blob: ea851d363220aa07c7102461be79f7ba5abdd44f [file] [log] [blame]
<!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> * Licensed to the Apache Software Foundation (ASF) under one<a name="line.2"></a>
<span class="sourceLineNo">003</span> * or more contributor license agreements. See the NOTICE file<a name="line.3"></a>
<span class="sourceLineNo">004</span> * distributed with this work for additional information<a name="line.4"></a>
<span class="sourceLineNo">005</span> * regarding copyright ownership. The ASF licenses this file<a name="line.5"></a>
<span class="sourceLineNo">006</span> * to you under the Apache License, Version 2.0 (the<a name="line.6"></a>
<span class="sourceLineNo">007</span> * "License"); you may not use this file except in compliance<a name="line.7"></a>
<span class="sourceLineNo">008</span> * with the License. You may obtain a copy of the License at<a name="line.8"></a>
<span class="sourceLineNo">009</span> *<a name="line.9"></a>
<span class="sourceLineNo">010</span> * http://www.apache.org/licenses/LICENSE-2.0<a name="line.10"></a>
<span class="sourceLineNo">011</span> *<a name="line.11"></a>
<span class="sourceLineNo">012</span> * Unless required by applicable law or agreed to in writing, software<a name="line.12"></a>
<span class="sourceLineNo">013</span> * distributed under the License is distributed on an "AS IS" BASIS,<a name="line.13"></a>
<span class="sourceLineNo">014</span> * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.<a name="line.14"></a>
<span class="sourceLineNo">015</span> * See the License for the specific language governing permissions and<a name="line.15"></a>
<span class="sourceLineNo">016</span> * limitations under the License.<a name="line.16"></a>
<span class="sourceLineNo">017</span> */<a name="line.17"></a>
<span class="sourceLineNo">018</span>package org.apache.hadoop.hbase.test;<a name="line.18"></a>
<span class="sourceLineNo">019</span><a name="line.19"></a>
<span class="sourceLineNo">020</span>import static org.junit.Assert.assertEquals;<a name="line.20"></a>
<span class="sourceLineNo">021</span>import static org.junit.Assert.assertTrue;<a name="line.21"></a>
<span class="sourceLineNo">022</span><a name="line.22"></a>
<span class="sourceLineNo">023</span>import java.io.BufferedReader;<a name="line.23"></a>
<span class="sourceLineNo">024</span>import java.io.FileNotFoundException;<a name="line.24"></a>
<span class="sourceLineNo">025</span>import java.io.IOException;<a name="line.25"></a>
<span class="sourceLineNo">026</span>import java.io.InputStream;<a name="line.26"></a>
<span class="sourceLineNo">027</span>import java.io.InputStreamReader;<a name="line.27"></a>
<span class="sourceLineNo">028</span>import java.io.InterruptedIOException;<a name="line.28"></a>
<span class="sourceLineNo">029</span>import java.nio.charset.StandardCharsets;<a name="line.29"></a>
<span class="sourceLineNo">030</span>import java.util.List;<a name="line.30"></a>
<span class="sourceLineNo">031</span>import java.util.Random;<a name="line.31"></a>
<span class="sourceLineNo">032</span>import java.util.Set;<a name="line.32"></a>
<span class="sourceLineNo">033</span>import java.util.SortedSet;<a name="line.33"></a>
<span class="sourceLineNo">034</span>import java.util.TreeSet;<a name="line.34"></a>
<span class="sourceLineNo">035</span>import java.util.concurrent.ThreadLocalRandom;<a name="line.35"></a>
<span class="sourceLineNo">036</span>import java.util.concurrent.atomic.AtomicInteger;<a name="line.36"></a>
<span class="sourceLineNo">037</span>import java.util.regex.Matcher;<a name="line.37"></a>
<span class="sourceLineNo">038</span>import java.util.regex.Pattern;<a name="line.38"></a>
<span class="sourceLineNo">039</span>import org.apache.hadoop.conf.Configuration;<a name="line.39"></a>
<span class="sourceLineNo">040</span>import org.apache.hadoop.fs.FileStatus;<a name="line.40"></a>
<span class="sourceLineNo">041</span>import org.apache.hadoop.fs.FileSystem;<a name="line.41"></a>
<span class="sourceLineNo">042</span>import org.apache.hadoop.fs.LocatedFileStatus;<a name="line.42"></a>
<span class="sourceLineNo">043</span>import org.apache.hadoop.fs.Path;<a name="line.43"></a>
<span class="sourceLineNo">044</span>import org.apache.hadoop.fs.RemoteIterator;<a name="line.44"></a>
<span class="sourceLineNo">045</span>import org.apache.hadoop.hbase.Cell;<a name="line.45"></a>
<span class="sourceLineNo">046</span>import org.apache.hadoop.hbase.HBaseConfiguration;<a name="line.46"></a>
<span class="sourceLineNo">047</span>import org.apache.hadoop.hbase.HConstants;<a name="line.47"></a>
<span class="sourceLineNo">048</span>import org.apache.hadoop.hbase.IntegrationTestBase;<a name="line.48"></a>
<span class="sourceLineNo">049</span>import org.apache.hadoop.hbase.IntegrationTestingUtility;<a name="line.49"></a>
<span class="sourceLineNo">050</span>import org.apache.hadoop.hbase.TableName;<a name="line.50"></a>
<span class="sourceLineNo">051</span>import org.apache.hadoop.hbase.client.Admin;<a name="line.51"></a>
<span class="sourceLineNo">052</span>import org.apache.hadoop.hbase.client.BufferedMutator;<a name="line.52"></a>
<span class="sourceLineNo">053</span>import org.apache.hadoop.hbase.client.BufferedMutatorParams;<a name="line.53"></a>
<span class="sourceLineNo">054</span>import org.apache.hadoop.hbase.client.ColumnFamilyDescriptorBuilder;<a name="line.54"></a>
<span class="sourceLineNo">055</span>import org.apache.hadoop.hbase.client.Connection;<a name="line.55"></a>
<span class="sourceLineNo">056</span>import org.apache.hadoop.hbase.client.ConnectionFactory;<a name="line.56"></a>
<span class="sourceLineNo">057</span>import org.apache.hadoop.hbase.client.Mutation;<a name="line.57"></a>
<span class="sourceLineNo">058</span>import org.apache.hadoop.hbase.client.Put;<a name="line.58"></a>
<span class="sourceLineNo">059</span>import org.apache.hadoop.hbase.client.Result;<a name="line.59"></a>
<span class="sourceLineNo">060</span>import org.apache.hadoop.hbase.client.Scan;<a name="line.60"></a>
<span class="sourceLineNo">061</span>import org.apache.hadoop.hbase.client.TableDescriptor;<a name="line.61"></a>
<span class="sourceLineNo">062</span>import org.apache.hadoop.hbase.client.TableDescriptorBuilder;<a name="line.62"></a>
<span class="sourceLineNo">063</span>import org.apache.hadoop.hbase.io.ImmutableBytesWritable;<a name="line.63"></a>
<span class="sourceLineNo">064</span>import org.apache.hadoop.hbase.mapreduce.NMapInputFormat;<a name="line.64"></a>
<span class="sourceLineNo">065</span>import org.apache.hadoop.hbase.mapreduce.TableMapReduceUtil;<a name="line.65"></a>
<span class="sourceLineNo">066</span>import org.apache.hadoop.hbase.mapreduce.TableMapper;<a name="line.66"></a>
<span class="sourceLineNo">067</span>import org.apache.hadoop.hbase.mapreduce.TableRecordReaderImpl;<a name="line.67"></a>
<span class="sourceLineNo">068</span>import org.apache.hadoop.hbase.mapreduce.WALPlayer;<a name="line.68"></a>
<span class="sourceLineNo">069</span>import org.apache.hadoop.hbase.testclassification.IntegrationTests;<a name="line.69"></a>
<span class="sourceLineNo">070</span>import org.apache.hadoop.hbase.util.AbstractHBaseTool;<a name="line.70"></a>
<span class="sourceLineNo">071</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.71"></a>
<span class="sourceLineNo">072</span>import org.apache.hadoop.hbase.util.CommonFSUtils;<a name="line.72"></a>
<span class="sourceLineNo">073</span>import org.apache.hadoop.hbase.wal.WALEdit;<a name="line.73"></a>
<span class="sourceLineNo">074</span>import org.apache.hadoop.hbase.wal.WALKey;<a name="line.74"></a>
<span class="sourceLineNo">075</span>import org.apache.hadoop.io.BytesWritable;<a name="line.75"></a>
<span class="sourceLineNo">076</span>import org.apache.hadoop.io.NullWritable;<a name="line.76"></a>
<span class="sourceLineNo">077</span>import org.apache.hadoop.io.Text;<a name="line.77"></a>
<span class="sourceLineNo">078</span>import org.apache.hadoop.mapreduce.Counter;<a name="line.78"></a>
<span class="sourceLineNo">079</span>import org.apache.hadoop.mapreduce.Job;<a name="line.79"></a>
<span class="sourceLineNo">080</span>import org.apache.hadoop.mapreduce.Mapper;<a name="line.80"></a>
<span class="sourceLineNo">081</span>import org.apache.hadoop.mapreduce.Reducer;<a name="line.81"></a>
<span class="sourceLineNo">082</span>import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;<a name="line.82"></a>
<span class="sourceLineNo">083</span>import org.apache.hadoop.mapreduce.lib.output.NullOutputFormat;<a name="line.83"></a>
<span class="sourceLineNo">084</span>import org.apache.hadoop.util.ToolRunner;<a name="line.84"></a>
<span class="sourceLineNo">085</span>import org.junit.Test;<a name="line.85"></a>
<span class="sourceLineNo">086</span>import org.junit.experimental.categories.Category;<a name="line.86"></a>
<span class="sourceLineNo">087</span>import org.slf4j.Logger;<a name="line.87"></a>
<span class="sourceLineNo">088</span>import org.slf4j.LoggerFactory;<a name="line.88"></a>
<span class="sourceLineNo">089</span><a name="line.89"></a>
<span class="sourceLineNo">090</span>import org.apache.hbase.thirdparty.com.google.common.base.Splitter;<a name="line.90"></a>
<span class="sourceLineNo">091</span>import org.apache.hbase.thirdparty.com.google.common.collect.Iterables;<a name="line.91"></a>
<span class="sourceLineNo">092</span>import org.apache.hbase.thirdparty.com.google.common.collect.Sets;<a name="line.92"></a>
<span class="sourceLineNo">093</span>import org.apache.hbase.thirdparty.org.apache.commons.cli.CommandLine;<a name="line.93"></a>
<span class="sourceLineNo">094</span><a name="line.94"></a>
<span class="sourceLineNo">095</span>/**<a name="line.95"></a>
<span class="sourceLineNo">096</span> * A large test which loads a lot of data that has internal references, and verifies the data. In<a name="line.96"></a>
<span class="sourceLineNo">097</span> * load step, 200 map tasks are launched, which in turn write loadmapper.num_to_write (default 100K)<a name="line.97"></a>
<span class="sourceLineNo">098</span> * rows to an hbase table. Rows are written in blocks, for a total of 100 blocks. Each row in a<a name="line.98"></a>
<span class="sourceLineNo">099</span> * block, contains loadmapper.backrefs (default 50) references to random rows in the prev block.<a name="line.99"></a>
<span class="sourceLineNo">100</span> * Verify step is scans the table, and verifies that for every referenced row, the row is actually<a name="line.100"></a>
<span class="sourceLineNo">101</span> * there (no data loss). Failed rows are output from reduce to be saved in the job output dir in<a name="line.101"></a>
<span class="sourceLineNo">102</span> * hdfs and inspected later. This class can be run as a unit test, as an integration test, or from<a name="line.102"></a>
<span class="sourceLineNo">103</span> * the command line Originally taken from Apache Bigtop.<a name="line.103"></a>
<span class="sourceLineNo">104</span> */<a name="line.104"></a>
<span class="sourceLineNo">105</span>@Category(IntegrationTests.class)<a name="line.105"></a>
<span class="sourceLineNo">106</span>public class IntegrationTestLoadAndVerify extends IntegrationTestBase {<a name="line.106"></a>
<span class="sourceLineNo">107</span><a name="line.107"></a>
<span class="sourceLineNo">108</span> private static final Logger LOG = LoggerFactory.getLogger(IntegrationTestLoadAndVerify.class);<a name="line.108"></a>
<span class="sourceLineNo">109</span><a name="line.109"></a>
<span class="sourceLineNo">110</span> private static final String TEST_NAME = "IntegrationTestLoadAndVerify";<a name="line.110"></a>
<span class="sourceLineNo">111</span> private static final byte[] TEST_FAMILY = Bytes.toBytes("f1");<a name="line.111"></a>
<span class="sourceLineNo">112</span> private static final byte[] TEST_QUALIFIER = Bytes.toBytes("q1");<a name="line.112"></a>
<span class="sourceLineNo">113</span><a name="line.113"></a>
<span class="sourceLineNo">114</span> private static final String NUM_TO_WRITE_KEY = "loadmapper.num_to_write";<a name="line.114"></a>
<span class="sourceLineNo">115</span> private static final long NUM_TO_WRITE_DEFAULT = 100 * 1000;<a name="line.115"></a>
<span class="sourceLineNo">116</span><a name="line.116"></a>
<span class="sourceLineNo">117</span> private static final String TABLE_NAME_KEY = "loadmapper.table";<a name="line.117"></a>
<span class="sourceLineNo">118</span> private static final String TABLE_NAME_DEFAULT = "table";<a name="line.118"></a>
<span class="sourceLineNo">119</span><a name="line.119"></a>
<span class="sourceLineNo">120</span> private static final String NUM_BACKREFS_KEY = "loadmapper.backrefs";<a name="line.120"></a>
<span class="sourceLineNo">121</span> private static final int NUM_BACKREFS_DEFAULT = 50;<a name="line.121"></a>
<span class="sourceLineNo">122</span><a name="line.122"></a>
<span class="sourceLineNo">123</span> private static final String NUM_MAP_TASKS_KEY = "loadmapper.map.tasks";<a name="line.123"></a>
<span class="sourceLineNo">124</span> private static final String NUM_REDUCE_TASKS_KEY = "verify.reduce.tasks";<a name="line.124"></a>
<span class="sourceLineNo">125</span> private static final int NUM_MAP_TASKS_DEFAULT = 200;<a name="line.125"></a>
<span class="sourceLineNo">126</span> private static final int NUM_REDUCE_TASKS_DEFAULT = 35;<a name="line.126"></a>
<span class="sourceLineNo">127</span><a name="line.127"></a>
<span class="sourceLineNo">128</span> private static final int SCANNER_CACHING = 500;<a name="line.128"></a>
<span class="sourceLineNo">129</span><a name="line.129"></a>
<span class="sourceLineNo">130</span> private static final int MISSING_ROWS_TO_LOG = 10; // YARN complains when too many counters<a name="line.130"></a>
<span class="sourceLineNo">131</span><a name="line.131"></a>
<span class="sourceLineNo">132</span> private String toRun = null;<a name="line.132"></a>
<span class="sourceLineNo">133</span> private String keysDir = null;<a name="line.133"></a>
<span class="sourceLineNo">134</span><a name="line.134"></a>
<span class="sourceLineNo">135</span> private enum Counters {<a name="line.135"></a>
<span class="sourceLineNo">136</span> ROWS_WRITTEN,<a name="line.136"></a>
<span class="sourceLineNo">137</span> REFERENCES_WRITTEN,<a name="line.137"></a>
<span class="sourceLineNo">138</span> REFERENCES_CHECKED<a name="line.138"></a>
<span class="sourceLineNo">139</span> }<a name="line.139"></a>
<span class="sourceLineNo">140</span><a name="line.140"></a>
<span class="sourceLineNo">141</span> @Override<a name="line.141"></a>
<span class="sourceLineNo">142</span> public void setUpCluster() throws Exception {<a name="line.142"></a>
<span class="sourceLineNo">143</span> util = getTestingUtil(getConf());<a name="line.143"></a>
<span class="sourceLineNo">144</span> util.initializeCluster(3);<a name="line.144"></a>
<span class="sourceLineNo">145</span> this.setConf(util.getConfiguration());<a name="line.145"></a>
<span class="sourceLineNo">146</span> if (!util.isDistributedCluster()) {<a name="line.146"></a>
<span class="sourceLineNo">147</span> getConf().setLong(NUM_TO_WRITE_KEY, NUM_TO_WRITE_DEFAULT / 100);<a name="line.147"></a>
<span class="sourceLineNo">148</span> getConf().setInt(NUM_MAP_TASKS_KEY, NUM_MAP_TASKS_DEFAULT / 100);<a name="line.148"></a>
<span class="sourceLineNo">149</span> getConf().setInt(NUM_REDUCE_TASKS_KEY, NUM_REDUCE_TASKS_DEFAULT / 10);<a name="line.149"></a>
<span class="sourceLineNo">150</span> util.startMiniMapReduceCluster();<a name="line.150"></a>
<span class="sourceLineNo">151</span> }<a name="line.151"></a>
<span class="sourceLineNo">152</span> }<a name="line.152"></a>
<span class="sourceLineNo">153</span><a name="line.153"></a>
<span class="sourceLineNo">154</span> @Override<a name="line.154"></a>
<span class="sourceLineNo">155</span> public void cleanUpCluster() throws Exception {<a name="line.155"></a>
<span class="sourceLineNo">156</span> super.cleanUpCluster();<a name="line.156"></a>
<span class="sourceLineNo">157</span> if (!util.isDistributedCluster()) {<a name="line.157"></a>
<span class="sourceLineNo">158</span> util.shutdownMiniMapReduceCluster();<a name="line.158"></a>
<span class="sourceLineNo">159</span> }<a name="line.159"></a>
<span class="sourceLineNo">160</span> }<a name="line.160"></a>
<span class="sourceLineNo">161</span><a name="line.161"></a>
<span class="sourceLineNo">162</span> /**<a name="line.162"></a>
<span class="sourceLineNo">163</span> * Converts a "long" value between endian systems. Borrowed from Apache Commons IO<a name="line.163"></a>
<span class="sourceLineNo">164</span> * @param value value to convert<a name="line.164"></a>
<span class="sourceLineNo">165</span> * @return the converted value<a name="line.165"></a>
<span class="sourceLineNo">166</span> */<a name="line.166"></a>
<span class="sourceLineNo">167</span> public static long swapLong(long value) {<a name="line.167"></a>
<span class="sourceLineNo">168</span> return (((value &gt;&gt; 0) &amp; 0xff) &lt;&lt; 56) + (((value &gt;&gt; 8) &amp; 0xff) &lt;&lt; 48)<a name="line.168"></a>
<span class="sourceLineNo">169</span> + (((value &gt;&gt; 16) &amp; 0xff) &lt;&lt; 40) + (((value &gt;&gt; 24) &amp; 0xff) &lt;&lt; 32)<a name="line.169"></a>
<span class="sourceLineNo">170</span> + (((value &gt;&gt; 32) &amp; 0xff) &lt;&lt; 24) + (((value &gt;&gt; 40) &amp; 0xff) &lt;&lt; 16)<a name="line.170"></a>
<span class="sourceLineNo">171</span> + (((value &gt;&gt; 48) &amp; 0xff) &lt;&lt; 8) + (((value &gt;&gt; 56) &amp; 0xff) &lt;&lt; 0);<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> public static class LoadMapper<a name="line.174"></a>
<span class="sourceLineNo">175</span> extends Mapper&lt;NullWritable, NullWritable, NullWritable, NullWritable&gt; {<a name="line.175"></a>
<span class="sourceLineNo">176</span> protected long recordsToWrite;<a name="line.176"></a>
<span class="sourceLineNo">177</span> protected Connection connection;<a name="line.177"></a>
<span class="sourceLineNo">178</span> protected BufferedMutator mutator;<a name="line.178"></a>
<span class="sourceLineNo">179</span> protected Configuration conf;<a name="line.179"></a>
<span class="sourceLineNo">180</span> protected int numBackReferencesPerRow;<a name="line.180"></a>
<span class="sourceLineNo">181</span> protected String shortTaskId;<a name="line.181"></a>
<span class="sourceLineNo">182</span> protected Counter rowsWritten, refsWritten;<a name="line.182"></a>
<span class="sourceLineNo">183</span><a name="line.183"></a>
<span class="sourceLineNo">184</span> @Override<a name="line.184"></a>
<span class="sourceLineNo">185</span> public void setup(Context context) throws IOException {<a name="line.185"></a>
<span class="sourceLineNo">186</span> conf = context.getConfiguration();<a name="line.186"></a>
<span class="sourceLineNo">187</span> recordsToWrite = conf.getLong(NUM_TO_WRITE_KEY, NUM_TO_WRITE_DEFAULT);<a name="line.187"></a>
<span class="sourceLineNo">188</span> String tableName = conf.get(TABLE_NAME_KEY, TABLE_NAME_DEFAULT);<a name="line.188"></a>
<span class="sourceLineNo">189</span> numBackReferencesPerRow = conf.getInt(NUM_BACKREFS_KEY, NUM_BACKREFS_DEFAULT);<a name="line.189"></a>
<span class="sourceLineNo">190</span> this.connection = ConnectionFactory.createConnection(conf);<a name="line.190"></a>
<span class="sourceLineNo">191</span> mutator = connection.getBufferedMutator(<a name="line.191"></a>
<span class="sourceLineNo">192</span> new BufferedMutatorParams(TableName.valueOf(tableName)).writeBufferSize(4 * 1024 * 1024));<a name="line.192"></a>
<span class="sourceLineNo">193</span><a name="line.193"></a>
<span class="sourceLineNo">194</span> String taskId = conf.get("mapreduce.task.attempt.id");<a name="line.194"></a>
<span class="sourceLineNo">195</span> Matcher matcher = Pattern.compile(".+_m_(\\d+_\\d+)").matcher(taskId);<a name="line.195"></a>
<span class="sourceLineNo">196</span> if (!matcher.matches()) {<a name="line.196"></a>
<span class="sourceLineNo">197</span> throw new RuntimeException("Strange task ID: " + taskId);<a name="line.197"></a>
<span class="sourceLineNo">198</span> }<a name="line.198"></a>
<span class="sourceLineNo">199</span> shortTaskId = matcher.group(1);<a name="line.199"></a>
<span class="sourceLineNo">200</span><a name="line.200"></a>
<span class="sourceLineNo">201</span> rowsWritten = context.getCounter(Counters.ROWS_WRITTEN);<a name="line.201"></a>
<span class="sourceLineNo">202</span> refsWritten = context.getCounter(Counters.REFERENCES_WRITTEN);<a name="line.202"></a>
<span class="sourceLineNo">203</span> }<a name="line.203"></a>
<span class="sourceLineNo">204</span><a name="line.204"></a>
<span class="sourceLineNo">205</span> @Override<a name="line.205"></a>
<span class="sourceLineNo">206</span> public void cleanup(Context context) throws IOException {<a name="line.206"></a>
<span class="sourceLineNo">207</span> mutator.close();<a name="line.207"></a>
<span class="sourceLineNo">208</span> connection.close();<a name="line.208"></a>
<span class="sourceLineNo">209</span> }<a name="line.209"></a>
<span class="sourceLineNo">210</span><a name="line.210"></a>
<span class="sourceLineNo">211</span> @Override<a name="line.211"></a>
<span class="sourceLineNo">212</span> protected void map(NullWritable key, NullWritable value, Context context)<a name="line.212"></a>
<span class="sourceLineNo">213</span> throws IOException, InterruptedException {<a name="line.213"></a>
<span class="sourceLineNo">214</span><a name="line.214"></a>
<span class="sourceLineNo">215</span> String suffix = "/" + shortTaskId;<a name="line.215"></a>
<span class="sourceLineNo">216</span> byte[] row = Bytes.add(new byte[8], Bytes.toBytes(suffix));<a name="line.216"></a>
<span class="sourceLineNo">217</span> int BLOCK_SIZE = (int) (recordsToWrite / 100);<a name="line.217"></a>
<span class="sourceLineNo">218</span> Random rand = ThreadLocalRandom.current();<a name="line.218"></a>
<span class="sourceLineNo">219</span><a name="line.219"></a>
<span class="sourceLineNo">220</span> for (long i = 0; i &lt; recordsToWrite;) {<a name="line.220"></a>
<span class="sourceLineNo">221</span> long blockStart = i;<a name="line.221"></a>
<span class="sourceLineNo">222</span> for (long idxInBlock = 0; idxInBlock &lt; BLOCK_SIZE<a name="line.222"></a>
<span class="sourceLineNo">223</span> &amp;&amp; i &lt; recordsToWrite; idxInBlock++, i++) {<a name="line.223"></a>
<span class="sourceLineNo">224</span><a name="line.224"></a>
<span class="sourceLineNo">225</span> long byteSwapped = swapLong(i);<a name="line.225"></a>
<span class="sourceLineNo">226</span> Bytes.putLong(row, 0, byteSwapped);<a name="line.226"></a>
<span class="sourceLineNo">227</span><a name="line.227"></a>
<span class="sourceLineNo">228</span> Put p = new Put(row);<a name="line.228"></a>
<span class="sourceLineNo">229</span> p.addColumn(TEST_FAMILY, TEST_QUALIFIER, HConstants.EMPTY_BYTE_ARRAY);<a name="line.229"></a>
<span class="sourceLineNo">230</span> if (blockStart &gt; 0) {<a name="line.230"></a>
<span class="sourceLineNo">231</span> for (int j = 0; j &lt; numBackReferencesPerRow; j++) {<a name="line.231"></a>
<span class="sourceLineNo">232</span> long referredRow = blockStart - BLOCK_SIZE + rand.nextInt(BLOCK_SIZE);<a name="line.232"></a>
<span class="sourceLineNo">233</span> Bytes.putLong(row, 0, swapLong(referredRow));<a name="line.233"></a>
<span class="sourceLineNo">234</span> p.addColumn(TEST_FAMILY, row, HConstants.EMPTY_BYTE_ARRAY);<a name="line.234"></a>
<span class="sourceLineNo">235</span> }<a name="line.235"></a>
<span class="sourceLineNo">236</span> refsWritten.increment(1);<a name="line.236"></a>
<span class="sourceLineNo">237</span> }<a name="line.237"></a>
<span class="sourceLineNo">238</span> rowsWritten.increment(1);<a name="line.238"></a>
<span class="sourceLineNo">239</span> mutator.mutate(p);<a name="line.239"></a>
<span class="sourceLineNo">240</span><a name="line.240"></a>
<span class="sourceLineNo">241</span> if (i % 100 == 0) {<a name="line.241"></a>
<span class="sourceLineNo">242</span> context.setStatus("Written " + i + "/" + recordsToWrite + " records");<a name="line.242"></a>
<span class="sourceLineNo">243</span> context.progress();<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> // End of block, flush all of them before we start writing anything<a name="line.246"></a>
<span class="sourceLineNo">247</span> // pointing to these!<a name="line.247"></a>
<span class="sourceLineNo">248</span> mutator.flush();<a name="line.248"></a>
<span class="sourceLineNo">249</span> }<a name="line.249"></a>
<span class="sourceLineNo">250</span> }<a name="line.250"></a>
<span class="sourceLineNo">251</span> }<a name="line.251"></a>
<span class="sourceLineNo">252</span><a name="line.252"></a>
<span class="sourceLineNo">253</span> public static class VerifyMapper extends TableMapper&lt;BytesWritable, BytesWritable&gt; {<a name="line.253"></a>
<span class="sourceLineNo">254</span> static final BytesWritable EMPTY = new BytesWritable(HConstants.EMPTY_BYTE_ARRAY);<a name="line.254"></a>
<span class="sourceLineNo">255</span><a name="line.255"></a>
<span class="sourceLineNo">256</span> @Override<a name="line.256"></a>
<span class="sourceLineNo">257</span> protected void map(ImmutableBytesWritable key, Result value, Context context)<a name="line.257"></a>
<span class="sourceLineNo">258</span> throws IOException, InterruptedException {<a name="line.258"></a>
<span class="sourceLineNo">259</span> BytesWritable bwKey = new BytesWritable(key.get());<a name="line.259"></a>
<span class="sourceLineNo">260</span> BytesWritable bwVal = new BytesWritable();<a name="line.260"></a>
<span class="sourceLineNo">261</span> for (Cell kv : value.listCells()) {<a name="line.261"></a>
<span class="sourceLineNo">262</span> if (<a name="line.262"></a>
<span class="sourceLineNo">263</span> Bytes.compareTo(TEST_QUALIFIER, 0, TEST_QUALIFIER.length, kv.getQualifierArray(),<a name="line.263"></a>
<span class="sourceLineNo">264</span> kv.getQualifierOffset(), kv.getQualifierLength()) == 0<a name="line.264"></a>
<span class="sourceLineNo">265</span> ) {<a name="line.265"></a>
<span class="sourceLineNo">266</span> context.write(bwKey, EMPTY);<a name="line.266"></a>
<span class="sourceLineNo">267</span> } else {<a name="line.267"></a>
<span class="sourceLineNo">268</span> bwVal.set(kv.getQualifierArray(), kv.getQualifierOffset(), kv.getQualifierLength());<a name="line.268"></a>
<span class="sourceLineNo">269</span> context.write(bwVal, bwKey);<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> }<a name="line.272"></a>
<span class="sourceLineNo">273</span> }<a name="line.273"></a>
<span class="sourceLineNo">274</span><a name="line.274"></a>
<span class="sourceLineNo">275</span> public static class VerifyReducer extends Reducer&lt;BytesWritable, BytesWritable, Text, Text&gt; {<a name="line.275"></a>
<span class="sourceLineNo">276</span> private Counter refsChecked;<a name="line.276"></a>
<span class="sourceLineNo">277</span> private Counter rowsWritten;<a name="line.277"></a>
<span class="sourceLineNo">278</span><a name="line.278"></a>
<span class="sourceLineNo">279</span> @Override<a name="line.279"></a>
<span class="sourceLineNo">280</span> public void setup(Context context) throws IOException {<a name="line.280"></a>
<span class="sourceLineNo">281</span> refsChecked = context.getCounter(Counters.REFERENCES_CHECKED);<a name="line.281"></a>
<span class="sourceLineNo">282</span> rowsWritten = context.getCounter(Counters.ROWS_WRITTEN);<a name="line.282"></a>
<span class="sourceLineNo">283</span> }<a name="line.283"></a>
<span class="sourceLineNo">284</span><a name="line.284"></a>
<span class="sourceLineNo">285</span> @Override<a name="line.285"></a>
<span class="sourceLineNo">286</span> protected void reduce(BytesWritable referredRow, Iterable&lt;BytesWritable&gt; referrers,<a name="line.286"></a>
<span class="sourceLineNo">287</span> VerifyReducer.Context ctx) throws IOException, InterruptedException {<a name="line.287"></a>
<span class="sourceLineNo">288</span> boolean gotOriginalRow = false;<a name="line.288"></a>
<span class="sourceLineNo">289</span> int refCount = 0;<a name="line.289"></a>
<span class="sourceLineNo">290</span><a name="line.290"></a>
<span class="sourceLineNo">291</span> for (BytesWritable ref : referrers) {<a name="line.291"></a>
<span class="sourceLineNo">292</span> if (ref.getLength() == 0) {<a name="line.292"></a>
<span class="sourceLineNo">293</span> assert !gotOriginalRow;<a name="line.293"></a>
<span class="sourceLineNo">294</span> gotOriginalRow = true;<a name="line.294"></a>
<span class="sourceLineNo">295</span> } else {<a name="line.295"></a>
<span class="sourceLineNo">296</span> refCount++;<a name="line.296"></a>
<span class="sourceLineNo">297</span> }<a name="line.297"></a>
<span class="sourceLineNo">298</span> }<a name="line.298"></a>
<span class="sourceLineNo">299</span> refsChecked.increment(refCount);<a name="line.299"></a>
<span class="sourceLineNo">300</span><a name="line.300"></a>
<span class="sourceLineNo">301</span> if (!gotOriginalRow) {<a name="line.301"></a>
<span class="sourceLineNo">302</span> String parsedRow = makeRowReadable(referredRow.getBytes(), referredRow.getLength());<a name="line.302"></a>
<span class="sourceLineNo">303</span> String binRow = Bytes.toStringBinary(referredRow.getBytes(), 0, referredRow.getLength());<a name="line.303"></a>
<span class="sourceLineNo">304</span> LOG.error("Reference error row " + parsedRow);<a name="line.304"></a>
<span class="sourceLineNo">305</span> ctx.write(new Text(binRow), new Text(parsedRow));<a name="line.305"></a>
<span class="sourceLineNo">306</span> rowsWritten.increment(1);<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> private String makeRowReadable(byte[] bytes, int length) {<a name="line.310"></a>
<span class="sourceLineNo">311</span> long rowIdx = swapLong(Bytes.toLong(bytes, 0));<a name="line.311"></a>
<span class="sourceLineNo">312</span> String suffix = Bytes.toString(bytes, 8, length - 8);<a name="line.312"></a>
<span class="sourceLineNo">313</span><a name="line.313"></a>
<span class="sourceLineNo">314</span> return "Row #" + rowIdx + " suffix " + suffix;<a name="line.314"></a>
<span class="sourceLineNo">315</span> }<a name="line.315"></a>
<span class="sourceLineNo">316</span> }<a name="line.316"></a>
<span class="sourceLineNo">317</span><a name="line.317"></a>
<span class="sourceLineNo">318</span> protected Job doLoad(Configuration conf, TableDescriptor tableDescriptor) throws Exception {<a name="line.318"></a>
<span class="sourceLineNo">319</span> Path outputDir = getTestDir(TEST_NAME, "load-output");<a name="line.319"></a>
<span class="sourceLineNo">320</span> LOG.info("Load output dir: " + outputDir);<a name="line.320"></a>
<span class="sourceLineNo">321</span><a name="line.321"></a>
<span class="sourceLineNo">322</span> NMapInputFormat.setNumMapTasks(conf, conf.getInt(NUM_MAP_TASKS_KEY, NUM_MAP_TASKS_DEFAULT));<a name="line.322"></a>
<span class="sourceLineNo">323</span> conf.set(TABLE_NAME_KEY, tableDescriptor.getTableName().getNameAsString());<a name="line.323"></a>
<span class="sourceLineNo">324</span><a name="line.324"></a>
<span class="sourceLineNo">325</span> Job job = Job.getInstance(conf);<a name="line.325"></a>
<span class="sourceLineNo">326</span> job.setJobName(TEST_NAME + " Load for " + tableDescriptor.getTableName());<a name="line.326"></a>
<span class="sourceLineNo">327</span> job.setJarByClass(this.getClass());<a name="line.327"></a>
<span class="sourceLineNo">328</span> setMapperClass(job);<a name="line.328"></a>
<span class="sourceLineNo">329</span> job.setInputFormatClass(NMapInputFormat.class);<a name="line.329"></a>
<span class="sourceLineNo">330</span> job.setNumReduceTasks(0);<a name="line.330"></a>
<span class="sourceLineNo">331</span> setJobScannerConf(job);<a name="line.331"></a>
<span class="sourceLineNo">332</span> FileOutputFormat.setOutputPath(job, outputDir);<a name="line.332"></a>
<span class="sourceLineNo">333</span><a name="line.333"></a>
<span class="sourceLineNo">334</span> TableMapReduceUtil.addDependencyJars(job);<a name="line.334"></a>
<span class="sourceLineNo">335</span><a name="line.335"></a>
<span class="sourceLineNo">336</span> TableMapReduceUtil.addDependencyJarsForClasses(job.getConfiguration(), AbstractHBaseTool.class);<a name="line.336"></a>
<span class="sourceLineNo">337</span> TableMapReduceUtil.initCredentials(job);<a name="line.337"></a>
<span class="sourceLineNo">338</span> assertTrue(job.waitForCompletion(true));<a name="line.338"></a>
<span class="sourceLineNo">339</span> return job;<a name="line.339"></a>
<span class="sourceLineNo">340</span> }<a name="line.340"></a>
<span class="sourceLineNo">341</span><a name="line.341"></a>
<span class="sourceLineNo">342</span> protected void setMapperClass(Job job) {<a name="line.342"></a>
<span class="sourceLineNo">343</span> job.setMapperClass(LoadMapper.class);<a name="line.343"></a>
<span class="sourceLineNo">344</span> }<a name="line.344"></a>
<span class="sourceLineNo">345</span><a name="line.345"></a>
<span class="sourceLineNo">346</span> protected void doVerify(Configuration conf, TableDescriptor tableDescriptor) throws Exception {<a name="line.346"></a>
<span class="sourceLineNo">347</span> Path outputDir = getTestDir(TEST_NAME, "verify-output");<a name="line.347"></a>
<span class="sourceLineNo">348</span> LOG.info("Verify output dir: " + outputDir);<a name="line.348"></a>
<span class="sourceLineNo">349</span><a name="line.349"></a>
<span class="sourceLineNo">350</span> Job job = Job.getInstance(conf);<a name="line.350"></a>
<span class="sourceLineNo">351</span> job.setJarByClass(this.getClass());<a name="line.351"></a>
<span class="sourceLineNo">352</span> job.setJobName(TEST_NAME + " Verification for " + tableDescriptor.getTableName());<a name="line.352"></a>
<span class="sourceLineNo">353</span> setJobScannerConf(job);<a name="line.353"></a>
<span class="sourceLineNo">354</span><a name="line.354"></a>
<span class="sourceLineNo">355</span> Scan scan = new Scan();<a name="line.355"></a>
<span class="sourceLineNo">356</span><a name="line.356"></a>
<span class="sourceLineNo">357</span> TableMapReduceUtil.initTableMapperJob(tableDescriptor.getTableName().getNameAsString(), scan,<a name="line.357"></a>
<span class="sourceLineNo">358</span> VerifyMapper.class, BytesWritable.class, BytesWritable.class, job);<a name="line.358"></a>
<span class="sourceLineNo">359</span> TableMapReduceUtil.addDependencyJarsForClasses(job.getConfiguration(), AbstractHBaseTool.class);<a name="line.359"></a>
<span class="sourceLineNo">360</span> int scannerCaching = conf.getInt("verify.scannercaching", SCANNER_CACHING);<a name="line.360"></a>
<span class="sourceLineNo">361</span> TableMapReduceUtil.setScannerCaching(job, scannerCaching);<a name="line.361"></a>
<span class="sourceLineNo">362</span><a name="line.362"></a>
<span class="sourceLineNo">363</span> job.setReducerClass(VerifyReducer.class);<a name="line.363"></a>
<span class="sourceLineNo">364</span> job.setNumReduceTasks(conf.getInt(NUM_REDUCE_TASKS_KEY, NUM_REDUCE_TASKS_DEFAULT));<a name="line.364"></a>
<span class="sourceLineNo">365</span> FileOutputFormat.setOutputPath(job, outputDir);<a name="line.365"></a>
<span class="sourceLineNo">366</span> assertTrue(job.waitForCompletion(true));<a name="line.366"></a>
<span class="sourceLineNo">367</span><a name="line.367"></a>
<span class="sourceLineNo">368</span> long numOutputRecords = job.getCounters().findCounter(Counters.ROWS_WRITTEN).getValue();<a name="line.368"></a>
<span class="sourceLineNo">369</span> assertEquals(0, numOutputRecords);<a name="line.369"></a>
<span class="sourceLineNo">370</span> }<a name="line.370"></a>
<span class="sourceLineNo">371</span><a name="line.371"></a>
<span class="sourceLineNo">372</span> /**<a name="line.372"></a>
<span class="sourceLineNo">373</span> * Tool to search missing rows in WALs and hfiles. Pass in file or dir of keys to search for. Key<a name="line.373"></a>
<span class="sourceLineNo">374</span> * file must have been written by Verify step (we depend on the format it writes out. We'll read<a name="line.374"></a>
<span class="sourceLineNo">375</span> * them in and then search in hbase WALs and oldWALs dirs (Some of this is TODO).<a name="line.375"></a>
<span class="sourceLineNo">376</span> */<a name="line.376"></a>
<span class="sourceLineNo">377</span> public static class WALSearcher extends WALPlayer {<a name="line.377"></a>
<span class="sourceLineNo">378</span> public WALSearcher(Configuration conf) {<a name="line.378"></a>
<span class="sourceLineNo">379</span> super(conf);<a name="line.379"></a>
<span class="sourceLineNo">380</span> }<a name="line.380"></a>
<span class="sourceLineNo">381</span><a name="line.381"></a>
<span class="sourceLineNo">382</span> /**<a name="line.382"></a>
<span class="sourceLineNo">383</span> * The actual searcher mapper.<a name="line.383"></a>
<span class="sourceLineNo">384</span> */<a name="line.384"></a>
<span class="sourceLineNo">385</span> public static class WALMapperSearcher extends WALMapper {<a name="line.385"></a>
<span class="sourceLineNo">386</span> private SortedSet&lt;byte[]&gt; keysToFind;<a name="line.386"></a>
<span class="sourceLineNo">387</span> private AtomicInteger rows = new AtomicInteger(0);<a name="line.387"></a>
<span class="sourceLineNo">388</span><a name="line.388"></a>
<span class="sourceLineNo">389</span> @Override<a name="line.389"></a>
<span class="sourceLineNo">390</span> public void setup(Mapper&lt;WALKey, WALEdit, ImmutableBytesWritable, Mutation&gt;.Context context)<a name="line.390"></a>
<span class="sourceLineNo">391</span> throws IOException {<a name="line.391"></a>
<span class="sourceLineNo">392</span> super.setup(context);<a name="line.392"></a>
<span class="sourceLineNo">393</span> try {<a name="line.393"></a>
<span class="sourceLineNo">394</span> this.keysToFind = readKeysToSearch(context.getConfiguration());<a name="line.394"></a>
<span class="sourceLineNo">395</span> LOG.info("Loaded keys to find: count=" + this.keysToFind.size());<a name="line.395"></a>
<span class="sourceLineNo">396</span> } catch (InterruptedException e) {<a name="line.396"></a>
<span class="sourceLineNo">397</span> throw new InterruptedIOException(e.toString());<a name="line.397"></a>
<span class="sourceLineNo">398</span> }<a name="line.398"></a>
<span class="sourceLineNo">399</span> }<a name="line.399"></a>
<span class="sourceLineNo">400</span><a name="line.400"></a>
<span class="sourceLineNo">401</span> @Override<a name="line.401"></a>
<span class="sourceLineNo">402</span> protected boolean filter(Context context, Cell cell) {<a name="line.402"></a>
<span class="sourceLineNo">403</span> // TODO: Can I do a better compare than this copying out key?<a name="line.403"></a>
<span class="sourceLineNo">404</span> byte[] row = new byte[cell.getRowLength()];<a name="line.404"></a>
<span class="sourceLineNo">405</span> System.arraycopy(cell.getRowArray(), cell.getRowOffset(), row, 0, cell.getRowLength());<a name="line.405"></a>
<span class="sourceLineNo">406</span> boolean b = this.keysToFind.contains(row);<a name="line.406"></a>
<span class="sourceLineNo">407</span> if (b) {<a name="line.407"></a>
<span class="sourceLineNo">408</span> String keyStr = Bytes.toStringBinary(row);<a name="line.408"></a>
<span class="sourceLineNo">409</span> try {<a name="line.409"></a>
<span class="sourceLineNo">410</span> LOG.info("Found cell=" + cell + " , walKey=" + context.getCurrentKey());<a name="line.410"></a>
<span class="sourceLineNo">411</span> } catch (IOException | InterruptedException e) {<a name="line.411"></a>
<span class="sourceLineNo">412</span> LOG.warn(e.toString(), e);<a name="line.412"></a>
<span class="sourceLineNo">413</span> }<a name="line.413"></a>
<span class="sourceLineNo">414</span> if (rows.addAndGet(1) &lt; MISSING_ROWS_TO_LOG) {<a name="line.414"></a>
<span class="sourceLineNo">415</span> context.getCounter(FOUND_GROUP_KEY, keyStr).increment(1);<a name="line.415"></a>
<span class="sourceLineNo">416</span> }<a name="line.416"></a>
<span class="sourceLineNo">417</span> context.getCounter(FOUND_GROUP_KEY, "CELL_WITH_MISSING_ROW").increment(1);<a name="line.417"></a>
<span class="sourceLineNo">418</span> }<a name="line.418"></a>
<span class="sourceLineNo">419</span> return b;<a name="line.419"></a>
<span class="sourceLineNo">420</span> }<a name="line.420"></a>
<span class="sourceLineNo">421</span> }<a name="line.421"></a>
<span class="sourceLineNo">422</span><a name="line.422"></a>
<span class="sourceLineNo">423</span> // Put in place the above WALMapperSearcher.<a name="line.423"></a>
<span class="sourceLineNo">424</span> @Override<a name="line.424"></a>
<span class="sourceLineNo">425</span> public Job createSubmittableJob(String[] args) throws IOException {<a name="line.425"></a>
<span class="sourceLineNo">426</span> Job job = super.createSubmittableJob(args);<a name="line.426"></a>
<span class="sourceLineNo">427</span> // Call my class instead.<a name="line.427"></a>
<span class="sourceLineNo">428</span> job.setJarByClass(WALMapperSearcher.class);<a name="line.428"></a>
<span class="sourceLineNo">429</span> job.setMapperClass(WALMapperSearcher.class);<a name="line.429"></a>
<span class="sourceLineNo">430</span> job.setOutputFormatClass(NullOutputFormat.class);<a name="line.430"></a>
<span class="sourceLineNo">431</span> return job;<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> static final String FOUND_GROUP_KEY = "Found";<a name="line.435"></a>
<span class="sourceLineNo">436</span> static final String SEARCHER_INPUTDIR_KEY = "searcher.keys.inputdir";<a name="line.436"></a>
<span class="sourceLineNo">437</span><a name="line.437"></a>
<span class="sourceLineNo">438</span> static SortedSet&lt;byte[]&gt; readKeysToSearch(final Configuration conf)<a name="line.438"></a>
<span class="sourceLineNo">439</span> throws IOException, InterruptedException {<a name="line.439"></a>
<span class="sourceLineNo">440</span> Path keysInputDir = new Path(conf.get(SEARCHER_INPUTDIR_KEY));<a name="line.440"></a>
<span class="sourceLineNo">441</span> FileSystem fs = FileSystem.get(conf);<a name="line.441"></a>
<span class="sourceLineNo">442</span> SortedSet&lt;byte[]&gt; result = new TreeSet&lt;&gt;(Bytes.BYTES_COMPARATOR);<a name="line.442"></a>
<span class="sourceLineNo">443</span> if (!fs.exists(keysInputDir)) {<a name="line.443"></a>
<span class="sourceLineNo">444</span> throw new FileNotFoundException(keysInputDir.toString());<a name="line.444"></a>
<span class="sourceLineNo">445</span> }<a name="line.445"></a>
<span class="sourceLineNo">446</span> if (!fs.isDirectory(keysInputDir)) {<a name="line.446"></a>
<span class="sourceLineNo">447</span> FileStatus keyFileStatus = fs.getFileStatus(keysInputDir);<a name="line.447"></a>
<span class="sourceLineNo">448</span> readFileToSearch(fs, keyFileStatus, result);<a name="line.448"></a>
<span class="sourceLineNo">449</span> } else {<a name="line.449"></a>
<span class="sourceLineNo">450</span> RemoteIterator&lt;LocatedFileStatus&gt; iterator = fs.listFiles(keysInputDir, false);<a name="line.450"></a>
<span class="sourceLineNo">451</span> while (iterator.hasNext()) {<a name="line.451"></a>
<span class="sourceLineNo">452</span> LocatedFileStatus keyFileStatus = iterator.next();<a name="line.452"></a>
<span class="sourceLineNo">453</span> // Skip "_SUCCESS" file.<a name="line.453"></a>
<span class="sourceLineNo">454</span> if (keyFileStatus.getPath().getName().startsWith("_")) continue;<a name="line.454"></a>
<span class="sourceLineNo">455</span> readFileToSearch(fs, keyFileStatus, result);<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> return result;<a name="line.458"></a>
<span class="sourceLineNo">459</span> }<a name="line.459"></a>
<span class="sourceLineNo">460</span><a name="line.460"></a>
<span class="sourceLineNo">461</span> private static SortedSet&lt;byte[]&gt; readFileToSearch(final FileSystem fs,<a name="line.461"></a>
<span class="sourceLineNo">462</span> final FileStatus keyFileStatus, SortedSet&lt;byte[]&gt; result)<a name="line.462"></a>
<span class="sourceLineNo">463</span> throws IOException, InterruptedException {<a name="line.463"></a>
<span class="sourceLineNo">464</span> // verify uses file output format and writes &lt;Text, Text&gt;. We can read it as a text file<a name="line.464"></a>
<span class="sourceLineNo">465</span> try (InputStream in = fs.open(keyFileStatus.getPath()); BufferedReader reader =<a name="line.465"></a>
<span class="sourceLineNo">466</span> new BufferedReader(new InputStreamReader(in, StandardCharsets.UTF_8))) {<a name="line.466"></a>
<span class="sourceLineNo">467</span> // extract out the key and return that missing as a missing key<a name="line.467"></a>
<span class="sourceLineNo">468</span> String line;<a name="line.468"></a>
<span class="sourceLineNo">469</span> while ((line = reader.readLine()) != null) {<a name="line.469"></a>
<span class="sourceLineNo">470</span> if (line.isEmpty()) continue;<a name="line.470"></a>
<span class="sourceLineNo">471</span> List&lt;String&gt; parts = Splitter.onPattern("\\s+").splitToList(line);<a name="line.471"></a>
<span class="sourceLineNo">472</span> if (parts.size() &gt;= 1) {<a name="line.472"></a>
<span class="sourceLineNo">473</span> result.add(Bytes.toBytesBinary(Iterables.get(parts, 0)));<a name="line.473"></a>
<span class="sourceLineNo">474</span> } else {<a name="line.474"></a>
<span class="sourceLineNo">475</span> LOG.info("Cannot parse key from: " + line);<a name="line.475"></a>
<span class="sourceLineNo">476</span> }<a name="line.476"></a>
<span class="sourceLineNo">477</span> }<a name="line.477"></a>
<span class="sourceLineNo">478</span> }<a name="line.478"></a>
<span class="sourceLineNo">479</span> return result;<a name="line.479"></a>
<span class="sourceLineNo">480</span> }<a name="line.480"></a>
<span class="sourceLineNo">481</span><a name="line.481"></a>
<span class="sourceLineNo">482</span> private int doSearch(String keysDir) throws Exception {<a name="line.482"></a>
<span class="sourceLineNo">483</span> Path inputDir = new Path(keysDir);<a name="line.483"></a>
<span class="sourceLineNo">484</span><a name="line.484"></a>
<span class="sourceLineNo">485</span> getConf().set(SEARCHER_INPUTDIR_KEY, inputDir.toString());<a name="line.485"></a>
<span class="sourceLineNo">486</span> SortedSet&lt;byte[]&gt; keys = readKeysToSearch(getConf());<a name="line.486"></a>
<span class="sourceLineNo">487</span> if (keys.isEmpty()) throw new RuntimeException("No keys to find");<a name="line.487"></a>
<span class="sourceLineNo">488</span> LOG.info("Count of keys to find: " + keys.size());<a name="line.488"></a>
<span class="sourceLineNo">489</span> for (byte[] key : keys)<a name="line.489"></a>
<span class="sourceLineNo">490</span> LOG.info("Key: " + Bytes.toStringBinary(key));<a name="line.490"></a>
<span class="sourceLineNo">491</span> // Now read all WALs. In two dirs. Presumes certain layout.<a name="line.491"></a>
<span class="sourceLineNo">492</span> Path walsDir = new Path(CommonFSUtils.getWALRootDir(getConf()), HConstants.HREGION_LOGDIR_NAME);<a name="line.492"></a>
<span class="sourceLineNo">493</span> Path oldWalsDir =<a name="line.493"></a>
<span class="sourceLineNo">494</span> new Path(CommonFSUtils.getWALRootDir(getConf()), HConstants.HREGION_OLDLOGDIR_NAME);<a name="line.494"></a>
<span class="sourceLineNo">495</span> LOG.info("Running Search with keys inputDir=" + inputDir + " against "<a name="line.495"></a>
<span class="sourceLineNo">496</span> + getConf().get(HConstants.HBASE_DIR));<a name="line.496"></a>
<span class="sourceLineNo">497</span> int ret = ToolRunner.run(new WALSearcher(getConf()), new String[] { walsDir.toString(), "" });<a name="line.497"></a>
<span class="sourceLineNo">498</span> if (ret != 0) return ret;<a name="line.498"></a>
<span class="sourceLineNo">499</span> return ToolRunner.run(new WALSearcher(getConf()), new String[] { oldWalsDir.toString(), "" });<a name="line.499"></a>
<span class="sourceLineNo">500</span> }<a name="line.500"></a>
<span class="sourceLineNo">501</span><a name="line.501"></a>
<span class="sourceLineNo">502</span> private static void setJobScannerConf(Job job) {<a name="line.502"></a>
<span class="sourceLineNo">503</span> long lpr = job.getConfiguration().getLong(NUM_TO_WRITE_KEY, NUM_TO_WRITE_DEFAULT) / 100;<a name="line.503"></a>
<span class="sourceLineNo">504</span> job.getConfiguration().setInt(TableRecordReaderImpl.LOG_PER_ROW_COUNT, (int) lpr);<a name="line.504"></a>
<span class="sourceLineNo">505</span> }<a name="line.505"></a>
<span class="sourceLineNo">506</span><a name="line.506"></a>
<span class="sourceLineNo">507</span> public Path getTestDir(String testName, String subdir) throws IOException {<a name="line.507"></a>
<span class="sourceLineNo">508</span> Path testDir = util.getDataTestDirOnTestFS(testName);<a name="line.508"></a>
<span class="sourceLineNo">509</span> FileSystem fs = FileSystem.get(getConf());<a name="line.509"></a>
<span class="sourceLineNo">510</span> fs.deleteOnExit(testDir);<a name="line.510"></a>
<span class="sourceLineNo">511</span><a name="line.511"></a>
<span class="sourceLineNo">512</span> return new Path(new Path(testDir, testName), subdir);<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> @Test<a name="line.515"></a>
<span class="sourceLineNo">516</span> public void testLoadAndVerify() throws Exception {<a name="line.516"></a>
<span class="sourceLineNo">517</span> TableDescriptor tableDescriptor =<a name="line.517"></a>
<span class="sourceLineNo">518</span> TableDescriptorBuilder.newBuilder(TableName.valueOf(TEST_NAME))<a name="line.518"></a>
<span class="sourceLineNo">519</span> .setColumnFamily(ColumnFamilyDescriptorBuilder.of(TEST_FAMILY)).build();<a name="line.519"></a>
<span class="sourceLineNo">520</span><a name="line.520"></a>
<span class="sourceLineNo">521</span> Admin admin = getTestingUtil(getConf()).getAdmin();<a name="line.521"></a>
<span class="sourceLineNo">522</span> admin.createTable(tableDescriptor, Bytes.toBytes(0L), Bytes.toBytes(-1L), 40);<a name="line.522"></a>
<span class="sourceLineNo">523</span><a name="line.523"></a>
<span class="sourceLineNo">524</span> doLoad(getConf(), tableDescriptor);<a name="line.524"></a>
<span class="sourceLineNo">525</span> doVerify(getConf(), tableDescriptor);<a name="line.525"></a>
<span class="sourceLineNo">526</span><a name="line.526"></a>
<span class="sourceLineNo">527</span> // Only disable and drop if we succeeded to verify - otherwise it's useful<a name="line.527"></a>
<span class="sourceLineNo">528</span> // to leave it around for post-mortem<a name="line.528"></a>
<span class="sourceLineNo">529</span> getTestingUtil(getConf()).deleteTable(tableDescriptor.getTableName());<a name="line.529"></a>
<span class="sourceLineNo">530</span> }<a name="line.530"></a>
<span class="sourceLineNo">531</span><a name="line.531"></a>
<span class="sourceLineNo">532</span> @Override<a name="line.532"></a>
<span class="sourceLineNo">533</span> public void printUsage() {<a name="line.533"></a>
<span class="sourceLineNo">534</span> printUsage(this.getClass().getSimpleName() + " &lt;options&gt;"<a name="line.534"></a>
<span class="sourceLineNo">535</span> + " [-Doptions] &lt;load|verify|loadAndVerify|search&gt;", "Options", "");<a name="line.535"></a>
<span class="sourceLineNo">536</span> System.err.println("");<a name="line.536"></a>
<span class="sourceLineNo">537</span> System.err.println(" Loads a table with row dependencies and verifies the dependency chains");<a name="line.537"></a>
<span class="sourceLineNo">538</span> System.err.println("Options");<a name="line.538"></a>
<span class="sourceLineNo">539</span> System.err<a name="line.539"></a>
<span class="sourceLineNo">540</span> .println(" -Dloadmapper.table=&lt;name&gt; Table to write/verify (default autogen)");<a name="line.540"></a>
<span class="sourceLineNo">541</span> System.err<a name="line.541"></a>
<span class="sourceLineNo">542</span> .println(" -Dloadmapper.backrefs=&lt;n&gt; Number of backreferences per row (default 50)");<a name="line.542"></a>
<span class="sourceLineNo">543</span> System.err.println(<a name="line.543"></a>
<span class="sourceLineNo">544</span> " -Dloadmapper.num_to_write=&lt;n&gt; Number of rows per mapper (default 100,000 per mapper)");<a name="line.544"></a>
<span class="sourceLineNo">545</span> System.err.println(<a name="line.545"></a>
<span class="sourceLineNo">546</span> " -Dloadmapper.deleteAfter=&lt;bool&gt; Delete after a successful verify (default true)");<a name="line.546"></a>
<span class="sourceLineNo">547</span> System.err.println(<a name="line.547"></a>
<span class="sourceLineNo">548</span> " -Dloadmapper.numPresplits=&lt;n&gt; Number of presplit regions to start with (default 40)");<a name="line.548"></a>
<span class="sourceLineNo">549</span> System.err<a name="line.549"></a>
<span class="sourceLineNo">550</span> .println(" -Dloadmapper.map.tasks=&lt;n&gt; Number of map tasks for load (default 200)");<a name="line.550"></a>
<span class="sourceLineNo">551</span> System.err<a name="line.551"></a>
<span class="sourceLineNo">552</span> .println(" -Dverify.reduce.tasks=&lt;n&gt; Number of reduce tasks for verify (default 35)");<a name="line.552"></a>
<span class="sourceLineNo">553</span> System.err.println(<a name="line.553"></a>
<span class="sourceLineNo">554</span> " -Dverify.scannercaching=&lt;n&gt; Number hbase scanner caching rows to read (default 50)");<a name="line.554"></a>
<span class="sourceLineNo">555</span> }<a name="line.555"></a>
<span class="sourceLineNo">556</span><a name="line.556"></a>
<span class="sourceLineNo">557</span> @Override<a name="line.557"></a>
<span class="sourceLineNo">558</span> protected void processOptions(CommandLine cmd) {<a name="line.558"></a>
<span class="sourceLineNo">559</span> super.processOptions(cmd);<a name="line.559"></a>
<span class="sourceLineNo">560</span><a name="line.560"></a>
<span class="sourceLineNo">561</span> String[] args = cmd.getArgs();<a name="line.561"></a>
<span class="sourceLineNo">562</span> if (args == null || args.length &lt; 1) {<a name="line.562"></a>
<span class="sourceLineNo">563</span> printUsage();<a name="line.563"></a>
<span class="sourceLineNo">564</span> throw new RuntimeException("Incorrect Number of args.");<a name="line.564"></a>
<span class="sourceLineNo">565</span> }<a name="line.565"></a>
<span class="sourceLineNo">566</span> toRun = args[0];<a name="line.566"></a>
<span class="sourceLineNo">567</span> if (toRun.equalsIgnoreCase("search")) {<a name="line.567"></a>
<span class="sourceLineNo">568</span> if (args.length &gt; 1) {<a name="line.568"></a>
<span class="sourceLineNo">569</span> keysDir = args[1];<a name="line.569"></a>
<span class="sourceLineNo">570</span> }<a name="line.570"></a>
<span class="sourceLineNo">571</span> }<a name="line.571"></a>
<span class="sourceLineNo">572</span> }<a name="line.572"></a>
<span class="sourceLineNo">573</span><a name="line.573"></a>
<span class="sourceLineNo">574</span> @Override<a name="line.574"></a>
<span class="sourceLineNo">575</span> public int runTestFromCommandLine() throws Exception {<a name="line.575"></a>
<span class="sourceLineNo">576</span> IntegrationTestingUtility.setUseDistributedCluster(getConf());<a name="line.576"></a>
<span class="sourceLineNo">577</span> boolean doLoad = false;<a name="line.577"></a>
<span class="sourceLineNo">578</span> boolean doVerify = false;<a name="line.578"></a>
<span class="sourceLineNo">579</span> boolean doSearch = false;<a name="line.579"></a>
<span class="sourceLineNo">580</span> boolean doDelete = getConf().getBoolean("loadmapper.deleteAfter", true);<a name="line.580"></a>
<span class="sourceLineNo">581</span> int numPresplits = getConf().getInt("loadmapper.numPresplits", 40);<a name="line.581"></a>
<span class="sourceLineNo">582</span><a name="line.582"></a>
<span class="sourceLineNo">583</span> if (toRun.equalsIgnoreCase("load")) {<a name="line.583"></a>
<span class="sourceLineNo">584</span> doLoad = true;<a name="line.584"></a>
<span class="sourceLineNo">585</span> } else if (toRun.equalsIgnoreCase("verify")) {<a name="line.585"></a>
<span class="sourceLineNo">586</span> doVerify = true;<a name="line.586"></a>
<span class="sourceLineNo">587</span> } else if (toRun.equalsIgnoreCase("loadAndVerify")) {<a name="line.587"></a>
<span class="sourceLineNo">588</span> doLoad = true;<a name="line.588"></a>
<span class="sourceLineNo">589</span> doVerify = true;<a name="line.589"></a>
<span class="sourceLineNo">590</span> } else if (toRun.equalsIgnoreCase("search")) {<a name="line.590"></a>
<span class="sourceLineNo">591</span> doLoad = false;<a name="line.591"></a>
<span class="sourceLineNo">592</span> doVerify = false;<a name="line.592"></a>
<span class="sourceLineNo">593</span> doSearch = true;<a name="line.593"></a>
<span class="sourceLineNo">594</span> if (keysDir == null) {<a name="line.594"></a>
<span class="sourceLineNo">595</span> System.err.println("Usage: search &lt;KEYS_DIR&gt;]");<a name="line.595"></a>
<span class="sourceLineNo">596</span> return 1;<a name="line.596"></a>
<span class="sourceLineNo">597</span> }<a name="line.597"></a>
<span class="sourceLineNo">598</span> } else {<a name="line.598"></a>
<span class="sourceLineNo">599</span> System.err.println("Invalid argument " + toRun);<a name="line.599"></a>
<span class="sourceLineNo">600</span> printUsage();<a name="line.600"></a>
<span class="sourceLineNo">601</span> return 1;<a name="line.601"></a>
<span class="sourceLineNo">602</span> }<a name="line.602"></a>
<span class="sourceLineNo">603</span><a name="line.603"></a>
<span class="sourceLineNo">604</span> // create HTableDescriptor for specified table<a name="line.604"></a>
<span class="sourceLineNo">605</span> TableName table = getTablename();<a name="line.605"></a>
<span class="sourceLineNo">606</span> TableDescriptor tableDescriptor = TableDescriptorBuilder.newBuilder(table)<a name="line.606"></a>
<span class="sourceLineNo">607</span> .setColumnFamily(ColumnFamilyDescriptorBuilder.of(TEST_FAMILY)).build();<a name="line.607"></a>
<span class="sourceLineNo">608</span><a name="line.608"></a>
<span class="sourceLineNo">609</span> if (doLoad) {<a name="line.609"></a>
<span class="sourceLineNo">610</span> try (Connection conn = ConnectionFactory.createConnection(getConf());<a name="line.610"></a>
<span class="sourceLineNo">611</span> Admin admin = conn.getAdmin()) {<a name="line.611"></a>
<span class="sourceLineNo">612</span> admin.createTable(tableDescriptor, Bytes.toBytes(0L), Bytes.toBytes(-1L), numPresplits);<a name="line.612"></a>
<span class="sourceLineNo">613</span> doLoad(getConf(), tableDescriptor);<a name="line.613"></a>
<span class="sourceLineNo">614</span> }<a name="line.614"></a>
<span class="sourceLineNo">615</span> }<a name="line.615"></a>
<span class="sourceLineNo">616</span> if (doVerify) {<a name="line.616"></a>
<span class="sourceLineNo">617</span> doVerify(getConf(), tableDescriptor);<a name="line.617"></a>
<span class="sourceLineNo">618</span> if (doDelete) {<a name="line.618"></a>
<span class="sourceLineNo">619</span> getTestingUtil(getConf()).deleteTable(tableDescriptor.getTableName());<a name="line.619"></a>
<span class="sourceLineNo">620</span> }<a name="line.620"></a>
<span class="sourceLineNo">621</span> }<a name="line.621"></a>
<span class="sourceLineNo">622</span> if (doSearch) {<a name="line.622"></a>
<span class="sourceLineNo">623</span> return doSearch(keysDir);<a name="line.623"></a>
<span class="sourceLineNo">624</span> }<a name="line.624"></a>
<span class="sourceLineNo">625</span> return 0;<a name="line.625"></a>
<span class="sourceLineNo">626</span> }<a name="line.626"></a>
<span class="sourceLineNo">627</span><a name="line.627"></a>
<span class="sourceLineNo">628</span> @Override<a name="line.628"></a>
<span class="sourceLineNo">629</span> public TableName getTablename() {<a name="line.629"></a>
<span class="sourceLineNo">630</span> return TableName.valueOf(getConf().get(TABLE_NAME_KEY, TEST_NAME));<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> @Override<a name="line.633"></a>
<span class="sourceLineNo">634</span> protected Set&lt;String&gt; getColumnFamilies() {<a name="line.634"></a>
<span class="sourceLineNo">635</span> return Sets.newHashSet(Bytes.toString(TEST_FAMILY));<a name="line.635"></a>
<span class="sourceLineNo">636</span> }<a name="line.636"></a>
<span class="sourceLineNo">637</span><a name="line.637"></a>
<span class="sourceLineNo">638</span> public static void main(String argv[]) throws Exception {<a name="line.638"></a>
<span class="sourceLineNo">639</span> Configuration conf = HBaseConfiguration.create();<a name="line.639"></a>
<span class="sourceLineNo">640</span> IntegrationTestingUtility.setUseDistributedCluster(conf);<a name="line.640"></a>
<span class="sourceLineNo">641</span> int ret = ToolRunner.run(conf, new IntegrationTestLoadAndVerify(), argv);<a name="line.641"></a>
<span class="sourceLineNo">642</span> System.exit(ret);<a name="line.642"></a>
<span class="sourceLineNo">643</span> }<a name="line.643"></a>
<span class="sourceLineNo">644</span>}<a name="line.644"></a>
</pre>
</div>
</body>
</html>