org.apache.hadoop.hbase.TableName;<a name="line.42"></a> |
| <span class="sourceLineNo">043</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.43"></a> |
| <span class="sourceLineNo">044</span>import org.slf4j.Logger;<a name="line.44"></a> |
| <span class="sourceLineNo">045</span>import org.slf4j.LoggerFactory;<a name="line.45"></a> |
| <span class="sourceLineNo">046</span>import org.apache.hadoop.hbase.client.Connection;<a name="line.46"></a> |
| <span class="sourceLineNo">047</span>import org.apache.hadoop.hbase.client.ConnectionFactory;<a name="line.47"></a> |
| <span class="sourceLineNo">048</span>import org.apache.hadoop.hbase.client.Put;<a name="line.48"></a> |
| <span class="sourceLineNo">049</span>import org.apache.hadoop.hbase.client.RegionLocator;<a name="line.49"></a> |
| <span class="sourceLineNo">050</span>import org.apache.hadoop.hbase.client.Scan;<a name="line.50"></a> |
| <span class="sourceLineNo">051</span>import org.apache.hadoop.hbase.io.ImmutableBytesWritable;<a name="line.51"></a> |
| <span class="sourceLineNo">052</span>import org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil;<a name="line.52"></a> |
| <span class="sourceLineNo">053</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.ClientProtos;<a name="line.53"></a> |
| <span class="sourceLineNo">054</span>import org.apache.hadoop.hbase.security.User;<a name="line.54"></a> |
| <span class="sourceLineNo">055</span>import org.apache.hadoop.hbase.security.UserProvider;<a name="line.55"></a> |
| <span class="sourceLineNo">056</span>import org.apache.hadoop.hbase.security.token.TokenUtil;<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.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 Logger LOG = LoggerFactory.getLogger(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> /**<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> /**<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> CellSerialization.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 and read directly<a name="line.348"></a> |
| <span class="sourceLineNo">349</span> * from snapshot files.<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 carrying all necessary HBase<a name="line.355"></a> |
| <span class="sourceLineNo">356</span> * configuration.<a name="line.356"></a> |
| <span class="sourceLineNo">357</span> * @param addDependencyJars upload HBase jars and jars for any of the configured job classes via<a name="line.357"></a> |
| <span class="sourceLineNo">358</span> * the distributed cache (tmpjars).<a name="line.358"></a> |
| <span class="sourceLineNo">359</span> * @param tmpRestoreDir a temporary directory to copy the snapshot files into. Current user should<a name="line.359"></a> |
| <span class="sourceLineNo">360</span> * have write permissions to this directory, and this should not be a subdirectory of<a name="line.360"></a> |
| <span class="sourceLineNo">361</span> * rootdir. After the job is finished, restore directory can be deleted.<a name="line.361"></a> |
| <span class="sourceLineNo">362</span> * @throws IOException When setting up the details fails.<a name="line.362"></a> |
| <span class="sourceLineNo">363</span> * @see TableSnapshotInputFormat<a name="line.363"></a> |
| <span class="sourceLineNo">364</span> */<a name="line.364"></a> |
| <span class="sourceLineNo">365</span> public static void initTableSnapshotMapperJob(String snapshotName, Scan scan,<a name="line.365"></a> |
| <span class="sourceLineNo">366</span> Class<? extends TableMapper> mapper,<a name="line.366"></a> |
| <span class="sourceLineNo">367</span> Class<?> outputKeyClass,<a name="line.367"></a> |
| <span class="sourceLineNo">368</span> Class<?> outputValueClass, Job job,<a name="line.368"></a> |
| <span class="sourceLineNo">369</span> boolean addDependencyJars, Path tmpRestoreDir)<a name="line.369"></a> |
| <span class="sourceLineNo">370</span> throws IOException {<a name="line.370"></a> |
| <span class="sourceLineNo">371</span> TableSnapshotInputFormat.setInput(job, snapshotName, tmpRestoreDir);<a name="line.371"></a> |
| <span class="sourceLineNo">372</span> initTableMapperJob(snapshotName, scan, mapper, outputKeyClass, outputValueClass, job,<a name="line.372"></a> |
| <span class="sourceLineNo">373</span> addDependencyJars, false, TableSnapshotInputFormat.class);<a name="line.373"></a> |
| <span class="sourceLineNo">374</span> resetCacheConfig(job.getConfiguration());<a name="line.374"></a> |
| <span class="sourceLineNo">375</span> }<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> * Sets up the job for reading from a table snapshot. It bypasses hbase servers<a name="line.378"></a> |
| <span class="sourceLineNo">379</span> * and read directly from snapshot files.<a name="line.379"></a> |
| <span class="sourceLineNo">380</span> *<a name="line.380"></a> |
| <span class="sourceLineNo">381</span> * @param snapshotName The name of the snapshot (of a table) to read from.<a name="line.381"></a> |
| <span class="sourceLineNo">382</span> * @param scan The scan instance with the columns, time range etc.<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<a name="line.386"></a> |
| <span class="sourceLineNo">387</span> * carrying all necessary HBase configuration.<a name="line.387"></a> |
| <span class="sourceLineNo">388</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.388"></a> |
| <span class="sourceLineNo">389</span> * job classes via the distributed cache (tmpjars).<a name="line.389"></a> |
| <span class="sourceLineNo">390</span> *<a name="line.390"></a> |
| <span class="sourceLineNo">391</span> * @param tmpRestoreDir a temporary directory to copy the snapshot files into. Current user should<a name="line.391"></a> |
| <span class="sourceLineNo">392</span> * have write permissions to this directory, and this should not be a subdirectory of rootdir.<a name="line.392"></a> |
| <span class="sourceLineNo">393</span> * After the job is finished, restore directory can be deleted.<a name="line.393"></a> |
| <span class="sourceLineNo">394</span> * @param splitAlgo algorithm to split<a name="line.394"></a> |
| <span class="sourceLineNo">395</span> * @param numSplitsPerRegion how many input splits to generate per one region<a name="line.395"></a> |
| <span class="sourceLineNo">396</span> * @throws IOException When setting up the details fails.<a name="line.396"></a> |
| <span class="sourceLineNo">397</span> * @see TableSnapshotInputFormat<a name="line.397"></a> |
| <span class="sourceLineNo">398</span> */<a name="line.398"></a> |
| <span class="sourceLineNo">399</span> public static void initTableSnapshotMapperJob(String snapshotName, Scan scan,<a name="line.399"></a> |
| <span class="sourceLineNo">400</span> Class<? extends TableMapper> mapper,<a name="line.400"></a> |
| <span class="sourceLineNo">401</span> Class<?> outputKeyClass,<a name="line.401"></a> |
| <span class="sourceLineNo">402</span> Class<?> outputValueClass, Job job,<a name="line.402"></a> |
| <span class="sourceLineNo">403</span> boolean addDependencyJars, Path tmpRestoreDir,<a name="line.403"></a> |
| <span class="sourceLineNo">404</span> RegionSplitter.SplitAlgorithm splitAlgo,<a name="line.404"></a> |
| <span class="sourceLineNo">405</span> int numSplitsPerRegion)<a name="line.405"></a> |
| <span class="sourceLineNo">406</span> throws IOException {<a name="line.406"></a> |
| <span class="sourceLineNo">407</span> TableSnapshotInputFormat.setInput(job, snapshotName, tmpRestoreDir, splitAlgo,<a name="line.407"></a> |
| <span class="sourceLineNo">408</span> numSplitsPerRegion);<a name="line.408"></a> |
| <span class="sourceLineNo">409</span> initTableMapperJob(snapshotName, scan, mapper, outputKeyClass,<a name="line.409"></a> |
| <span class="sourceLineNo">410</span> outputValueClass, job, addDependencyJars, false, TableSnapshotInputFormat.class);<a name="line.410"></a> |
| <span class="sourceLineNo">411</span> resetCacheConfig(job.getConfiguration());<a name="line.411"></a> |
| <span class="sourceLineNo">412</span> }<a name="line.412"></a> |
| <span class="sourceLineNo">413</span><a name="line.413"></a> |
| <span class="sourceLineNo">414</span> /**<a name="line.414"></a> |
| <span class="sourceLineNo">415</span> * Use this before submitting a Multi TableMap job. It will appropriately set<a name="line.415"></a> |
| <span class="sourceLineNo">416</span> * up the job.<a name="line.416"></a> |
| <span class="sourceLineNo">417</span> *<a name="line.417"></a> |
| <span class="sourceLineNo">418</span> * @param scans The list of {@link Scan} objects to read from.<a name="line.418"></a> |
| <span class="sourceLineNo">419</span> * @param mapper The mapper class to use.<a name="line.419"></a> |
| <span class="sourceLineNo">420</span> * @param outputKeyClass The class of the output key.<a name="line.420"></a> |
| <span class="sourceLineNo">421</span> * @param outputValueClass The class of the output value.<a name="line.421"></a> |
| <span class="sourceLineNo">422</span> * @param job The current job to adjust. Make sure the passed job is carrying<a name="line.422"></a> |
| <span class="sourceLineNo">423</span> * all necessary HBase configuration.<a name="line.423"></a> |
| <span class="sourceLineNo">424</span> * @throws IOException When setting up the details fails.<a name="line.424"></a> |
| <span class="sourceLineNo">425</span> */<a name="line.425"></a> |
| <span class="sourceLineNo">426</span> public static void initTableMapperJob(List<Scan> scans,<a name="line.426"></a> |
| <span class="sourceLineNo">427</span> Class<? extends TableMapper> mapper,<a name="line.427"></a> |
| <span class="sourceLineNo">428</span> Class<?> outputKeyClass,<a name="line.428"></a> |
| <span class="sourceLineNo">429</span> Class<?> outputValueClass, Job job) throws IOException {<a name="line.429"></a> |
| <span class="sourceLineNo">430</span> initTableMapperJob(scans, mapper, outputKeyClass, outputValueClass, job,<a name="line.430"></a> |
| <span class="sourceLineNo">431</span> true);<a name="line.431"></a> |
| <span class="sourceLineNo">432</span> }<a name="line.432"></a> |
| <span class="sourceLineNo">433</span><a name="line.433"></a> |
| <span class="sourceLineNo">434</span> /**<a name="line.434"></a> |
| <span class="sourceLineNo">435</span> * Use this before submitting a Multi TableMap job. It will appropriately set<a name="line.435"></a> |
| <span class="sourceLineNo">436</span> * up the job.<a name="line.436"></a> |
| <span class="sourceLineNo">437</span> *<a name="line.437"></a> |
| <span class="sourceLineNo">438</span> * @param scans The list of {@link Scan} objects to read from.<a name="line.438"></a> |
| <span class="sourceLineNo">439</span> * @param mapper The mapper class to use.<a name="line.439"></a> |
| <span class="sourceLineNo">440</span> * @param outputKeyClass The class of the output key.<a name="line.440"></a> |
| <span class="sourceLineNo">441</span> * @param outputValueClass The class of the output value.<a name="line.441"></a> |
| <span class="sourceLineNo">442</span> * @param job The current job to adjust. Make sure the passed job is carrying<a name="line.442"></a> |
| <span class="sourceLineNo">443</span> * all necessary HBase configuration.<a name="line.443"></a> |
| <span class="sourceLineNo">444</span> * @param addDependencyJars upload HBase jars and jars for any of the<a name="line.444"></a> |
| <span class="sourceLineNo">445</span> * configured job classes via the distributed cache (tmpjars).<a name="line.445"></a> |
| <span class="sourceLineNo">446</span> * @throws IOException When setting up the details fails.<a name="line.446"></a> |
| <span class="sourceLineNo">447</span> */<a name="line.447"></a> |
| <span class="sourceLineNo">448</span> public static void initTableMapperJob(List<Scan> scans,<a name="line.448"></a> |
| <span class="sourceLineNo">449</span> Class<? extends TableMapper> mapper,<a name="line.449"></a> |
| <span class="sourceLineNo">450</span> Class<?> outputKeyClass,<a name="line.450"></a> |
| <span class="sourceLineNo">451</span> Class<?> outputValueClass, Job job,<a name="line.451"></a> |
| <span class="sourceLineNo">452</span> boolean addDependencyJars) throws IOException {<a name="line.452"></a> |
| <span class="sourceLineNo">453</span> initTableMapperJob(scans, mapper, outputKeyClass, outputValueClass, job,<a name="line.453"></a> |
| <span class="sourceLineNo">454</span> addDependencyJars, true);<a name="line.454"></a> |
| <span class="sourceLineNo">455</span> }<a name="line.455"></a> |
| <span class="sourceLineNo">456</span><a name="line.456"></a> |
| <span class="sourceLineNo">457</span> /**<a name="line.457"></a> |
| <span class="sourceLineNo">458</span> * Use this before submitting a Multi TableMap job. It will appropriately set<a name="line.458"></a> |
| <span class="sourceLineNo">459</span> * up the job.<a name="line.459"></a> |
| <span class="sourceLineNo">460</span> *<a name="line.460"></a> |
| <span class="sourceLineNo">461</span> * @param scans The list of {@link Scan} objects to read from.<a name="line.461"></a> |
| <span class="sourceLineNo">462</span> * @param mapper The mapper class to use.<a name="line.462"></a> |
| <span class="sourceLineNo">463</span> * @param outputKeyClass The class of the output key.<a name="line.463"></a> |
| <span class="sourceLineNo">464</span> * @param outputValueClass The class of the output value.<a name="line.464"></a> |
| <span class="sourceLineNo">465</span> * @param job The current job to adjust. Make sure the passed job is carrying<a name="line.465"></a> |
| <span class="sourceLineNo">466</span> * all necessary HBase configuration.<a name="line.466"></a> |
| <span class="sourceLineNo">467</span> * @param addDependencyJars upload HBase jars and jars for any of the<a name="line.467"></a> |
| <span class="sourceLineNo">468</span> * configured job classes via the distributed cache (tmpjars).<a name="line.468"></a> |
| <span class="sourceLineNo">469</span> * @param initCredentials whether to initialize hbase auth credentials for the job<a name="line.469"></a> |
| <span class="sourceLineNo">470</span> * @throws IOException When setting up the details fails.<a name="line.470"></a> |
| <span class="sourceLineNo">471</span> */<a name="line.471"></a> |
| <span class="sourceLineNo">472</span> public static void initTableMapperJob(List<Scan> scans,<a name="line.472"></a> |
| <span class="sourceLineNo">473</span> Class<? extends TableMapper> mapper,<a name="line.473"></a> |
| <span class="sourceLineNo">474</span> Class<?> outputKeyClass,<a name="line.474"></a> |
| <span class="sourceLineNo">475</span> Class<?> outputValueClass, Job job,<a name="line.475"></a> |
| <span class="sourceLineNo">476</span> boolean addDependencyJars,<a name="line.476"></a> |
| <span class="sourceLineNo">477</span> boolean initCredentials) throws IOException {<a name="line.477"></a> |
| <span class="sourceLineNo">478</span> job.setInputFormatClass(MultiTableInputFormat.class);<a name="line.478"></a> |
| <span class="sourceLineNo">479</span> if (outputValueClass != null) {<a name="line.479"></a> |
| <span class="sourceLineNo">480</span> job.setMapOutputValueClass(outputValueClass);<a name="line.480"></a> |
| <span class="sourceLineNo">481</span> }<a name="line.481"></a> |
| <span class="sourceLineNo">482</span> if (outputKeyClass != null) {<a name="line.482"></a> |
| <span class="sourceLineNo">483</span> job.setMapOutputKeyClass(outputKeyClass);<a name="line.483"></a> |
| <span class="sourceLineNo">484</span> }<a name="line.484"></a> |
| <span class="sourceLineNo">485</span> job.setMapperClass(mapper);<a name="line.485"></a> |
| <span class="sourceLineNo">486</span> Configuration conf = job.getConfiguration();<a name="line.486"></a> |
| <span class="sourceLineNo">487</span> HBaseConfiguration.merge(conf, HBaseConfiguration.create(conf));<a name="line.487"></a> |
| <span class="sourceLineNo">488</span> List<String> scanStrings = new ArrayList<>();<a name="line.488"></a> |
| <span class="sourceLineNo">489</span><a name="line.489"></a> |
| <span class="sourceLineNo">490</span> for (Scan scan : scans) {<a name="line.490"></a> |
| <span class="sourceLineNo">491</span> scanStrings.add(convertScanToString(scan));<a name="line.491"></a> |
| <span class="sourceLineNo">492</span> }<a name="line.492"></a> |
| <span class="sourceLineNo">493</span> job.getConfiguration().setStrings(MultiTableInputFormat.SCANS,<a name="line.493"></a> |
| <span class="sourceLineNo">494</span> scanStrings.toArray(new String[scanStrings.size()]));<a name="line.494"></a> |
| <span class="sourceLineNo">495</span><a name="line.495"></a> |
| <span class="sourceLineNo">496</span> if (addDependencyJars) {<a name="line.496"></a> |
| <span class="sourceLineNo">497</span> addDependencyJars(job);<a name="line.497"></a> |
| <span class="sourceLineNo">498</span> }<a name="line.498"></a> |
| <span class="sourceLineNo">499</span><a name="line.499"></a> |
| <span class="sourceLineNo">500</span> if (initCredentials) {<a name="line.500"></a> |
| <span class="sourceLineNo">501</span> initCredentials(job);<a name="line.501"></a> |
| <span class="sourceLineNo">502</span> }<a name="line.502"></a> |
| <span class="sourceLineNo">503</span> }<a name="line.503"></a> |
| <span class="sourceLineNo">504</span><a name="line.504"></a> |
| <span class="sourceLineNo">505</span> public static void initCredentials(Job job) throws IOException {<a name="line.505"></a> |
| <span class="sourceLineNo">506</span> UserProvider userProvider = UserProvider.instantiate(job.getConfiguration());<a name="line.506"></a> |
| <span class="sourceLineNo">507</span> if (userProvider.isHadoopSecurityEnabled()) {<a name="line.507"></a> |
| <span class="sourceLineNo">508</span> // propagate delegation related props from launcher job to MR job<a name="line.508"></a> |
| <span class="sourceLineNo">509</span> if (System.getenv("HADOOP_TOKEN_FILE_LOCATION") != null) {<a name="line.509"></a> |
| <span class="sourceLineNo">510</span> job.getConfiguration().set("mapreduce.job.credentials.binary",<a name="line.510"></a> |
| <span class="sourceLineNo">511</span> System.getenv("HADOOP_TOKEN_FILE_LOCATION"));<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> if (userProvider.isHBaseSecurityEnabled()) {<a name="line.515"></a> |
| <span class="sourceLineNo">516</span> try {<a name="line.516"></a> |
| <span class="sourceLineNo">517</span> // init credentials for remote cluster<a name="line.517"></a> |
| <span class="sourceLineNo">518</span> String quorumAddress = job.getConfiguration().get(TableOutputFormat.QUORUM_ADDRESS);<a name="line.518"></a> |
| <span class="sourceLineNo">519</span> User user = userProvider.getCurrent();<a name="line.519"></a> |
| <span class="sourceLineNo">520</span> if (quorumAddress != null) {<a name="line.520"></a> |
| <span class="sourceLineNo">521</span> Configuration peerConf = HBaseConfiguration.createClusterConf(job.getConfiguration(),<a name="line.521"></a> |
| <span class="sourceLineNo">522</span> quorumAddress, TableOutputFormat.OUTPUT_CONF_PREFIX);<a name="line.522"></a> |
| <span class="sourceLineNo">523</span> Connection peerConn = ConnectionFactory.createConnection(peerConf);<a name="line.523"></a> |
| <span class="sourceLineNo">524</span> try {<a name="line.524"></a> |
| <span class="sourceLineNo">525</span> TokenUtil.addTokenForJob(peerConn, user, job);<a name="line.525"></a> |
| <span class="sourceLineNo">526</span> } finally {<a name="line.526"></a> |
| <span class="sourceLineNo">527</span> peerConn.close();<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><a name="line.530"></a> |
| <span class="sourceLineNo">531</span> Connection conn = ConnectionFactory.createConnection(job.getConfiguration());<a name="line.531"></a> |
| <span class="sourceLineNo">532</span> try {<a name="line.532"></a> |
| <span class="sourceLineNo">533</span> TokenUtil.addTokenForJob(conn, user, job);<a name="line.533"></a> |
| <span class="sourceLineNo">534</span> } finally {<a name="line.534"></a> |
| <span class="sourceLineNo">535</span> conn.close();<a name="line.535"></a> |
| <span class="sourceLineNo">536</span> }<a name="line.536"></a> |
| <span class="sourceLineNo">537</span> } catch (InterruptedException ie) {<a name="line.537"></a> |
| <span class="sourceLineNo">538</span> LOG.info("Interrupted obtaining user authentication token");<a name="line.538"></a> |
| <span class="sourceLineNo">539</span> Thread.currentThread().interrupt();<a name="line.539"></a> |
| <span class="sourceLineNo">540</span> }<a name="line.540"></a> |
| <span class="sourceLineNo">541</span> }<a name="line.541"></a> |
| <span class="sourceLineNo">542</span> }<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> * Obtain an authentication token, for the specified cluster, on behalf of the current user<a name="line.545"></a> |
| <span class="sourceLineNo">546</span> * and add it to the credentials for the given map reduce job.<a name="line.546"></a> |
| <span class="sourceLineNo">547</span> *<a name="line.547"></a> |
| <span class="sourceLineNo">548</span> * The quorumAddress is the key to the ZK ensemble, which contains:<a name="line.548"></a> |
| <span class="sourceLineNo">549</span> * hbase.zookeeper.quorum, hbase.zookeeper.client.port and<a name="line.549"></a> |
| <span class="sourceLineNo">550</span> * zookeeper.znode.parent<a name="line.550"></a> |
| <span class="sourceLineNo">551</span> *<a name="line.551"></a> |
| <span class="sourceLineNo">552</span> * @param job The job that requires the permission.<a name="line.552"></a> |
| <span class="sourceLineNo">553</span> * @param quorumAddress string that contains the 3 required configuratins<a name="line.553"></a> |
| <span class="sourceLineNo">554</span> * @throws IOException When the authentication token cannot be obtained.<a name="line.554"></a> |
| <span class="sourceLineNo">555</span> * @deprecated Since 1.2.0 and will be removed in 3.0.0. Use<a name="line.555"></a> |
| <span class="sourceLineNo">556</span> * {@link #initCredentialsForCluster(Job, Configuration)} instead.<a name="line.556"></a> |
| <span class="sourceLineNo">557</span> * @see #initCredentialsForCluster(Job, Configuration)<a name="line.557"></a> |
| <span class="sourceLineNo">558</span> * @see <a href="https://issues.apache.org/jira/browse/HBASE-14886">HBASE-14886</a><a name="line.558"></a> |
| <span class="sourceLineNo">559</span> */<a name="line.559"></a> |
| <span class="sourceLineNo">560</span> @Deprecated<a name="line.560"></a> |
| <span class="sourceLineNo">561</span> public static void initCredentialsForCluster(Job job, String quorumAddress)<a name="line.561"></a> |
| <span class="sourceLineNo">562</span> throws IOException {<a name="line.562"></a> |
| <span class="sourceLineNo">563</span> Configuration peerConf = HBaseConfiguration.createClusterConf(job.getConfiguration(),<a name="line.563"></a> |
| <span class="sourceLineNo">564</span> quorumAddress);<a name="line.564"></a> |
| <span class="sourceLineNo">565</span> initCredentialsForCluster(job, peerConf);<a name="line.565"></a> |
| <span class="sourceLineNo">566</span> }<a name="line.566"></a> |
| <span class="sourceLineNo">567</span><a name="line.567"></a> |
| <span class="sourceLineNo">568</span> /**<a name="line.568"></a> |
| <span class="sourceLineNo">569</span> * Obtain an authentication token, for the specified cluster, on behalf of the current user<a name="line.569"></a> |
| <span class="sourceLineNo">570</span> * and add it to the credentials for the given map reduce job.<a name="line.570"></a> |
| <span class="sourceLineNo">571</span> *<a name="line.571"></a> |
| <span class="sourceLineNo">572</span> * @param job The job that requires the permission.<a name="line.572"></a> |
| <span class="sourceLineNo">573</span> * @param conf The configuration to use in connecting to the peer cluster<a name="line.573"></a> |
| <span class="sourceLineNo">574</span> * @throws IOException When the authentication token cannot be obtained.<a name="line.574"></a> |
| <span class="sourceLineNo">575</span> */<a name="line.575"></a> |
| <span class="sourceLineNo">576</span> public static void initCredentialsForCluster(Job job, Configuration conf)<a name="line.576"></a> |
| <span class="sourceLineNo">577</span> throws IOException {<a name="line.577"></a> |
| <span class="sourceLineNo">578</span> UserProvider userProvider = UserProvider.instantiate(job.getConfiguration());<a name="line.578"></a> |
| <span class="sourceLineNo">579</span> if (userProvider.isHBaseSecurityEnabled()) {<a name="line.579"></a> |
| <span class="sourceLineNo">580</span> try {<a name="line.580"></a> |
| <span class="sourceLineNo">581</span> Connection peerConn = ConnectionFactory.createConnection(conf);<a name="line.581"></a> |
| <span class="sourceLineNo">582</span> try {<a name="line.582"></a> |
| <span class="sourceLineNo">583</span> TokenUtil.addTokenForJob(peerConn, userProvider.getCurrent(), job);<a name="line.583"></a> |
| <span class="sourceLineNo">584</span> } finally {<a name="line.584"></a> |
| <span class="sourceLineNo">585</span> peerConn.close();<a name="line.585"></a> |
| <span class="sourceLineNo">586</span> }<a name="line.586"></a> |
| <span class="sourceLineNo">587</span> } catch (InterruptedException e) {<a name="line.587"></a> |
| <span class="sourceLineNo">588</span> LOG.info("Interrupted obtaining user authentication token");<a name="line.588"></a> |
| <span class="sourceLineNo">589</span> Thread.interrupted();<a name="line.589"></a> |
| <span class="sourceLineNo">590</span> }<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> /**<a name="line.594"></a> |
| <span class="sourceLineNo">595</span> * Writes the given scan into a Base64 encoded string.<a name="line.595"></a> |
| <span class="sourceLineNo">596</span> *<a name="line.596"></a> |
| <span class="sourceLineNo">597</span> * @param scan The scan to write out.<a name="line.597"></a> |
| <span class="sourceLineNo">598</span> * @return The scan saved in a Base64 encoded string.<a name="line.598"></a> |
| <span class="sourceLineNo">599</span> * @throws IOException When writing the scan fails.<a name="line.599"></a> |
| <span class="sourceLineNo">600</span> */<a name="line.600"></a> |
| <span class="sourceLineNo">601</span> public static String convertScanToString(Scan scan) throws IOException {<a name="line.601"></a> |
| <span class="sourceLineNo">602</span> ClientProtos.Scan proto = ProtobufUtil.toScan(scan);<a name="line.602"></a> |
| <span class="sourceLineNo">603</span> return Bytes.toString(Base64.getEncoder().encode(proto.toByteArray()));<a name="line.603"></a> |
| <span class="sourceLineNo">604</span> }<a name="line.604"></a> |
| <span class="sourceLineNo">605</span><a name="line.605"></a> |
| <span class="sourceLineNo">606</span> /**<a name="line.606"></a> |
| <span class="sourceLineNo">607</span> * Converts the given Base64 string back into a Scan instance.<a name="line.607"></a> |
| <span class="sourceLineNo">608</span> *<a name="line.608"></a> |
| <span class="sourceLineNo">609</span> * @param base64 The scan details.<a name="line.609"></a> |
| <span class="sourceLineNo">610</span> * @return The newly created Scan instance.<a name="line.610"></a> |
| <span class="sourceLineNo">611</span> * @throws IOException When reading the scan instance fails.<a name="line.611"></a> |
| <span class="sourceLineNo">612</span> */<a name="line.612"></a> |
| <span class="sourceLineNo">613</span> public static Scan convertStringToScan(String base64) throws IOException {<a name="line.613"></a> |
| <span class="sourceLineNo">614</span> byte [] decoded = Base64.getDecoder().decode(base64);<a name="line.614"></a> |
| <span class="sourceLineNo">615</span> return ProtobufUtil.toScan(ClientProtos.Scan.parseFrom(decoded));<a name="line.615"></a> |
| <span class="sourceLineNo">616</span> }<a name="line.616"></a> |
| <span class="sourceLineNo">617</span><a name="line.617"></a> |
| <span class="sourceLineNo">618</span> /**<a name="line.618"></a> |
| <span class="sourceLineNo">619</span> * Use this before submitting a TableReduce job. It will<a name="line.619"></a> |
| <span class="sourceLineNo">620</span> * appropriately set up the JobConf.<a name="line.620"></a> |
| <span class="sourceLineNo">621</span> *<a name="line.621"></a> |
| <span class="sourceLineNo">622</span> * @param table The output table.<a name="line.622"></a> |
| <span class="sourceLineNo">623</span> * @param reducer The reducer class to use.<a name="line.623"></a> |
| <span class="sourceLineNo">624</span> * @param job The current job to adjust.<a name="line.624"></a> |
| <span class="sourceLineNo">625</span> * @throws IOException When determining the region count fails.<a name="line.625"></a> |
| <span class="sourceLineNo">626</span> */<a name="line.626"></a> |
| <span class="sourceLineNo">627</span> public static void initTableReducerJob(String table,<a name="line.627"></a> |
| <span class="sourceLineNo">628</span> Class<? extends TableReducer> reducer, Job job)<a name="line.628"></a> |
| <span class="sourceLineNo">629</span> throws IOException {<a name="line.629"></a> |
| <span class="sourceLineNo">630</span> initTableReducerJob(table, reducer, job, null);<a name="line.630"></a> |
| <span class="sourceLineNo">631</span> }<a name="line.631"></a> |
| <span class="sourceLineNo">632</span><a name="line.632"></a> |
| <span class="sourceLineNo">633</span> /**<a name="line.633"></a> |
| <span class="sourceLineNo">634</span> * Use this before submitting a TableReduce job. It will<a name="line.634"></a> |
| <span class="sourceLineNo">635</span> * appropriately set up the JobConf.<a name="line.635"></a> |
| <span class="sourceLineNo">636</span> *<a name="line.636"></a> |
| <span class="sourceLineNo">637</span> * @param table The output table.<a name="line.637"></a> |
| <span class="sourceLineNo">638</span> * @param reducer The reducer class to use.<a name="line.638"></a> |
| <span class="sourceLineNo">639</span> * @param job The current job to adjust.<a name="line.639"></a> |
| <span class="sourceLineNo">640</span> * @param partitioner Partitioner to use. Pass <code>null</code> to use<a name="line.640"></a> |
| <span class="sourceLineNo">641</span> * default partitioner.<a name="line.641"></a> |
| <span class="sourceLineNo">642</span> * @throws IOException When determining the region count fails.<a name="line.642"></a> |
| <span class="sourceLineNo">643</span> */<a name="line.643"></a> |
| <span class="sourceLineNo">644</span> public static void initTableReducerJob(String table,<a name="line.644"></a> |
| <span class="sourceLineNo">645</span> Class<? extends TableReducer> reducer, Job job,<a name="line.645"></a> |
| <span class="sourceLineNo">646</span> Class partitioner) throws IOException {<a name="line.646"></a> |
| <span class="sourceLineNo">647</span> initTableReducerJob(table, reducer, job, partitioner, null, null, null);<a name="line.647"></a> |
| <span class="sourceLineNo">648</span> }<a name="line.648"></a> |
| <span class="sourceLineNo">649</span><a name="line.649"></a> |
| <span class="sourceLineNo">650</span> /**<a name="line.650"></a> |
| <span class="sourceLineNo">651</span> * Use this before submitting a TableReduce job. It will<a name="line.651"></a> |
| <span class="sourceLineNo">652</span> * appropriately set up the JobConf.<a name="line.652"></a> |
| <span class="sourceLineNo">653</span> *<a name="line.653"></a> |
| <span class="sourceLineNo">654</span> * @param table The output table.<a name="line.654"></a> |
| <span class="sourceLineNo">655</span> * @param reducer The reducer class to use.<a name="line.655"></a> |
| <span class="sourceLineNo">656</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.656"></a> |
| <span class="sourceLineNo">657</span> * carrying all necessary HBase configuration.<a name="line.657"></a> |
| <span class="sourceLineNo">658</span> * @param partitioner Partitioner to use. Pass <code>null</code> to use<a name="line.658"></a> |
| <span class="sourceLineNo">659</span> * default partitioner.<a name="line.659"></a> |
| <span class="sourceLineNo">660</span> * @param quorumAddress Distant cluster to write to; default is null for<a name="line.660"></a> |
| <span class="sourceLineNo">661</span> * output to the cluster that is designated in <code>hbase-site.xml</code>.<a name="line.661"></a> |
| <span class="sourceLineNo">662</span> * Set this String to the zookeeper ensemble of an alternate remote cluster<a name="line.662"></a> |
| <span class="sourceLineNo">663</span> * when you would have the reduce write a cluster that is other than the<a name="line.663"></a> |
| <span class="sourceLineNo">664</span> * default; e.g. copying tables between clusters, the source would be<a name="line.664"></a> |
| <span class="sourceLineNo">665</span> * designated by <code>hbase-site.xml</code> and this param would have the<a name="line.665"></a> |
| <span class="sourceLineNo">666</span> * ensemble address of the remote cluster. The format to pass is particular.<a name="line.666"></a> |
| <span class="sourceLineNo">667</span> * Pass <code> &lt;hbase.zookeeper.quorum&gt;:&lt;<a name="line.667"></a> |
| <span class="sourceLineNo">668</span> * hbase.zookeeper.client.port&gt;:&lt;zookeeper.znode.parent&gt;<a name="line.668"></a> |
| <span class="sourceLineNo">669</span> * </code> such as <code>server,server2,server3:2181:/hbase</code>.<a name="line.669"></a> |
| <span class="sourceLineNo">670</span> * @param serverClass redefined hbase.regionserver.class<a name="line.670"></a> |
| <span class="sourceLineNo">671</span> * @param serverImpl redefined hbase.regionserver.impl<a name="line.671"></a> |
| <span class="sourceLineNo">672</span> * @throws IOException When determining the region count fails.<a name="line.672"></a> |
| <span class="sourceLineNo">673</span> */<a name="line.673"></a> |
| <span class="sourceLineNo">674</span> public static void initTableReducerJob(String table,<a name="line.674"></a> |
| <span class="sourceLineNo">675</span> Class<? extends TableReducer> reducer, Job job,<a name="line.675"></a> |
| <span class="sourceLineNo">676</span> Class partitioner, String quorumAddress, String serverClass,<a name="line.676"></a> |
| <span class="sourceLineNo">677</span> String serverImpl) throws IOException {<a name="line.677"></a> |
| <span class="sourceLineNo">678</span> initTableReducerJob(table, reducer, job, partitioner, quorumAddress,<a name="line.678"></a> |
| <span class="sourceLineNo">679</span> serverClass, serverImpl, true);<a name="line.679"></a> |
| <span class="sourceLineNo">680</span> }<a name="line.680"></a> |
| <span class="sourceLineNo">681</span><a name="line.681"></a> |
| <span class="sourceLineNo">682</span> /**<a name="line.682"></a> |
| <span class="sourceLineNo">683</span> * Use this before submitting a TableReduce job. It will<a name="line.683"></a> |
| <span class="sourceLineNo">684</span> * appropriately set up the JobConf.<a name="line.684"></a> |
| <span class="sourceLineNo">685</span> *<a name="line.685"></a> |
| <span class="sourceLineNo">686</span> * @param table The output table.<a name="line.686"></a> |
| <span class="sourceLineNo">687</span> * @param reducer The reducer class to use.<a name="line.687"></a> |
| <span class="sourceLineNo">688</span> * @param job The current job to adjust. Make sure the passed job is<a name="line.688"></a> |
| <span class="sourceLineNo">689</span> * carrying all necessary HBase configuration.<a name="line.689"></a> |
| <span class="sourceLineNo">690</span> * @param partitioner Partitioner to use. Pass <code>null</code> to use<a name="line.690"></a> |
| <span class="sourceLineNo">691</span> * default partitioner.<a name="line.691"></a> |
| <span class="sourceLineNo">692</span> * @param quorumAddress Distant cluster to write to; default is null for<a name="line.692"></a> |
| <span class="sourceLineNo">693</span> * output to the cluster that is designated in <code>hbase-site.xml</code>.<a name="line.693"></a> |
| <span class="sourceLineNo">694</span> * Set this String to the zookeeper ensemble of an alternate remote cluster<a name="line.694"></a> |
| <span class="sourceLineNo">695</span> * when you would have the reduce write a cluster that is other than the<a name="line.695"></a> |
| <span class="sourceLineNo">696</span> * default; e.g. copying tables between clusters, the source would be<a name="line.696"></a> |
| <span class="sourceLineNo">697</span> * designated by <code>hbase-site.xml</code> and this param would have the<a name="line.697"></a> |
| <span class="sourceLineNo">698</span> * ensemble address of the remote cluster. The format to pass is particular.<a name="line.698"></a> |
| <span class="sourceLineNo">699</span> * Pass <code> &lt;hbase.zookeeper.quorum&gt;:&lt;<a name="line.699"></a> |
| <span class="sourceLineNo">700</span> * hbase.zookeeper.client.port&gt;:&lt;zookeeper.znode.parent&gt;<a name="line.700"></a> |
| <span class="sourceLineNo">701</span> * </code> such as <code>server,server2,server3:2181:/hbase</code>.<a name="line.701"></a> |
| <span class="sourceLineNo">702</span> * @param serverClass redefined hbase.regionserver.class<a name="line.702"></a> |
| <span class="sourceLineNo">703</span> * @param serverImpl redefined hbase.regionserver.impl<a name="line.703"></a> |
| <span class="sourceLineNo">704</span> * @param addDependencyJars upload HBase jars and jars for any of the configured<a name="line.704"></a> |
| <span class="sourceLineNo">705</span> * job classes via the distributed cache (tmpjars).<a name="line.705"></a> |
| <span class="sourceLineNo">706</span> * @throws IOException When determining the region count fails.<a name="line.706"></a> |
| <span class="sourceLineNo">707</span> */<a name="line.707"></a> |
| <span class="sourceLineNo">708</span> public static void initTableReducerJob(String table,<a name="line.708"></a> |
| <span class="sourceLineNo">709</span> Class<? extends TableReducer> reducer, Job job,<a name="line.709"></a> |
| <span class="sourceLineNo">710</span> Class partitioner, String quorumAddress, String serverClass,<a name="line.710"></a> |
| <span class="sourceLineNo">711</span> String serverImpl, boolean addDependencyJars) throws IOException {<a name="line.711"></a> |
| <span class="sourceLineNo">712</span><a name="line.712"></a> |
| <span class="sourceLineNo">713</span> Configuration conf = job.getConfiguration();<a name="line.713"></a> |
| <span class="sourceLineNo">714</span> HBaseConfiguration.merge(conf, HBaseConfiguration.create(conf));<a name="line.714"></a> |
| <span class="sourceLineNo">715</span> job.setOutputFormatClass(TableOutputFormat.class);<a name="line.715"></a> |
| <span class="sourceLineNo">716</span> if (reducer != null) job.setReducerClass(reducer);<a name="line.716"></a> |
| <span class="sourceLineNo">717</span> conf.set(TableOutputFormat.OUTPUT_TABLE, table);<a name="line.717"></a> |
| <span class="sourceLineNo">718</span> conf.setStrings("io.serializations", conf.get("io.serializations"),<a name="line.718"></a> |
| <span class="sourceLineNo">719</span> MutationSerialization.class.getName(), ResultSerialization.class.getName());<a name="line.719"></a> |
| <span class="sourceLineNo">720</span> // If passed a quorum/ensemble address, pass it on to TableOutputFormat.<a name="line.720"></a> |
| <span class="sourceLineNo">721</span> if (quorumAddress != null) {<a name="line.721"></a> |
| <span class="sourceLineNo">722</span> // Calling this will validate the format<a name="line.722"></a> |
| <span class="sourceLineNo">723</span> ZKConfig.validateClusterKey(quorumAddress);<a name="line.723"></a> |
| <span class="sourceLineNo">724</span> conf.set(TableOutputFormat.QUORUM_ADDRESS,quorumAddress);<a name="line.724"></a> |
| <span class="sourceLineNo">725</span> }<a name="line.725"></a> |
| <span class="sourceLineNo">726</span> if (serverClass != null && serverImpl != null) {<a name="line.726"></a> |
| <span class="sourceLineNo">727</span> conf.set(TableOutputFormat.REGION_SERVER_CLASS, serverClass);<a name="line.727"></a> |
| <span class="sourceLineNo">728</span> conf.set(TableOutputFormat.REGION_SERVER_IMPL, serverImpl);<a name="line.728"></a> |
| <span class="sourceLineNo">729</span> }<a name="line.729"></a> |
| <span class="sourceLineNo">730</span> job.setOutputKeyClass(ImmutableBytesWritable.class);<a name="line.730"></a> |
| <span class="sourceLineNo">731</span> job.setOutputValueClass(Writable.class);<a name="line.731"></a> |
| <span class="sourceLineNo">732</span> if (partitioner == HRegionPartitioner.class) {<a name="line.732"></a> |
| <span class="sourceLineNo">733</span> job.setPartitionerClass(HRegionPartitioner.class);<a name="line.733"></a> |
| <span class="sourceLineNo">734</span> int regions = getRegionCount(conf, TableName.valueOf(table));<a name="line.734"></a> |
| <span class="sourceLineNo">735</span> if (job.getNumReduceTasks() > regions) {<a name="line.735"></a> |
| <span class="sourceLineNo">736</span> job.setNumReduceTasks(regions);<a name="line.736"></a> |
| <span class="sourceLineNo">737</span> }<a name="line.737"></a> |
| <span class="sourceLineNo">738</span> } else if (partitioner != null) {<a name="line.738"></a> |
| <span class="sourceLineNo">739</span> job.setPartitionerClass(partitioner);<a name="line.739"></a> |
| <span class="sourceLineNo">740</span> }<a name="line.740"></a> |
| <span class="sourceLineNo">741</span><a name="line.741"></a> |
| <span class="sourceLineNo">742</span> if (addDependencyJars) {<a name="line.742"></a> |
| <span class="sourceLineNo">743</span> addDependencyJars(job);<a name="line.743"></a> |
| <span class="sourceLineNo">744</span> }<a name="line.744"></a> |
| <span class="sourceLineNo">745</span><a name="line.745"></a> |
| <span class="sourceLineNo">746</span> initCredentials(job);<a name="line.746"></a> |
| <span class="sourceLineNo">747</span> }<a name="line.747"></a> |
| <span class="sourceLineNo">748</span><a name="line.748"></a> |
| <span class="sourceLineNo">749</span> /**<a name="line.749"></a> |
| <span class="sourceLineNo">750</span> * Ensures that the given number of reduce tasks for the given job<a name="line.750"></a> |
| <span class="sourceLineNo">751</span> * configuration does not exceed the number of regions for the given table.<a name="line.751"></a> |
| <span class="sourceLineNo">752</span> *<a name="line.752"></a> |
| <span class="sourceLineNo">753</span> * @param table The table to get the region count for.<a name="line.753"></a> |
| <span class="sourceLineNo">754</span> * @param job The current job to adjust.<a name="line.754"></a> |
| <span class="sourceLineNo">755</span> * @throws IOException When retrieving the table details fails.<a name="line.755"></a> |
| <span class="sourceLineNo">756</span> */<a name="line.756"></a> |
| <span class="sourceLineNo">757</span> public static void limitNumReduceTasks(String table, Job job) throws IOException {<a name="line.757"></a> |
| <span class="sourceLineNo">758</span> int regions = getRegionCount(job.getConfiguration(), TableName.valueOf(table));<a name="line.758"></a> |
| <span class="sourceLineNo">759</span> if (job.getNumReduceTasks() > regions) {<a name="line.759"></a> |
| <span class="sourceLineNo">760</span> job.setNumReduceTasks(regions);<a name="line.760"></a> |
| <span class="sourceLineNo">761</span> }<a name="line.761"></a> |
| <span class="sourceLineNo">762</span> }<a name="line.762"></a> |
| <span class="sourceLineNo">763</span><a name="line.763"></a> |
| <span class="sourceLineNo">764</span> /**<a name="line.764"></a> |
| <span class="sourceLineNo">765</span> * Sets the number of reduce tasks for the given job configuration to the<a name="line.765"></a> |
| <span class="sourceLineNo">766</span> * number of regions the given table has.<a name="line.766"></a> |
| <span class="sourceLineNo">767</span> *<a name="line.767"></a> |
| <span class="sourceLineNo">768</span> * @param table The table to get the region count for.<a name="line.768"></a> |
| <span class="sourceLineNo">769</span> * @param job The current job to adjust.<a name="line.769"></a> |
| <span class="sourceLineNo">770</span> * @throws IOException When retrieving the table details fails.<a name="line.770"></a> |
| <span class="sourceLineNo">771</span> */<a name="line.771"></a> |
| <span class="sourceLineNo">772</span> public static void setNumReduceTasks(String table, Job job) throws IOException {<a name="line.772"></a> |
| <span class="sourceLineNo">773</span> job.setNumReduceTasks(getRegionCount(job.getConfiguration(), TableName.valueOf(table)));<a name="line.773"></a> |
| <span class="sourceLineNo">774</span> }<a name="line.774"></a> |
| <span class="sourceLineNo">775</span><a name="line.775"></a> |
| <span class="sourceLineNo">776</span> /**<a name="line.776"></a> |
| <span class="sourceLineNo">777</span> * Sets the number of rows to return and cache with each scanner iteration.<a name="line.777"></a> |
| <span class="sourceLineNo">778</span> * Higher caching values will enable faster mapreduce jobs at the expense of<a name="line.778"></a> |
| <span class="sourceLineNo">779</span> * requiring more heap to contain the cached rows.<a name="line.779"></a> |
| <span class="sourceLineNo">780</span> *<a name="line.780"></a> |
| <span class="sourceLineNo">781</span> * @param job The current job to adjust.<a name="line.781"></a> |
| <span class="sourceLineNo">782</span> * @param batchSize The number of rows to return in batch with each scanner<a name="line.782"></a> |
| <span class="sourceLineNo">783</span> * iteration.<a name="line.783"></a> |
| <span class="sourceLineNo">784</span> */<a name="line.784"></a> |
| <span class="sourceLineNo">785</span> public static void setScannerCaching(Job job, int batchSize) {<a name="line.785"></a> |
| <span class="sourceLineNo">786</span> job.getConfiguration().setInt("hbase.client.scanner.caching", batchSize);<a name="line.786"></a> |
| <span class="sourceLineNo">787</span> }<a name="line.787"></a> |
| <span class="sourceLineNo">788</span><a name="line.788"></a> |
| <span class="sourceLineNo">789</span> /**<a name="line.789"></a> |
| <span class="sourceLineNo">790</span> * Add HBase and its dependencies (only) to the job configuration.<a name="line.790"></a> |
| <span class="sourceLineNo">791</span> * <p><a name="line.791"></a> |
| <span class="sourceLineNo">792</span> * This is intended as a low-level API, facilitating code reuse between this<a name="line.792"></a> |
| <span class="sourceLineNo">793</span> * class and its mapred counterpart. It also of use to external tools that<a name="line.793"></a> |
| <span class="sourceLineNo">794</span> * need to build a MapReduce job that interacts with HBase but want<a name="line.794"></a> |
| <span class="sourceLineNo">795</span> * fine-grained control over the jars shipped to the cluster.<a name="line.795"></a> |
| <span class="sourceLineNo">796</span> * </p><a name="line.796"></a> |
| <span class="sourceLineNo">797</span> * @param conf The Configuration object to extend with dependencies.<a name="line.797"></a> |
| <span class="sourceLineNo">798</span> * @see org.apache.hadoop.hbase.mapred.TableMapReduceUtil<a name="line.798"></a> |
| <span class="sourceLineNo">799</span> * @see <a href="https://issues.apache.org/jira/browse/PIG-3285">PIG-3285</a><a name="line.799"></a> |
| <span class="sourceLineNo">800</span> */<a name="line.800"></a> |
| <span class="sourceLineNo">801</span> public static void addHBaseDependencyJars(Configuration conf) throws IOException {<a name="line.801"></a> |
| <span class="sourceLineNo">802</span> addDependencyJarsForClasses(conf,<a name="line.802"></a> |
| <span class="sourceLineNo">803</span> // explicitly pull a class from each module<a name="line.803"></a> |
| <span class="sourceLineNo">804</span> org.apache.hadoop.hbase.HConstants.class, // hbase-common<a name="line.804"></a> |
| <span class="sourceLineNo">805</span> org.apache.hadoop.hbase.shaded.protobuf.generated.ClientProtos.class, // hbase-protocol-shaded<a name="line.805"></a> |
| <span class="sourceLineNo">806</span> org.apache.hadoop.hbase.client.Put.class, // hbase-client<a name="line.806"></a> |
| <span class="sourceLineNo">807</span> org.apache.hadoop.hbase.ipc.RpcServer.class, // hbase-server<a name="line.807"></a> |
| <span class="sourceLineNo">808</span> org.apache.hadoop.hbase.CompatibilityFactory.class, // hbase-hadoop-compat<a name="line.808"></a> |
| <span class="sourceLineNo">809</span> org.apache.hadoop.hbase.mapreduce.JobUtil.class, // hbase-hadoop2-compat<a name="line.809"></a> |
| <span class="sourceLineNo">810</span> org.apache.hadoop.hbase.mapreduce.TableMapper.class, // hbase-mapreduce<a name="line.810"></a> |
| <span class="sourceLineNo">811</span> org.apache.hadoop.hbase.metrics.impl.FastLongHistogram.class, // hbase-metrics<a name="line.811"></a> |
| <span class="sourceLineNo">812</span> org.apache.hadoop.hbase.metrics.Snapshot.class, // hbase-metrics-api<a name="line.812"></a> |
| <span class="sourceLineNo">813</span> org.apache.hadoop.hbase.replication.ReplicationUtils.class, // hbase-replication<a name="line.813"></a> |
| <span class="sourceLineNo">814</span> org.apache.hadoop.hbase.http.HttpServer.class, // hbase-http<a name="line.814"></a> |
| <span class="sourceLineNo">815</span> org.apache.hadoop.hbase.procedure2.Procedure.class, // hbase-procedure<a name="line.815"></a> |
| <span class="sourceLineNo">816</span> org.apache.hadoop.hbase.zookeeper.ZKWatcher.class, // hbase-zookeeper<a name="line.816"></a> |
| <span class="sourceLineNo">817</span> org.apache.hbase.thirdparty.com.google.common.collect.Lists.class, // hb-shaded-miscellaneous<a name="line.817"></a> |
| <span class="sourceLineNo">818</span> org.apache.hbase.thirdparty.com.google.gson.GsonBuilder.class, // hbase-shaded-gson<a name="line.818"></a> |
| <span class="sourceLineNo">819</span> org.apache.hbase.thirdparty.com.google.protobuf.UnsafeByteOperations.class, // hb-sh-protobuf<a name="line.819"></a> |
| <span class="sourceLineNo">820</span> org.apache.hbase.thirdparty.io.netty.channel.Channel.class, // hbase-shaded-netty<a name="line.820"></a> |
| <span class="sourceLineNo">821</span> org.apache.zookeeper.ZooKeeper.class, // zookeeper<a name="line.821"></a> |
| <span class="sourceLineNo">822</span> org.apache.htrace.core.Tracer.class, // htrace<a name="line.822"></a> |
| <span class="sourceLineNo">823</span> com.codahale.metrics.MetricRegistry.class, // metrics-core<a name="line.823"></a> |
| <span class="sourceLineNo">824</span> org.apache.commons.lang3.ArrayUtils.class); // commons-lang<a name="line.824"></a> |
| <span class="sourceLineNo">825</span> }<a name="line.825"></a> |
| <span class="sourceLineNo">826</span><a name="line.826"></a> |
| <span class="sourceLineNo">827</span> /**<a name="line.827"></a> |
| <span class="sourceLineNo">828</span> * Returns a classpath string built from the content of the "tmpjars" value in {@code conf}.<a name="line.828"></a> |
| <span class="sourceLineNo">829</span> * Also exposed to shell scripts via `bin/hbase mapredcp`.<a name="line.829"></a> |
| <span class="sourceLineNo">830</span> */<a name="line.830"></a> |
| <span class="sourceLineNo">831</span> public static String buildDependencyClasspath(Configuration conf) {<a name="line.831"></a> |
| <span class="sourceLineNo">832</span> if (conf == null) {<a name="line.832"></a> |
| <span class="sourceLineNo">833</span> throw new IllegalArgumentException("Must provide a configuration object.");<a name="line.833"></a> |
| <span class="sourceLineNo">834</span> }<a name="line.834"></a> |
| <span class="sourceLineNo">835</span> Set<String> paths = new HashSet<>(conf.getStringCollection("tmpjars"));<a name="line.835"></a> |
| <span class="sourceLineNo">836</span> if (paths.isEmpty()) {<a name="line.836"></a> |
| <span class="sourceLineNo">837</span> throw new IllegalArgumentException("Configuration contains no tmpjars.");<a name="line.837"></a> |
| <span class="sourceLineNo">838</span> }<a name="line.838"></a> |
| <span class="sourceLineNo">839</span> StringBuilder sb = new StringBuilder();<a name="line.839"></a> |
| <span class="sourceLineNo">840</span> for (String s : paths) {<a name="line.840"></a> |
| <span class="sourceLineNo">841</span> // entries can take the form 'file:/path/to/file.jar'.<a name="line.841"></a> |
| <span class="sourceLineNo">842</span> int idx = s.indexOf(":");<a name="line.842"></a> |
| <span class="sourceLineNo">843</span> if (idx != -1) s = s.substring(idx + 1);<a name="line.843"></a> |
| <span class="sourceLineNo">844</span> if (sb.length() > 0) sb.append(File.pathSeparator);<a name="line.844"></a> |
| <span class="sourceLineNo">845</span> sb.append(s);<a name="line.845"></a> |
| <span class="sourceLineNo">846</span> }<a name="line.846"></a> |
| <span class="sourceLineNo">847</span> return sb.toString();<a name="line.847"></a> |
| <span class="sourceLineNo">848</span> }<a name="line.848"></a> |
| <span class="sourceLineNo">849</span><a name="line.849"></a> |
| <span class="sourceLineNo">850</span> /**<a name="line.850"></a> |
| <span class="sourceLineNo">851</span> * Add the HBase dependency jars as well as jars for any of the configured<a name="line.851"></a> |
| <span class="sourceLineNo">852</span> * job classes to the job configuration, so that JobClient will ship them<a name="line.852"></a> |
| <span class="sourceLineNo">853</span> * to the cluster and add them to the DistributedCache.<a name="line.853"></a> |
| <span class="sourceLineNo">854</span> */<a name="line.854"></a> |
| <span class="sourceLineNo">855</span> public static void addDependencyJars(Job job) throws IOException {<a name="line.855"></a> |
| <span class="sourceLineNo">856</span> addHBaseDependencyJars(job.getConfiguration());<a name="line.856"></a> |
| <span class="sourceLineNo">857</span> try {<a name="line.857"></a> |
| <span class="sourceLineNo">858</span> addDependencyJarsForClasses(job.getConfiguration(),<a name="line.858"></a> |
| <span class="sourceLineNo">859</span> // when making changes here, consider also mapred.TableMapReduceUtil<a name="line.859"></a> |
| <span class="sourceLineNo">860</span> // pull job classes<a name="line.860"></a> |
| <span class="sourceLineNo">861</span> job.getMapOutputKeyClass(),<a name="line.861"></a> |
| <span class="sourceLineNo">862</span> job.getMapOutputValueClass(),<a name="line.862"></a> |
| <span class="sourceLineNo">863</span> job.getInputFormatClass(),<a name="line.863"></a> |
| <span class="sourceLineNo">864</span> job.getOutputKeyClass(),<a name="line.864"></a> |
| <span class="sourceLineNo">865</span> job.getOutputValueClass(),<a name="line.865"></a> |
| <span class="sourceLineNo">866</span> job.getOutputFormatClass(),<a name="line.866"></a> |
| <span class="sourceLineNo">867</span> job.getPartitionerClass(),<a name="line.867"></a> |
| <span class="sourceLineNo">868</span> job.getCombinerClass());<a name="line.868"></a> |
| <span class="sourceLineNo">869</span> } catch (ClassNotFoundException e) {<a name="line.869"></a> |
| <span class="sourceLineNo">870</span> throw new IOException(e);<a name="line.870"></a> |
| <span class="sourceLineNo">871</span> }<a name="line.871"></a> |
| <span class="sourceLineNo">872</span> }<a name="line.872"></a> |
| <span class="sourceLineNo">873</span><a name="line.873"></a> |
| <span class="sourceLineNo">874</span> /**<a name="line.874"></a> |
| <span class="sourceLineNo">875</span> * Add the jars containing the given classes to the job's configuration<a name="line.875"></a> |
| <span class="sourceLineNo">876</span> * such that JobClient will ship them to the cluster and add them to<a name="line.876"></a> |
| <span class="sourceLineNo">877</span> * the DistributedCache.<a name="line.877"></a> |
| <span class="sourceLineNo">878</span> * @deprecated since 1.3.0 and will be removed in 3.0.0. Use {@link #addDependencyJars(Job)}<a name="line.878"></a> |
| <span class="sourceLineNo">879</span> * instead.<a name="line.879"></a> |
| <span class="sourceLineNo">880</span> * @see #addDependencyJars(Job)<a name="line.880"></a> |
| <span class="sourceLineNo">881</span> * @see <a href="https://issues.apache.org/jira/browse/HBASE-8386">HBASE-8386</a><a name="line.881"></a> |
| <span class="sourceLineNo">882</span> */<a name="line.882"></a> |
| <span class="sourceLineNo">883</span> @Deprecated<a name="line.883"></a> |
| <span class="sourceLineNo">884</span> public static void addDependencyJars(Configuration conf,<a name="line.884"></a> |
| <span class="sourceLineNo">885</span> Class<?>... classes) throws IOException {<a name="line.885"></a> |
| <span class="sourceLineNo">886</span> LOG.warn("The addDependencyJars(Configuration, Class<?>...) method has been deprecated since it"<a name="line.886"></a> |
| <span class="sourceLineNo">887</span> + " is easy to use incorrectly. Most users should rely on addDependencyJars(Job) " +<a name="line.887"></a> |
| <span class="sourceLineNo">888</span> "instead. See HBASE-8386 for more details.");<a name="line.888"></a> |
| <span class="sourceLineNo">889</span> addDependencyJarsForClasses(conf, classes);<a name="line.889"></a> |
| <span class="sourceLineNo">890</span> }<a name="line.890"></a> |
| <span class="sourceLineNo">891</span><a name="line.891"></a> |
| <span class="sourceLineNo">892</span> /**<a name="line.892"></a> |
| <span class="sourceLineNo">893</span> * Add the jars containing the given classes to the job's configuration<a name="line.893"></a> |
| <span class="sourceLineNo">894</span> * such that JobClient will ship them to the cluster and add them to<a name="line.894"></a> |
| <span class="sourceLineNo">895</span> * the DistributedCache.<a name="line.895"></a> |
| <span class="sourceLineNo">896</span> *<a name="line.896"></a> |
| <span class="sourceLineNo">897</span> * N.B. that this method at most adds one jar per class given. If there is more than one<a name="line.897"></a> |
| <span class="sourceLineNo">898</span> * jar available containing a class with the same name as a given class, we don't define<a name="line.898"></a> |
| <span class="sourceLineNo">899</span> * which of those jars might be chosen.<a name="line.899"></a> |
| <span class="sourceLineNo">900</span> *<a name="line.900"></a> |
| <span class="sourceLineNo">901</span> * @param conf The Hadoop Configuration to modify<a name="line.901"></a> |
| <span class="sourceLineNo">902</span> * @param classes will add just those dependencies needed to find the given classes<a name="line.902"></a> |
| <span class="sourceLineNo">903</span> * @throws IOException if an underlying library call fails.<a name="line.903"></a> |
| <span class="sourceLineNo">904</span> */<a name="line.904"></a> |
| <span class="sourceLineNo">905</span> @InterfaceAudience.Private<a name="line.905"></a> |
| <span class="sourceLineNo">906</span> public static void addDependencyJarsForClasses(Configuration conf,<a name="line.906"></a> |
| <span class="sourceLineNo">907</span> Class<?>... classes) throws IOException {<a name="line.907"></a> |
| <span class="sourceLineNo">908</span><a name="line.908"></a> |
| <span class="sourceLineNo">909</span> FileSystem localFs = FileSystem.getLocal(conf);<a name="line.909"></a> |
| <span class="sourceLineNo">910</span> Set<String> jars = new HashSet<>();<a name="line.910"></a> |
| <span class="sourceLineNo">911</span> // Add jars that are already in the tmpjars variable<a name="line.911"></a> |
| <span class="sourceLineNo">912</span> jars.addAll(conf.getStringCollection("tmpjars"));<a name="line.912"></a> |
| <span class="sourceLineNo">913</span><a name="line.913"></a> |
| <span class="sourceLineNo">914</span> // add jars as we find them to a map of contents jar name so that we can avoid<a name="line.914"></a> |
| <span class="sourceLineNo">915</span> // creating new jars for classes that have already been packaged.<a name="line.915"></a> |
| <span class="sourceLineNo">916</span> Map<String, String> packagedClasses = new HashMap<>();<a name="line.916"></a> |
| <span class="sourceLineNo">917</span><a name="line.917"></a> |
| <span class="sourceLineNo">918</span> // Add jars containing the specified classes<a name="line.918"></a> |
| <span class="sourceLineNo">919</span> for (Class<?> clazz : classes) {<a name="line.919"></a> |
| <span class="sourceLineNo">920</span> if (clazz == null) continue;<a name="line.920"></a> |
| <span class="sourceLineNo">921</span><a name="line.921"></a> |
| <span class="sourceLineNo">922</span> Path path = findOrCreateJar(clazz, localFs, packagedClasses);<a name="line.922"></a> |
| <span class="sourceLineNo">923</span> if (path == null) {<a name="line.923"></a> |
| <span class="sourceLineNo">924</span> LOG.warn("Could not find jar for class " + clazz +<a name="line.924"></a> |
| <span class="sourceLineNo">925</span> " in order to ship it to the cluster.");<a name="line.925"></a> |
| <span class="sourceLineNo">926</span> continue;<a name="line.926"></a> |
| <span class="sourceLineNo">927</span> }<a name="line.927"></a> |
| <span class="sourceLineNo">928</span> if (!localFs.exists(path)) {<a name="line.928"></a> |
| <span class="sourceLineNo">929</span> LOG.warn("Could not validate jar file " + path + " for class "<a name="line.929"></a> |
| <span class="sourceLineNo">930</span> + clazz);<a name="line.930"></a> |
| <span class="sourceLineNo">931</span> continue;<a name="line.931"></a> |
| <span class="sourceLineNo">932</span> }<a name="line.932"></a> |
| <span class="sourceLineNo">933</span> jars.add(path.toString());<a name="line.933"></a> |
| <span class="sourceLineNo">934</span> }<a name="line.934"></a> |
| <span class="sourceLineNo">935</span> if (jars.isEmpty()) return;<a name="line.935"></a> |
| <span class="sourceLineNo">936</span><a name="line.936"></a> |
| <span class="sourceLineNo">937</span> conf.set("tmpjars", StringUtils.arrayToString(jars.toArray(new String[jars.size()])));<a name="line.937"></a> |
| <span class="sourceLineNo">938</span> }<a name="line.938"></a> |
| <span class="sourceLineNo">939</span><a name="line.939"></a> |
| <span class="sourceLineNo">940</span> /**<a name="line.940"></a> |
| <span class="sourceLineNo">941</span> * Finds the Jar for a class or creates it if it doesn't exist. If the class is in<a name="line.941"></a> |
| <span class="sourceLineNo">942</span> * a directory in the classpath, it creates a Jar on the fly with the<a name="line.942"></a> |
| <span class="sourceLineNo">943</span> * contents of the directory and returns the path to that Jar. If a Jar is<a name="line.943"></a> |
| <span class="sourceLineNo">944</span> * created, it is created in the system temporary directory. Otherwise,<a name="line.944"></a> |
| <span class="sourceLineNo">945</span> * returns an existing jar that contains a class of the same name. Maintains<a name="line.945"></a> |
| <span class="sourceLineNo">946</span> * a mapping from jar contents to the tmp jar created.<a name="line.946"></a> |
| <span class="sourceLineNo">947</span> * @param my_class the class to find.<a name="line.947"></a> |
| <span class="sourceLineNo">948</span> * @param fs the FileSystem with which to qualify the returned path.<a name="line.948"></a> |
| <span class="sourceLineNo">949</span> * @param packagedClasses a map of class name to path.<a name="line.949"></a> |
| <span class="sourceLineNo">950</span> * @return a jar file that contains the class.<a name="line.950"></a> |
| <span class="sourceLineNo">951</span> * @throws IOException<a name="line.951"></a> |
| <span class="sourceLineNo">952</span> */<a name="line.952"></a> |
| <span class="sourceLineNo">953</span> private static Path findOrCreateJar(Class<?> my_class, FileSystem fs,<a name="line.953"></a> |
| <span class="sourceLineNo">954</span> Map<String, String> packagedClasses)<a name="line.954"></a> |
| <span class="sourceLineNo">955</span> throws IOException {<a name="line.955"></a> |
| <span class="sourceLineNo">956</span> // attempt to locate an existing jar for the class.<a name="line.956"></a> |
| <span class="sourceLineNo">957</span> String jar = findContainingJar(my_class, packagedClasses);<a name="line.957"></a> |
| <span class="sourceLineNo">958</span> if (null == jar || jar.isEmpty()) {<a name="line.958"></a> |
| <span class="sourceLineNo">959</span> jar = getJar(my_class);<a name="line.959"></a> |
| <span class="sourceLineNo">960</span> updateMap(jar, packagedClasses);<a name="line.960"></a> |
| <span class="sourceLineNo">961</span> }<a name="line.961"></a> |
| <span class="sourceLineNo">962</span><a name="line.962"></a> |
| <span class="sourceLineNo">963</span> if (null == jar || jar.isEmpty()) {<a name="line.963"></a> |
| <span class="sourceLineNo">964</span> return null;<a name="line.964"></a> |
| <span class="sourceLineNo">965</span> }<a name="line.965"></a> |
| <span class="sourceLineNo">966</span><a name="line.966"></a> |
| <span class="sourceLineNo">967</span> LOG.debug(String.format("For class %s, using jar %s", my_class.getName(), jar));<a name="line.967"></a> |
| <span class="sourceLineNo">968</span> return new Path(jar).makeQualified(fs.getUri(), fs.getWorkingDirectory());<a name="line.968"></a> |
| <span class="sourceLineNo">969</span> }<a name="line.969"></a> |
| <span class="sourceLineNo">970</span><a name="line.970"></a> |
| <span class="sourceLineNo">971</span> /**<a name="line.971"></a> |
| <span class="sourceLineNo">972</span> * Add entries to <code>packagedClasses</code> corresponding to class files<a name="line.972"></a> |
| <span class="sourceLineNo">973</span> * contained in <code>jar</code>.<a name="line.973"></a> |
| <span class="sourceLineNo">974</span> * @param jar The jar who's content to list.<a name="line.974"></a> |
| <span class="sourceLineNo">975</span> * @param packagedClasses map[class -> jar]<a name="line.975"></a> |
| <span class="sourceLineNo">976</span> */<a name="line.976"></a> |
| <span class="sourceLineNo">977</span> private static void updateMap(String jar, Map<String, String> packagedClasses) throws IOException {<a name="line.977"></a> |
| <span class="sourceLineNo">978</span> if (null == jar || jar.isEmpty()) {<a name="line.978"></a> |
| <span class="sourceLineNo">979</span> return;<a name="line.979"></a> |
| <span class="sourceLineNo">980</span> }<a name="line.980"></a> |
| <span class="sourceLineNo">981</span> ZipFile zip = null;<a name="line.981"></a> |
| <span class="sourceLineNo">982</span> try {<a name="line.982"></a> |
| <span class="sourceLineNo">983</span> zip = new ZipFile(jar);<a name="line.983"></a> |
| <span class="sourceLineNo">984</span> for (Enumeration<? extends ZipEntry> iter = zip.entries(); iter.hasMoreElements();) {<a name="line.984"></a> |
| <span class="sourceLineNo">985</span> ZipEntry entry = iter.nextElement();<a name="line.985"></a> |
| <span class="sourceLineNo">986</span> if (entry.getName().endsWith("class")) {<a name="line.986"></a> |
| <span class="sourceLineNo">987</span> packagedClasses.put(entry.getName(), jar);<a name="line.987"></a> |
| <span class="sourceLineNo">988</span> }<a name="line.988"></a> |
| <span class="sourceLineNo">989</span> }<a name="line.989"></a> |
| <span class="sourceLineNo">990</span> } finally {<a name="line.990"></a> |
| <span class="sourceLineNo">991</span> if (null != zip) zip.close();<a name="line.991"></a> |
| <span class="sourceLineNo">992</span> }<a name="line.992"></a> |
| <span class="sourceLineNo">993</span> }<a name="line.993"></a> |
| <span class="sourceLineNo">994</span><a name="line.994"></a> |
| <span class="sourceLineNo">995</span> /**<a name="line.995"></a> |
| <span class="sourceLineNo">996</span> * Find a jar that contains a class of the same name, if any. It will return<a name="line.996"></a> |
| <span class="sourceLineNo">997</span> * a jar file, even if that is not the first thing on the class path that<a name="line.997"></a> |
| <span class="sourceLineNo">998</span> * has a class with the same name. Looks first on the classpath and then in<a name="line.998"></a> |
| <span class="sourceLineNo">999</span> * the <code>packagedClasses</code> map.<a name="line.999"></a> |
| <span class="sourceLineNo">1000</span> * @param my_class the class to find.<a name="line.1000"></a> |
| <span class="sourceLineNo">1001</span> * @return a jar file that contains the class, or null.<a name="line.1001"></a> |
| <span class="sourceLineNo">1002</span> * @throws IOException<a name="line.1002"></a> |
| <span class="sourceLineNo">1003</span> */<a name="line.1003"></a> |
| <span class="sourceLineNo">1004</span> private static String findContainingJar(Class<?> my_class, Map<String, String> packagedClasses)<a name="line.1004"></a> |
| <span class="sourceLineNo">1005</span> throws IOException {<a name="line.1005"></a> |
| <span class="sourceLineNo">1006</span> ClassLoader loader = my_class.getClassLoader();<a name="line.1006"></a> |
| <span class="sourceLineNo">1007</span><a name="line.1007"></a> |
| <span class="sourceLineNo">1008</span> String class_file = my_class.getName().replaceAll("\\.", "/") + ".class";<a name="line.1008"></a> |
| <span class="sourceLineNo">1009</span><a name="line.1009"></a> |
| <span class="sourceLineNo">1010</span> if (loader != null) {<a name="line.1010"></a> |
| <span class="sourceLineNo">1011</span> // first search the classpath<a name="line.1011"></a> |
| <span class="sourceLineNo">1012</span> for (Enumeration<URL> itr = loader.getResources(class_file); itr.hasMoreElements();) {<a name="line.1012"></a> |
| <span class="sourceLineNo">1013</span> URL url = itr.nextElement();<a name="line.1013"></a> |
| <span class="sourceLineNo">1014</span> if ("jar".equals(url.getProtocol())) {<a name="line.1014"></a> |
| <span class="sourceLineNo">1015</span> String toReturn = url.getPath();<a name="line.1015"></a> |
| <span class="sourceLineNo">1016</span> if (toReturn.startsWith("file:")) {<a name="line.1016"></a> |
| <span class="sourceLineNo">1017</span> toReturn = toReturn.substring("file:".length());<a name="line.1017"></a> |
| <span class="sourceLineNo">1018</span> }<a name="line.1018"></a> |
| <span class="sourceLineNo">1019</span> // URLDecoder is a misnamed class, since it actually decodes<a name="line.1019"></a> |
| <span class="sourceLineNo">1020</span> // x-www-form-urlencoded MIME type rather than actual<a name="line.1020"></a> |
| <span class="sourceLineNo">1021</span> // URL encoding (which the file path has). Therefore it would<a name="line.1021"></a> |
| <span class="sourceLineNo">1022</span> // decode +s to ' 's which is incorrect (spaces are actually<a name="line.1022"></a> |
| <span class="sourceLineNo">1023</span> // either unencoded or encoded as "%20"). Replace +s first, so<a name="line.1023"></a> |
| <span class="sourceLineNo">1024</span> // that they are kept sacred during the decoding process.<a name="line.1024"></a> |
| <span class="sourceLineNo">1025</span> toReturn = toReturn.replaceAll("\\+", "%2B");<a name="line.1025"></a> |
| <span class="sourceLineNo">1026</span> toReturn = URLDecoder.decode(toReturn, "UTF-8");<a name="line.1026"></a> |
| <span class="sourceLineNo">1027</span> return toReturn.replaceAll("!.*$", "");<a name="line.1027"></a> |
| <span class="sourceLineNo">1028</span> }<a name="line.1028"></a> |
| <span class="sourceLineNo">1029</span> }<a name="line.1029"></a> |
| <span class="sourceLineNo">1030</span> }<a name="line.1030"></a> |
| <span class="sourceLineNo">1031</span><a name="line.1031"></a> |
| <span class="sourceLineNo">1032</span> // now look in any jars we've packaged using JarFinder. Returns null when<a name="line.1032"></a> |
| <span class="sourceLineNo">1033</span> // no jar is found.<a name="line.1033"></a> |
| <span class="sourceLineNo">1034</span> return packagedClasses.get(class_file);<a name="line.1034"></a> |
| <span class="sourceLineNo">1035</span> }<a name="line.1035"></a> |
| <span class="sourceLineNo">1036</span><a name="line.1036"></a> |
| <span class="sourceLineNo">1037</span> /**<a name="line.1037"></a> |
| <span class="sourceLineNo">1038</span> * Invoke 'getJar' on a custom JarFinder implementation. Useful for some job<a name="line.1038"></a> |
| <span class="sourceLineNo">1039</span> * configuration contexts (HBASE-8140) and also for testing on MRv2.<a name="line.1039"></a> |
| <span class="sourceLineNo">1040</span> * check if we have HADOOP-9426.<a name="line.1040"></a> |
| <span class="sourceLineNo">1041</span> * @param my_class the class to find.<a name="line.1041"></a> |
| <span class="sourceLineNo">1042</span> * @return a jar file that contains the class, or null.<a name="line.1042"></a> |
| <span class="sourceLineNo">1043</span> */<a name="line.1043"></a> |
| <span class="sourceLineNo">1044</span> private static String getJar(Class<?> my_class) {<a name="line.1044"></a> |
| <span class="sourceLineNo">1045</span> String ret = null;<a name="line.1045"></a> |
| <span class="sourceLineNo">1046</span> try {<a name="line.1046"></a> |
| <span class="sourceLineNo">1047</span> ret = JarFinder.getJar(my_class);<a name="line.1047"></a> |
| <span class="sourceLineNo">1048</span> } catch (Exception e) {<a name="line.1048"></a> |
| <span class="sourceLineNo">1049</span> // toss all other exceptions, related to reflection failure<a name="line.1049"></a> |
| <span class="sourceLineNo">1050</span> throw new RuntimeException("getJar invocation failed.", e);<a name="line.1050"></a> |
| <span class="sourceLineNo">1051</span> }<a name="line.1051"></a> |
| <span class="sourceLineNo">1052</span><a name="line.1052"></a> |
| <span class="sourceLineNo">1053</span> return ret;<a name="line.1053"></a> |
| <span class="sourceLineNo">1054</span> }<a name="line.1054"></a> |
| <span class="sourceLineNo">1055</span><a name="line.1055"></a> |
| <span class="sourceLineNo">1056</span> private static int getRegionCount(Configuration conf, TableName tableName) throws IOException {<a name="line.1056"></a> |
| <span class="sourceLineNo">1057</span> try (Connection conn = ConnectionFactory.createConnection(conf);<a name="line.1057"></a> |
| <span class="sourceLineNo">1058</span> RegionLocator locator = conn.getRegionLocator(tableName)) {<a name="line.1058"></a> |
| <span class="sourceLineNo">1059</span> return locator.getAllRegionLocations().size();<a name="line.1059"></a> |
| <span class="sourceLineNo">1060</span> }<a name="line.1060"></a> |
| <span class="sourceLineNo">1061</span> }<a name="line.1061"></a> |
| <span class="sourceLineNo">1062</span>}<a name="line.1062"></a> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| </pre> |
| </div> |
| </body> |
| </html> |
