blob: c641613b9d7e6c610960490359a53efdc03b930e [file] [log] [blame]
<!DOCTYPE HTML>
<html lang="en">
<head>
<!-- Generated by javadoc (17) -->
<title>Source code</title>
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta name="description" content="source: package: org.apache.hadoop.hbase.backup.mapreduce, class: MapReduceHFileSplitterJob, class: HFileCellMapper">
<meta name="generator" content="javadoc/SourceToHTMLConverter">
<link rel="stylesheet" type="text/css" href="../../../../../../../stylesheet.css" title="Style">
</head>
<body class="source-page">
<main role="main">
<div class="source-container">
<pre><span class="source-line-no">001</span><span id="line-1">/*</span>
<span class="source-line-no">002</span><span id="line-2"> * Licensed to the Apache Software Foundation (ASF) under one</span>
<span class="source-line-no">003</span><span id="line-3"> * or more contributor license agreements. See the NOTICE file</span>
<span class="source-line-no">004</span><span id="line-4"> * distributed with this work for additional information</span>
<span class="source-line-no">005</span><span id="line-5"> * regarding copyright ownership. The ASF licenses this file</span>
<span class="source-line-no">006</span><span id="line-6"> * to you under the Apache License, Version 2.0 (the</span>
<span class="source-line-no">007</span><span id="line-7"> * "License"); you may not use this file except in compliance</span>
<span class="source-line-no">008</span><span id="line-8"> * with the License. You may obtain a copy of the License at</span>
<span class="source-line-no">009</span><span id="line-9"> *</span>
<span class="source-line-no">010</span><span id="line-10"> * http://www.apache.org/licenses/LICENSE-2.0</span>
<span class="source-line-no">011</span><span id="line-11"> *</span>
<span class="source-line-no">012</span><span id="line-12"> * Unless required by applicable law or agreed to in writing, software</span>
<span class="source-line-no">013</span><span id="line-13"> * distributed under the License is distributed on an "AS IS" BASIS,</span>
<span class="source-line-no">014</span><span id="line-14"> * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.</span>
<span class="source-line-no">015</span><span id="line-15"> * See the License for the specific language governing permissions and</span>
<span class="source-line-no">016</span><span id="line-16"> * limitations under the License.</span>
<span class="source-line-no">017</span><span id="line-17"> */</span>
<span class="source-line-no">018</span><span id="line-18">package org.apache.hadoop.hbase.backup.mapreduce;</span>
<span class="source-line-no">019</span><span id="line-19"></span>
<span class="source-line-no">020</span><span id="line-20">import java.io.IOException;</span>
<span class="source-line-no">021</span><span id="line-21">import org.apache.hadoop.conf.Configuration;</span>
<span class="source-line-no">022</span><span id="line-22">import org.apache.hadoop.conf.Configured;</span>
<span class="source-line-no">023</span><span id="line-23">import org.apache.hadoop.fs.Path;</span>
<span class="source-line-no">024</span><span id="line-24">import org.apache.hadoop.hbase.Cell;</span>
<span class="source-line-no">025</span><span id="line-25">import org.apache.hadoop.hbase.CellUtil;</span>
<span class="source-line-no">026</span><span id="line-26">import org.apache.hadoop.hbase.ExtendedCell;</span>
<span class="source-line-no">027</span><span id="line-27">import org.apache.hadoop.hbase.HBaseConfiguration;</span>
<span class="source-line-no">028</span><span id="line-28">import org.apache.hadoop.hbase.PrivateCellUtil;</span>
<span class="source-line-no">029</span><span id="line-29">import org.apache.hadoop.hbase.TableName;</span>
<span class="source-line-no">030</span><span id="line-30">import org.apache.hadoop.hbase.client.Connection;</span>
<span class="source-line-no">031</span><span id="line-31">import org.apache.hadoop.hbase.client.ConnectionFactory;</span>
<span class="source-line-no">032</span><span id="line-32">import org.apache.hadoop.hbase.client.RegionLocator;</span>
<span class="source-line-no">033</span><span id="line-33">import org.apache.hadoop.hbase.client.Table;</span>
<span class="source-line-no">034</span><span id="line-34">import org.apache.hadoop.hbase.io.ImmutableBytesWritable;</span>
<span class="source-line-no">035</span><span id="line-35">import org.apache.hadoop.hbase.mapreduce.CellSortReducer;</span>
<span class="source-line-no">036</span><span id="line-36">import org.apache.hadoop.hbase.mapreduce.HFileInputFormat;</span>
<span class="source-line-no">037</span><span id="line-37">import org.apache.hadoop.hbase.mapreduce.HFileOutputFormat2;</span>
<span class="source-line-no">038</span><span id="line-38">import org.apache.hadoop.hbase.mapreduce.KeyOnlyCellComparable;</span>
<span class="source-line-no">039</span><span id="line-39">import org.apache.hadoop.hbase.mapreduce.PreSortedCellsReducer;</span>
<span class="source-line-no">040</span><span id="line-40">import org.apache.hadoop.hbase.mapreduce.TableMapReduceUtil;</span>
<span class="source-line-no">041</span><span id="line-41">import org.apache.hadoop.hbase.snapshot.SnapshotRegionLocator;</span>
<span class="source-line-no">042</span><span id="line-42">import org.apache.hadoop.hbase.util.EnvironmentEdgeManager;</span>
<span class="source-line-no">043</span><span id="line-43">import org.apache.hadoop.hbase.util.MapReduceExtendedCell;</span>
<span class="source-line-no">044</span><span id="line-44">import org.apache.hadoop.io.NullWritable;</span>
<span class="source-line-no">045</span><span id="line-45">import org.apache.hadoop.io.WritableComparable;</span>
<span class="source-line-no">046</span><span id="line-46">import org.apache.hadoop.mapreduce.Job;</span>
<span class="source-line-no">047</span><span id="line-47">import org.apache.hadoop.mapreduce.Mapper;</span>
<span class="source-line-no">048</span><span id="line-48">import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;</span>
<span class="source-line-no">049</span><span id="line-49">import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;</span>
<span class="source-line-no">050</span><span id="line-50">import org.apache.hadoop.util.Tool;</span>
<span class="source-line-no">051</span><span id="line-51">import org.apache.hadoop.util.ToolRunner;</span>
<span class="source-line-no">052</span><span id="line-52">import org.apache.yetus.audience.InterfaceAudience;</span>
<span class="source-line-no">053</span><span id="line-53">import org.slf4j.Logger;</span>
<span class="source-line-no">054</span><span id="line-54">import org.slf4j.LoggerFactory;</span>
<span class="source-line-no">055</span><span id="line-55"></span>
<span class="source-line-no">056</span><span id="line-56">/**</span>
<span class="source-line-no">057</span><span id="line-57"> * A tool to split HFiles into new region boundaries as a MapReduce job. The tool generates HFiles</span>
<span class="source-line-no">058</span><span id="line-58"> * for later bulk importing.</span>
<span class="source-line-no">059</span><span id="line-59"> */</span>
<span class="source-line-no">060</span><span id="line-60">@InterfaceAudience.Private</span>
<span class="source-line-no">061</span><span id="line-61">public class MapReduceHFileSplitterJob extends Configured implements Tool {</span>
<span class="source-line-no">062</span><span id="line-62"> private static final Logger LOG = LoggerFactory.getLogger(MapReduceHFileSplitterJob.class);</span>
<span class="source-line-no">063</span><span id="line-63"> final static String NAME = "HFileSplitterJob";</span>
<span class="source-line-no">064</span><span id="line-64"> public final static String BULK_OUTPUT_CONF_KEY = "hfile.bulk.output";</span>
<span class="source-line-no">065</span><span id="line-65"> public final static String TABLES_KEY = "hfile.input.tables";</span>
<span class="source-line-no">066</span><span id="line-66"> public final static String TABLE_MAP_KEY = "hfile.input.tablesmap";</span>
<span class="source-line-no">067</span><span id="line-67"> private final static String JOB_NAME_CONF_KEY = "mapreduce.job.name";</span>
<span class="source-line-no">068</span><span id="line-68"></span>
<span class="source-line-no">069</span><span id="line-69"> public MapReduceHFileSplitterJob() {</span>
<span class="source-line-no">070</span><span id="line-70"> }</span>
<span class="source-line-no">071</span><span id="line-71"></span>
<span class="source-line-no">072</span><span id="line-72"> protected MapReduceHFileSplitterJob(final Configuration c) {</span>
<span class="source-line-no">073</span><span id="line-73"> super(c);</span>
<span class="source-line-no">074</span><span id="line-74"> }</span>
<span class="source-line-no">075</span><span id="line-75"></span>
<span class="source-line-no">076</span><span id="line-76"> /**</span>
<span class="source-line-no">077</span><span id="line-77"> * A mapper that just writes out cells. This one can be used together with {@link CellSortReducer}</span>
<span class="source-line-no">078</span><span id="line-78"> */</span>
<span class="source-line-no">079</span><span id="line-79"> static class HFileCellMapper extends Mapper&lt;NullWritable, Cell, WritableComparable&lt;?&gt;, Cell&gt; {</span>
<span class="source-line-no">080</span><span id="line-80"></span>
<span class="source-line-no">081</span><span id="line-81"> private boolean diskBasedSortingEnabled = false;</span>
<span class="source-line-no">082</span><span id="line-82"></span>
<span class="source-line-no">083</span><span id="line-83"> @Override</span>
<span class="source-line-no">084</span><span id="line-84"> public void map(NullWritable key, Cell value, Context context)</span>
<span class="source-line-no">085</span><span id="line-85"> throws IOException, InterruptedException {</span>
<span class="source-line-no">086</span><span id="line-86"> ExtendedCell extendedCell = PrivateCellUtil.ensureExtendedCell(value);</span>
<span class="source-line-no">087</span><span id="line-87"> context.write(wrap(extendedCell), new MapReduceExtendedCell(extendedCell));</span>
<span class="source-line-no">088</span><span id="line-88"> }</span>
<span class="source-line-no">089</span><span id="line-89"></span>
<span class="source-line-no">090</span><span id="line-90"> @Override</span>
<span class="source-line-no">091</span><span id="line-91"> public void setup(Context context) throws IOException {</span>
<span class="source-line-no">092</span><span id="line-92"> diskBasedSortingEnabled =</span>
<span class="source-line-no">093</span><span id="line-93"> HFileOutputFormat2.diskBasedSortingEnabled(context.getConfiguration());</span>
<span class="source-line-no">094</span><span id="line-94"> }</span>
<span class="source-line-no">095</span><span id="line-95"></span>
<span class="source-line-no">096</span><span id="line-96"> private WritableComparable&lt;?&gt; wrap(ExtendedCell cell) {</span>
<span class="source-line-no">097</span><span id="line-97"> if (diskBasedSortingEnabled) {</span>
<span class="source-line-no">098</span><span id="line-98"> return new KeyOnlyCellComparable(cell);</span>
<span class="source-line-no">099</span><span id="line-99"> }</span>
<span class="source-line-no">100</span><span id="line-100"> return new ImmutableBytesWritable(CellUtil.cloneRow(cell));</span>
<span class="source-line-no">101</span><span id="line-101"> }</span>
<span class="source-line-no">102</span><span id="line-102"> }</span>
<span class="source-line-no">103</span><span id="line-103"></span>
<span class="source-line-no">104</span><span id="line-104"> /**</span>
<span class="source-line-no">105</span><span id="line-105"> * Sets up the actual job.</span>
<span class="source-line-no">106</span><span id="line-106"> * @param args The command line parameters.</span>
<span class="source-line-no">107</span><span id="line-107"> * @return The newly created job.</span>
<span class="source-line-no">108</span><span id="line-108"> * @throws IOException When setting up the job fails.</span>
<span class="source-line-no">109</span><span id="line-109"> */</span>
<span class="source-line-no">110</span><span id="line-110"> public Job createSubmittableJob(String[] args) throws IOException {</span>
<span class="source-line-no">111</span><span id="line-111"> Configuration conf = getConf();</span>
<span class="source-line-no">112</span><span id="line-112"> String inputDirs = args[0];</span>
<span class="source-line-no">113</span><span id="line-113"> String tabName = args[1];</span>
<span class="source-line-no">114</span><span id="line-114"> conf.setStrings(TABLES_KEY, tabName);</span>
<span class="source-line-no">115</span><span id="line-115"> conf.set(FileInputFormat.INPUT_DIR, inputDirs);</span>
<span class="source-line-no">116</span><span id="line-116"> Job job = Job.getInstance(conf,</span>
<span class="source-line-no">117</span><span id="line-117"> conf.get(JOB_NAME_CONF_KEY, NAME + "_" + EnvironmentEdgeManager.currentTime()));</span>
<span class="source-line-no">118</span><span id="line-118"> // MapReduceHFileSplitter needs ExtendedCellSerialization so that sequenceId can be propagated</span>
<span class="source-line-no">119</span><span id="line-119"> // when sorting cells in CellSortReducer</span>
<span class="source-line-no">120</span><span id="line-120"> job.getConfiguration().setBoolean(HFileOutputFormat2.EXTENDED_CELL_SERIALIZATION_ENABLED_KEY,</span>
<span class="source-line-no">121</span><span id="line-121"> true);</span>
<span class="source-line-no">122</span><span id="line-122"> job.setJarByClass(MapReduceHFileSplitterJob.class);</span>
<span class="source-line-no">123</span><span id="line-123"> job.setInputFormatClass(HFileInputFormat.class);</span>
<span class="source-line-no">124</span><span id="line-124"> String hfileOutPath = conf.get(BULK_OUTPUT_CONF_KEY);</span>
<span class="source-line-no">125</span><span id="line-125"> boolean diskBasedSortingEnabled = HFileOutputFormat2.diskBasedSortingEnabled(conf);</span>
<span class="source-line-no">126</span><span id="line-126"> if (diskBasedSortingEnabled) {</span>
<span class="source-line-no">127</span><span id="line-127"> job.setMapOutputKeyClass(KeyOnlyCellComparable.class);</span>
<span class="source-line-no">128</span><span id="line-128"> job.setSortComparatorClass(KeyOnlyCellComparable.KeyOnlyCellComparator.class);</span>
<span class="source-line-no">129</span><span id="line-129"> } else {</span>
<span class="source-line-no">130</span><span id="line-130"> job.setMapOutputKeyClass(ImmutableBytesWritable.class);</span>
<span class="source-line-no">131</span><span id="line-131"> }</span>
<span class="source-line-no">132</span><span id="line-132"> if (hfileOutPath != null) {</span>
<span class="source-line-no">133</span><span id="line-133"> LOG.debug("add incremental job :" + hfileOutPath + " from " + inputDirs);</span>
<span class="source-line-no">134</span><span id="line-134"> TableName tableName = TableName.valueOf(tabName);</span>
<span class="source-line-no">135</span><span id="line-135"> job.setMapperClass(HFileCellMapper.class);</span>
<span class="source-line-no">136</span><span id="line-136"> if (diskBasedSortingEnabled) {</span>
<span class="source-line-no">137</span><span id="line-137"> job.setReducerClass(PreSortedCellsReducer.class);</span>
<span class="source-line-no">138</span><span id="line-138"> } else {</span>
<span class="source-line-no">139</span><span id="line-139"> job.setReducerClass(CellSortReducer.class);</span>
<span class="source-line-no">140</span><span id="line-140"> }</span>
<span class="source-line-no">141</span><span id="line-141"> Path outputDir = new Path(hfileOutPath);</span>
<span class="source-line-no">142</span><span id="line-142"> FileOutputFormat.setOutputPath(job, outputDir);</span>
<span class="source-line-no">143</span><span id="line-143"> job.setMapOutputValueClass(MapReduceExtendedCell.class);</span>
<span class="source-line-no">144</span><span id="line-144"> try (Connection conn = ConnectionFactory.createConnection(conf);</span>
<span class="source-line-no">145</span><span id="line-145"> Table table = conn.getTable(tableName);</span>
<span class="source-line-no">146</span><span id="line-146"> RegionLocator regionLocator = getRegionLocator(conf, conn, tableName)) {</span>
<span class="source-line-no">147</span><span id="line-147"> HFileOutputFormat2.configureIncrementalLoad(job, table.getDescriptor(), regionLocator);</span>
<span class="source-line-no">148</span><span id="line-148"> }</span>
<span class="source-line-no">149</span><span id="line-149"> LOG.debug("success configuring load incremental job");</span>
<span class="source-line-no">150</span><span id="line-150"></span>
<span class="source-line-no">151</span><span id="line-151"> TableMapReduceUtil.addDependencyJarsForClasses(job.getConfiguration(),</span>
<span class="source-line-no">152</span><span id="line-152"> org.apache.hbase.thirdparty.com.google.common.base.Preconditions.class);</span>
<span class="source-line-no">153</span><span id="line-153"> } else {</span>
<span class="source-line-no">154</span><span id="line-154"> throw new IOException("No bulk output directory specified");</span>
<span class="source-line-no">155</span><span id="line-155"> }</span>
<span class="source-line-no">156</span><span id="line-156"> return job;</span>
<span class="source-line-no">157</span><span id="line-157"> }</span>
<span class="source-line-no">158</span><span id="line-158"></span>
<span class="source-line-no">159</span><span id="line-159"> /**</span>
<span class="source-line-no">160</span><span id="line-160"> * Print usage</span>
<span class="source-line-no">161</span><span id="line-161"> * @param errorMsg Error message. Can be null.</span>
<span class="source-line-no">162</span><span id="line-162"> */</span>
<span class="source-line-no">163</span><span id="line-163"> private void usage(final String errorMsg) {</span>
<span class="source-line-no">164</span><span id="line-164"> if (errorMsg != null &amp;&amp; errorMsg.length() &gt; 0) {</span>
<span class="source-line-no">165</span><span id="line-165"> System.err.println("ERROR: " + errorMsg);</span>
<span class="source-line-no">166</span><span id="line-166"> }</span>
<span class="source-line-no">167</span><span id="line-167"> System.err.println("Usage: " + NAME + " [options] &lt;HFile inputdir(s)&gt; &lt;table&gt;");</span>
<span class="source-line-no">168</span><span id="line-168"> System.err.println("Read all HFile's for &lt;table&gt; and split them to &lt;table&gt; region boundaries.");</span>
<span class="source-line-no">169</span><span id="line-169"> System.err.println("&lt;table&gt; table to load.\n");</span>
<span class="source-line-no">170</span><span id="line-170"> System.err.println("To generate HFiles for a bulk data load, pass the option:");</span>
<span class="source-line-no">171</span><span id="line-171"> System.err.println(" -D" + BULK_OUTPUT_CONF_KEY + "=/path/for/output");</span>
<span class="source-line-no">172</span><span id="line-172"> System.err.println("Other options:");</span>
<span class="source-line-no">173</span><span id="line-173"> System.err.println(" -D " + JOB_NAME_CONF_KEY</span>
<span class="source-line-no">174</span><span id="line-174"> + "=jobName - use the specified mapreduce job name for the HFile splitter");</span>
<span class="source-line-no">175</span><span id="line-175"> System.err.println("For performance also consider the following options:\n"</span>
<span class="source-line-no">176</span><span id="line-176"> + " -Dmapreduce.map.speculative=false\n" + " -Dmapreduce.reduce.speculative=false");</span>
<span class="source-line-no">177</span><span id="line-177"> }</span>
<span class="source-line-no">178</span><span id="line-178"></span>
<span class="source-line-no">179</span><span id="line-179"> /**</span>
<span class="source-line-no">180</span><span id="line-180"> * Main entry point.</span>
<span class="source-line-no">181</span><span id="line-181"> * @param args The command line parameters.</span>
<span class="source-line-no">182</span><span id="line-182"> * @throws Exception When running the job fails.</span>
<span class="source-line-no">183</span><span id="line-183"> */</span>
<span class="source-line-no">184</span><span id="line-184"> public static void main(String[] args) throws Exception {</span>
<span class="source-line-no">185</span><span id="line-185"> int ret = ToolRunner.run(new MapReduceHFileSplitterJob(HBaseConfiguration.create()), args);</span>
<span class="source-line-no">186</span><span id="line-186"> System.exit(ret);</span>
<span class="source-line-no">187</span><span id="line-187"> }</span>
<span class="source-line-no">188</span><span id="line-188"></span>
<span class="source-line-no">189</span><span id="line-189"> @Override</span>
<span class="source-line-no">190</span><span id="line-190"> public int run(String[] args) throws Exception {</span>
<span class="source-line-no">191</span><span id="line-191"> if (args.length &lt; 2) {</span>
<span class="source-line-no">192</span><span id="line-192"> usage("Wrong number of arguments: " + args.length);</span>
<span class="source-line-no">193</span><span id="line-193"> return -1;</span>
<span class="source-line-no">194</span><span id="line-194"> }</span>
<span class="source-line-no">195</span><span id="line-195"> Job job = createSubmittableJob(args);</span>
<span class="source-line-no">196</span><span id="line-196"> int result = job.waitForCompletion(true) ? 0 : 1;</span>
<span class="source-line-no">197</span><span id="line-197"> return result;</span>
<span class="source-line-no">198</span><span id="line-198"> }</span>
<span class="source-line-no">199</span><span id="line-199"></span>
<span class="source-line-no">200</span><span id="line-200"> private static RegionLocator getRegionLocator(Configuration conf, Connection conn,</span>
<span class="source-line-no">201</span><span id="line-201"> TableName table) throws IOException {</span>
<span class="source-line-no">202</span><span id="line-202"> if (SnapshotRegionLocator.shouldUseSnapshotRegionLocator(conf, table)) {</span>
<span class="source-line-no">203</span><span id="line-203"> return SnapshotRegionLocator.create(conf, table);</span>
<span class="source-line-no">204</span><span id="line-204"> }</span>
<span class="source-line-no">205</span><span id="line-205"></span>
<span class="source-line-no">206</span><span id="line-206"> return conn.getRegionLocator(table);</span>
<span class="source-line-no">207</span><span id="line-207"> }</span>
<span class="source-line-no">208</span><span id="line-208">}</span>
</pre>
</div>
</main>
</body>
</html>