blob: 22b96336bfee46f4cd789e742936d121f5f98c6f [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 org.apache.hadoop.conf.Configuration;<a name="line.20"></a>
<span class="sourceLineNo">021</span>import org.apache.hadoop.hbase.HBaseConfiguration;<a name="line.21"></a>
<span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.IntegrationTestingUtility;<a name="line.22"></a>
<span class="sourceLineNo">023</span>import org.apache.hadoop.hbase.regionserver.storefiletracker.StoreFileTrackerFactory;<a name="line.23"></a>
<span class="sourceLineNo">024</span>import org.apache.hadoop.hbase.testclassification.IntegrationTests;<a name="line.24"></a>
<span class="sourceLineNo">025</span>import org.apache.hadoop.util.ToolRunner;<a name="line.25"></a>
<span class="sourceLineNo">026</span>import org.junit.Test;<a name="line.26"></a>
<span class="sourceLineNo">027</span>import org.junit.experimental.categories.Category;<a name="line.27"></a>
<span class="sourceLineNo">028</span>import org.slf4j.Logger;<a name="line.28"></a>
<span class="sourceLineNo">029</span>import org.slf4j.LoggerFactory;<a name="line.29"></a>
<span class="sourceLineNo">030</span><a name="line.30"></a>
<span class="sourceLineNo">031</span>/**<a name="line.31"></a>
<span class="sourceLineNo">032</span> * Test Bulk Load and MR on a distributed cluster. With FileBased StorefileTracker enabled. It<a name="line.32"></a>
<span class="sourceLineNo">033</span> * starts an MR job that creates linked chains The format of rows is like this: Row Key -&gt; Long L:&lt;&lt;<a name="line.33"></a>
<span class="sourceLineNo">034</span> * Chain Id &gt;&gt; -&gt; Row Key of the next link in the chain S:&lt;&lt; Chain Id &gt;&gt; -&gt; The step in the chain<a name="line.34"></a>
<span class="sourceLineNo">035</span> * that his link is. D:&lt;&lt; Chain Id &gt;&gt; -&gt; Random Data. All chains start on row 0. All rk's are &gt; 0.<a name="line.35"></a>
<span class="sourceLineNo">036</span> * After creating the linked lists they are walked over using a TableMapper based Mapreduce Job.<a name="line.36"></a>
<span class="sourceLineNo">037</span> * There are a few options exposed: hbase.IntegrationTestBulkLoad.chainLength The number of rows<a name="line.37"></a>
<span class="sourceLineNo">038</span> * that will be part of each and every chain. hbase.IntegrationTestBulkLoad.numMaps The number of<a name="line.38"></a>
<span class="sourceLineNo">039</span> * mappers that will be run. Each mapper creates on linked list chain.<a name="line.39"></a>
<span class="sourceLineNo">040</span> * hbase.IntegrationTestBulkLoad.numImportRounds How many jobs will be run to create linked lists.<a name="line.40"></a>
<span class="sourceLineNo">041</span> * hbase.IntegrationTestBulkLoad.tableName The name of the table.<a name="line.41"></a>
<span class="sourceLineNo">042</span> * hbase.IntegrationTestBulkLoad.replicaCount How many region replicas to configure for the table<a name="line.42"></a>
<span class="sourceLineNo">043</span> * under test.<a name="line.43"></a>
<span class="sourceLineNo">044</span> */<a name="line.44"></a>
<span class="sourceLineNo">045</span>@Category(IntegrationTests.class)<a name="line.45"></a>
<span class="sourceLineNo">046</span>public class IntegrationTestFileBasedSFTBulkLoad extends IntegrationTestBulkLoad {<a name="line.46"></a>
<span class="sourceLineNo">047</span><a name="line.47"></a>
<span class="sourceLineNo">048</span> private static final Logger LOG =<a name="line.48"></a>
<span class="sourceLineNo">049</span> LoggerFactory.getLogger(IntegrationTestFileBasedSFTBulkLoad.class);<a name="line.49"></a>
<span class="sourceLineNo">050</span><a name="line.50"></a>
<span class="sourceLineNo">051</span> private static String NUM_MAPS_KEY = "hbase.IntegrationTestBulkLoad.numMaps";<a name="line.51"></a>
<span class="sourceLineNo">052</span> private static String NUM_IMPORT_ROUNDS_KEY = "hbase.IntegrationTestBulkLoad.numImportRounds";<a name="line.52"></a>
<span class="sourceLineNo">053</span> private static String NUM_REPLICA_COUNT_KEY = "hbase.IntegrationTestBulkLoad.replicaCount";<a name="line.53"></a>
<span class="sourceLineNo">054</span> private static int NUM_REPLICA_COUNT_DEFAULT = 1;<a name="line.54"></a>
<span class="sourceLineNo">055</span><a name="line.55"></a>
<span class="sourceLineNo">056</span> @Test<a name="line.56"></a>
<span class="sourceLineNo">057</span> public void testFileBasedSFTBulkLoad() throws Exception {<a name="line.57"></a>
<span class="sourceLineNo">058</span> super.testBulkLoad();<a name="line.58"></a>
<span class="sourceLineNo">059</span> }<a name="line.59"></a>
<span class="sourceLineNo">060</span><a name="line.60"></a>
<span class="sourceLineNo">061</span> @Override<a name="line.61"></a>
<span class="sourceLineNo">062</span> public void setUpCluster() throws Exception {<a name="line.62"></a>
<span class="sourceLineNo">063</span> util = getTestingUtil(getConf());<a name="line.63"></a>
<span class="sourceLineNo">064</span> util.getConfiguration().set(StoreFileTrackerFactory.TRACKER_IMPL,<a name="line.64"></a>
<span class="sourceLineNo">065</span> "org.apache.hadoop.hbase.regionserver.storefiletracker.FileBasedStoreFileTracker");<a name="line.65"></a>
<span class="sourceLineNo">066</span> util.initializeCluster(1);<a name="line.66"></a>
<span class="sourceLineNo">067</span> int replicaCount = getConf().getInt(NUM_REPLICA_COUNT_KEY, NUM_REPLICA_COUNT_DEFAULT);<a name="line.67"></a>
<span class="sourceLineNo">068</span> if (LOG.isDebugEnabled() &amp;&amp; replicaCount != NUM_REPLICA_COUNT_DEFAULT) {<a name="line.68"></a>
<span class="sourceLineNo">069</span> LOG.debug("Region Replicas enabled: " + replicaCount);<a name="line.69"></a>
<span class="sourceLineNo">070</span> }<a name="line.70"></a>
<span class="sourceLineNo">071</span><a name="line.71"></a>
<span class="sourceLineNo">072</span> // Scale this up on a real cluster<a name="line.72"></a>
<span class="sourceLineNo">073</span> if (util.isDistributedCluster()) {<a name="line.73"></a>
<span class="sourceLineNo">074</span> util.getConfiguration().setIfUnset(NUM_MAPS_KEY,<a name="line.74"></a>
<span class="sourceLineNo">075</span> Integer.toString(util.getAdmin().getRegionServers().size() * 10));<a name="line.75"></a>
<span class="sourceLineNo">076</span> util.getConfiguration().setIfUnset(NUM_IMPORT_ROUNDS_KEY, "5");<a name="line.76"></a>
<span class="sourceLineNo">077</span> } else {<a name="line.77"></a>
<span class="sourceLineNo">078</span> util.startMiniMapReduceCluster();<a name="line.78"></a>
<span class="sourceLineNo">079</span> }<a name="line.79"></a>
<span class="sourceLineNo">080</span> }<a name="line.80"></a>
<span class="sourceLineNo">081</span><a name="line.81"></a>
<span class="sourceLineNo">082</span> public static void main(String[] args) throws Exception {<a name="line.82"></a>
<span class="sourceLineNo">083</span> Configuration conf = HBaseConfiguration.create();<a name="line.83"></a>
<span class="sourceLineNo">084</span> IntegrationTestingUtility.setUseDistributedCluster(conf);<a name="line.84"></a>
<span class="sourceLineNo">085</span> int status = ToolRunner.run(conf, new IntegrationTestFileBasedSFTBulkLoad(), args);<a name="line.85"></a>
<span class="sourceLineNo">086</span> System.exit(status);<a name="line.86"></a>
<span class="sourceLineNo">087</span> }<a name="line.87"></a>
<span class="sourceLineNo">088</span>}<a name="line.88"></a>
</pre>
</div>
</body>
</html>