blob: 2a1f4c65924e21ef6ca942350d1a9a2f62fa3f5e [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.mapreduce;<a name="line.18"></a>
<span class="sourceLineNo">019</span><a name="line.19"></a>
<span class="sourceLineNo">020</span>import java.util.Set;<a name="line.20"></a>
<span class="sourceLineNo">021</span>import org.apache.hadoop.conf.Configuration;<a name="line.21"></a>
<span class="sourceLineNo">022</span>import org.apache.hadoop.fs.Path;<a name="line.22"></a>
<span class="sourceLineNo">023</span>import org.apache.hadoop.hbase.HBaseConfiguration;<a name="line.23"></a>
<span class="sourceLineNo">024</span>import org.apache.hadoop.hbase.IntegrationTestBase;<a name="line.24"></a>
<span class="sourceLineNo">025</span>import org.apache.hadoop.hbase.IntegrationTestingUtility;<a name="line.25"></a>
<span class="sourceLineNo">026</span>import org.apache.hadoop.hbase.TableName;<a name="line.26"></a>
<span class="sourceLineNo">027</span>import org.apache.hadoop.hbase.testclassification.IntegrationTests;<a name="line.27"></a>
<span class="sourceLineNo">028</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.28"></a>
<span class="sourceLineNo">029</span>import org.apache.hadoop.hbase.util.EnvironmentEdgeManager;<a name="line.29"></a>
<span class="sourceLineNo">030</span>import org.apache.hadoop.util.ToolRunner;<a name="line.30"></a>
<span class="sourceLineNo">031</span>import org.junit.After;<a name="line.31"></a>
<span class="sourceLineNo">032</span>import org.junit.Before;<a name="line.32"></a>
<span class="sourceLineNo">033</span>import org.junit.experimental.categories.Category;<a name="line.33"></a>
<span class="sourceLineNo">034</span>import org.slf4j.Logger;<a name="line.34"></a>
<span class="sourceLineNo">035</span>import org.slf4j.LoggerFactory;<a name="line.35"></a>
<span class="sourceLineNo">036</span><a name="line.36"></a>
<span class="sourceLineNo">037</span>/**<a name="line.37"></a>
<span class="sourceLineNo">038</span> * An integration test to test {@link TableSnapshotInputFormat} which enables reading directly from<a name="line.38"></a>
<span class="sourceLineNo">039</span> * snapshot files without going through hbase servers. This test creates a table and loads the table<a name="line.39"></a>
<span class="sourceLineNo">040</span> * with the rows ranging from 'aaa' to 'zzz', and for each row, sets the columns f1:(null) and<a name="line.40"></a>
<span class="sourceLineNo">041</span> * f2:(null) to be the the same as the row value.<a name="line.41"></a>
<span class="sourceLineNo">042</span> *<a name="line.42"></a>
<span class="sourceLineNo">043</span> * &lt;pre&gt;<a name="line.43"></a>
<span class="sourceLineNo">044</span> * aaa, f1: =&amp;gt; aaa<a name="line.44"></a>
<span class="sourceLineNo">045</span> * aaa, f2: =&amp;gt; aaa<a name="line.45"></a>
<span class="sourceLineNo">046</span> * aab, f1: =&amp;gt; aab<a name="line.46"></a>
<span class="sourceLineNo">047</span> * ....<a name="line.47"></a>
<span class="sourceLineNo">048</span> * zzz, f2: =&amp;gt; zzz<a name="line.48"></a>
<span class="sourceLineNo">049</span> * &lt;/pre&gt;<a name="line.49"></a>
<span class="sourceLineNo">050</span> *<a name="line.50"></a>
<span class="sourceLineNo">051</span> * Then the test creates a snapshot from this table, and overrides the values in the original table<a name="line.51"></a>
<span class="sourceLineNo">052</span> * with values 'after_snapshot_value'. The test, then runs a mapreduce job over the snapshot with a<a name="line.52"></a>
<span class="sourceLineNo">053</span> * scan start row 'bbb' and stop row 'yyy'. The data is saved in a single reduce output file, and<a name="line.53"></a>
<span class="sourceLineNo">054</span> * inspected later to verify that the MR job has seen all the values from the snapshot.<a name="line.54"></a>
<span class="sourceLineNo">055</span> * &lt;p&gt;<a name="line.55"></a>
<span class="sourceLineNo">056</span> * These parameters can be used to configure the job: &lt;br&gt;<a name="line.56"></a>
<span class="sourceLineNo">057</span> * "IntegrationTestTableSnapshotInputFormat.table" =&amp;gt; the name of the table &lt;br&gt;<a name="line.57"></a>
<span class="sourceLineNo">058</span> * "IntegrationTestTableSnapshotInputFormat.snapshot" =&amp;gt; the name of the snapshot &lt;br&gt;<a name="line.58"></a>
<span class="sourceLineNo">059</span> * "IntegrationTestTableSnapshotInputFormat.numRegions" =&amp;gt; number of regions in the table to be<a name="line.59"></a>
<span class="sourceLineNo">060</span> * created (default, 32). &lt;br&gt;<a name="line.60"></a>
<span class="sourceLineNo">061</span> * "IntegrationTestTableSnapshotInputFormat.tableDir" =&amp;gt; temporary directory to restore the<a name="line.61"></a>
<span class="sourceLineNo">062</span> * snapshot files<a name="line.62"></a>
<span class="sourceLineNo">063</span> */<a name="line.63"></a>
<span class="sourceLineNo">064</span>@Category(IntegrationTests.class)<a name="line.64"></a>
<span class="sourceLineNo">065</span>// Not runnable as a unit test. See TestTableSnapshotInputFormat<a name="line.65"></a>
<span class="sourceLineNo">066</span>public class IntegrationTestTableSnapshotInputFormat extends IntegrationTestBase {<a name="line.66"></a>
<span class="sourceLineNo">067</span> private static final Logger LOG =<a name="line.67"></a>
<span class="sourceLineNo">068</span> LoggerFactory.getLogger(IntegrationTestTableSnapshotInputFormat.class);<a name="line.68"></a>
<span class="sourceLineNo">069</span><a name="line.69"></a>
<span class="sourceLineNo">070</span> private static final String TABLE_NAME_KEY = "IntegrationTestTableSnapshotInputFormat.table";<a name="line.70"></a>
<span class="sourceLineNo">071</span> private static final String DEFAULT_TABLE_NAME = "IntegrationTestTableSnapshotInputFormat";<a name="line.71"></a>
<span class="sourceLineNo">072</span><a name="line.72"></a>
<span class="sourceLineNo">073</span> private static final String SNAPSHOT_NAME_KEY =<a name="line.73"></a>
<span class="sourceLineNo">074</span> "IntegrationTestTableSnapshotInputFormat.snapshot";<a name="line.74"></a>
<span class="sourceLineNo">075</span> private static final String NUM_REGIONS_KEY =<a name="line.75"></a>
<span class="sourceLineNo">076</span> "IntegrationTestTableSnapshotInputFormat.numRegions";<a name="line.76"></a>
<span class="sourceLineNo">077</span><a name="line.77"></a>
<span class="sourceLineNo">078</span> private static final String MR_IMPLEMENTATION_KEY = "IntegrationTestTableSnapshotInputFormat.API";<a name="line.78"></a>
<span class="sourceLineNo">079</span> private static final String MAPRED_IMPLEMENTATION = "mapred";<a name="line.79"></a>
<span class="sourceLineNo">080</span> private static final String MAPREDUCE_IMPLEMENTATION = "mapreduce";<a name="line.80"></a>
<span class="sourceLineNo">081</span><a name="line.81"></a>
<span class="sourceLineNo">082</span> private static final int DEFAULT_NUM_REGIONS = 32;<a name="line.82"></a>
<span class="sourceLineNo">083</span> private static final String TABLE_DIR_KEY = "IntegrationTestTableSnapshotInputFormat.tableDir";<a name="line.83"></a>
<span class="sourceLineNo">084</span><a name="line.84"></a>
<span class="sourceLineNo">085</span> private static final byte[] START_ROW = Bytes.toBytes("bbb");<a name="line.85"></a>
<span class="sourceLineNo">086</span> private static final byte[] END_ROW = Bytes.toBytes("yyy");<a name="line.86"></a>
<span class="sourceLineNo">087</span><a name="line.87"></a>
<span class="sourceLineNo">088</span> // mapred API missing feature pairity with mapreduce. See comments in<a name="line.88"></a>
<span class="sourceLineNo">089</span> // mapred.TestTableSnapshotInputFormat<a name="line.89"></a>
<span class="sourceLineNo">090</span> private static final byte[] MAPRED_START_ROW = Bytes.toBytes("aaa");<a name="line.90"></a>
<span class="sourceLineNo">091</span> private static final byte[] MAPRED_END_ROW = Bytes.toBytes("zz{"); // 'z' + 1 =&gt; '{'<a name="line.91"></a>
<span class="sourceLineNo">092</span><a name="line.92"></a>
<span class="sourceLineNo">093</span> @Override<a name="line.93"></a>
<span class="sourceLineNo">094</span> public void setConf(Configuration conf) {<a name="line.94"></a>
<span class="sourceLineNo">095</span> super.setConf(conf);<a name="line.95"></a>
<span class="sourceLineNo">096</span> util = getTestingUtil(conf);<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> @Override<a name="line.99"></a>
<span class="sourceLineNo">100</span> @Before<a name="line.100"></a>
<span class="sourceLineNo">101</span> public void setUp() throws Exception {<a name="line.101"></a>
<span class="sourceLineNo">102</span> super.setUp();<a name="line.102"></a>
<span class="sourceLineNo">103</span> util = getTestingUtil(getConf());<a name="line.103"></a>
<span class="sourceLineNo">104</span> util.initializeCluster(1);<a name="line.104"></a>
<span class="sourceLineNo">105</span> this.setConf(util.getConfiguration());<a name="line.105"></a>
<span class="sourceLineNo">106</span> }<a name="line.106"></a>
<span class="sourceLineNo">107</span><a name="line.107"></a>
<span class="sourceLineNo">108</span> @Override<a name="line.108"></a>
<span class="sourceLineNo">109</span> @After<a name="line.109"></a>
<span class="sourceLineNo">110</span> public void cleanUp() throws Exception {<a name="line.110"></a>
<span class="sourceLineNo">111</span> util.restoreCluster();<a name="line.111"></a>
<span class="sourceLineNo">112</span> }<a name="line.112"></a>
<span class="sourceLineNo">113</span><a name="line.113"></a>
<span class="sourceLineNo">114</span> @Override<a name="line.114"></a>
<span class="sourceLineNo">115</span> public void setUpCluster() throws Exception {<a name="line.115"></a>
<span class="sourceLineNo">116</span> }<a name="line.116"></a>
<span class="sourceLineNo">117</span><a name="line.117"></a>
<span class="sourceLineNo">118</span> @Override<a name="line.118"></a>
<span class="sourceLineNo">119</span> public int runTestFromCommandLine() throws Exception {<a name="line.119"></a>
<span class="sourceLineNo">120</span> Configuration conf = getConf();<a name="line.120"></a>
<span class="sourceLineNo">121</span> TableName tableName = TableName.valueOf(conf.get(TABLE_NAME_KEY, DEFAULT_TABLE_NAME));<a name="line.121"></a>
<span class="sourceLineNo">122</span> String snapshotName = conf.get(SNAPSHOT_NAME_KEY,<a name="line.122"></a>
<span class="sourceLineNo">123</span> tableName.getQualifierAsString() + "_snapshot_" + EnvironmentEdgeManager.currentTime());<a name="line.123"></a>
<span class="sourceLineNo">124</span> int numRegions = conf.getInt(NUM_REGIONS_KEY, DEFAULT_NUM_REGIONS);<a name="line.124"></a>
<span class="sourceLineNo">125</span> String tableDirStr = conf.get(TABLE_DIR_KEY);<a name="line.125"></a>
<span class="sourceLineNo">126</span> Path tableDir;<a name="line.126"></a>
<span class="sourceLineNo">127</span> if (tableDirStr == null) {<a name="line.127"></a>
<span class="sourceLineNo">128</span> tableDir = util.getDataTestDirOnTestFS(tableName.getQualifierAsString());<a name="line.128"></a>
<span class="sourceLineNo">129</span> } else {<a name="line.129"></a>
<span class="sourceLineNo">130</span> tableDir = new Path(tableDirStr);<a name="line.130"></a>
<span class="sourceLineNo">131</span> }<a name="line.131"></a>
<span class="sourceLineNo">132</span><a name="line.132"></a>
<span class="sourceLineNo">133</span> final String mr = conf.get(MR_IMPLEMENTATION_KEY, MAPREDUCE_IMPLEMENTATION);<a name="line.133"></a>
<span class="sourceLineNo">134</span> if (mr.equalsIgnoreCase(MAPREDUCE_IMPLEMENTATION)) {<a name="line.134"></a>
<span class="sourceLineNo">135</span> /*<a name="line.135"></a>
<span class="sourceLineNo">136</span> * We create the table using HBaseAdmin#createTable(), which will create the table with<a name="line.136"></a>
<span class="sourceLineNo">137</span> * desired number of regions. We pass bbb as startKey and yyy as endKey, so if<a name="line.137"></a>
<span class="sourceLineNo">138</span> * desiredNumRegions is &gt; 2, we create regions empty - bbb and yyy - empty, and we create<a name="line.138"></a>
<span class="sourceLineNo">139</span> * numRegions - 2 regions between bbb - yyy. The test uses a Scan with startRow bbb and endRow<a name="line.139"></a>
<span class="sourceLineNo">140</span> * yyy, so, we expect the first and last region to be filtered out in the input format, and we<a name="line.140"></a>
<span class="sourceLineNo">141</span> * expect numRegions - 2 splits between bbb and yyy.<a name="line.141"></a>
<span class="sourceLineNo">142</span> */<a name="line.142"></a>
<span class="sourceLineNo">143</span> LOG.debug("Running job with mapreduce API.");<a name="line.143"></a>
<span class="sourceLineNo">144</span> int expectedNumSplits = numRegions &gt; 2 ? numRegions - 2 : numRegions;<a name="line.144"></a>
<span class="sourceLineNo">145</span><a name="line.145"></a>
<span class="sourceLineNo">146</span> org.apache.hadoop.hbase.mapreduce.TestTableSnapshotInputFormat.doTestWithMapReduce(util,<a name="line.146"></a>
<span class="sourceLineNo">147</span> tableName, snapshotName, START_ROW, END_ROW, tableDir, numRegions, 1, expectedNumSplits,<a name="line.147"></a>
<span class="sourceLineNo">148</span> false);<a name="line.148"></a>
<span class="sourceLineNo">149</span> } else if (mr.equalsIgnoreCase(MAPRED_IMPLEMENTATION)) {<a name="line.149"></a>
<span class="sourceLineNo">150</span> /*<a name="line.150"></a>
<span class="sourceLineNo">151</span> * Similar considerations to above. The difference is that mapred API does not support<a name="line.151"></a>
<span class="sourceLineNo">152</span> * specifying start/end rows (or a scan object at all). Thus the omission of first and last<a name="line.152"></a>
<span class="sourceLineNo">153</span> * regions are not performed. See comments in mapred.TestTableSnapshotInputFormat for details<a name="line.153"></a>
<span class="sourceLineNo">154</span> * of how that test works around the problem. This feature should be added in follow-on work.<a name="line.154"></a>
<span class="sourceLineNo">155</span> */<a name="line.155"></a>
<span class="sourceLineNo">156</span> LOG.debug("Running job with mapred API.");<a name="line.156"></a>
<span class="sourceLineNo">157</span> int expectedNumSplits = numRegions;<a name="line.157"></a>
<span class="sourceLineNo">158</span><a name="line.158"></a>
<span class="sourceLineNo">159</span> org.apache.hadoop.hbase.mapred.TestTableSnapshotInputFormat.doTestWithMapReduce(util,<a name="line.159"></a>
<span class="sourceLineNo">160</span> tableName, snapshotName, MAPRED_START_ROW, MAPRED_END_ROW, tableDir, numRegions, 1,<a name="line.160"></a>
<span class="sourceLineNo">161</span> expectedNumSplits, false);<a name="line.161"></a>
<span class="sourceLineNo">162</span> } else {<a name="line.162"></a>
<span class="sourceLineNo">163</span> throw new IllegalArgumentException("Unrecognized mapreduce implementation: " + mr + ".");<a name="line.163"></a>
<span class="sourceLineNo">164</span> }<a name="line.164"></a>
<span class="sourceLineNo">165</span><a name="line.165"></a>
<span class="sourceLineNo">166</span> return 0;<a name="line.166"></a>
<span class="sourceLineNo">167</span> }<a name="line.167"></a>
<span class="sourceLineNo">168</span><a name="line.168"></a>
<span class="sourceLineNo">169</span> @Override // Chaos Monkey is not intended to be run with this test<a name="line.169"></a>
<span class="sourceLineNo">170</span> public TableName getTablename() {<a name="line.170"></a>
<span class="sourceLineNo">171</span> return null;<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> @Override // Chaos Monkey is not inteded to be run with this test<a name="line.174"></a>
<span class="sourceLineNo">175</span> protected Set&lt;String&gt; getColumnFamilies() {<a name="line.175"></a>
<span class="sourceLineNo">176</span> return null;<a name="line.176"></a>
<span class="sourceLineNo">177</span> }<a name="line.177"></a>
<span class="sourceLineNo">178</span><a name="line.178"></a>
<span class="sourceLineNo">179</span> public static void main(String[] args) throws Exception {<a name="line.179"></a>
<span class="sourceLineNo">180</span> Configuration conf = HBaseConfiguration.create();<a name="line.180"></a>
<span class="sourceLineNo">181</span> IntegrationTestingUtility.setUseDistributedCluster(conf);<a name="line.181"></a>
<span class="sourceLineNo">182</span> int ret = ToolRunner.run(conf, new IntegrationTestTableSnapshotInputFormat(), args);<a name="line.182"></a>
<span class="sourceLineNo">183</span> System.exit(ret);<a name="line.183"></a>
<span class="sourceLineNo">184</span> }<a name="line.184"></a>
<span class="sourceLineNo">185</span><a name="line.185"></a>
<span class="sourceLineNo">186</span>}<a name="line.186"></a>
</pre>
</div>
</body>
</html>