http://git-wip-us.apache.org/repos/asf/hbase-site/blob/3b220124/devapidocs/src-html/org/apache/hadoop/hbase/tool/LoadIncrementalHFiles.BulkHFileVisitor.html ---------------------------------------------------------------------- diff --git a/devapidocs/src-html/org/apache/hadoop/hbase/tool/LoadIncrementalHFiles.BulkHFileVisitor.html b/devapidocs/src-html/org/apache/hadoop/hbase/tool/LoadIncrementalHFiles.BulkHFileVisitor.html new file mode 100644 index 0000000..79a160d --- /dev/null +++ b/devapidocs/src-html/org/apache/hadoop/hbase/tool/LoadIncrementalHFiles.BulkHFileVisitor.html @@ -0,0 +1,1323 @@ +<!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.tool;<a name="line.18"></a> +<span class="sourceLineNo">019</span><a name="line.19"></a> +<span class="sourceLineNo">020</span>import static java.lang.String.format;<a name="line.20"></a> +<span class="sourceLineNo">021</span><a name="line.21"></a> +<span class="sourceLineNo">022</span>import java.io.FileNotFoundException;<a name="line.22"></a> +<span class="sourceLineNo">023</span>import java.io.IOException;<a name="line.23"></a> +<span class="sourceLineNo">024</span>import java.io.InterruptedIOException;<a name="line.24"></a> +<span class="sourceLineNo">025</span>import java.nio.ByteBuffer;<a name="line.25"></a> +<span class="sourceLineNo">026</span>import java.util.ArrayDeque;<a name="line.26"></a> +<span class="sourceLineNo">027</span>import java.util.ArrayList;<a name="line.27"></a> +<span class="sourceLineNo">028</span>import java.util.Arrays;<a name="line.28"></a> +<span class="sourceLineNo">029</span>import java.util.Collection;<a name="line.29"></a> +<span class="sourceLineNo">030</span>import java.util.Collections;<a name="line.30"></a> +<span class="sourceLineNo">031</span>import java.util.Deque;<a name="line.31"></a> +<span class="sourceLineNo">032</span>import java.util.HashMap;<a name="line.32"></a> +<span class="sourceLineNo">033</span>import java.util.HashSet;<a name="line.33"></a> +<span class="sourceLineNo">034</span>import java.util.List;<a name="line.34"></a> +<span class="sourceLineNo">035</span>import java.util.Map;<a name="line.35"></a> +<span class="sourceLineNo">036</span>import java.util.Map.Entry;<a name="line.36"></a> +<span class="sourceLineNo">037</span>import java.util.Set;<a name="line.37"></a> +<span class="sourceLineNo">038</span>import java.util.SortedMap;<a name="line.38"></a> +<span class="sourceLineNo">039</span>import java.util.TreeMap;<a name="line.39"></a> +<span class="sourceLineNo">040</span>import java.util.UUID;<a name="line.40"></a> +<span class="sourceLineNo">041</span>import java.util.concurrent.Callable;<a name="line.41"></a> +<span class="sourceLineNo">042</span>import java.util.concurrent.ExecutionException;<a name="line.42"></a> +<span class="sourceLineNo">043</span>import java.util.concurrent.ExecutorService;<a name="line.43"></a> +<span class="sourceLineNo">044</span>import java.util.concurrent.Future;<a name="line.44"></a> +<span class="sourceLineNo">045</span>import java.util.concurrent.LinkedBlockingQueue;<a name="line.45"></a> +<span class="sourceLineNo">046</span>import java.util.concurrent.ThreadPoolExecutor;<a name="line.46"></a> +<span class="sourceLineNo">047</span>import java.util.concurrent.TimeUnit;<a name="line.47"></a> +<span class="sourceLineNo">048</span>import java.util.concurrent.atomic.AtomicInteger;<a name="line.48"></a> +<span class="sourceLineNo">049</span>import java.util.stream.Collectors;<a name="line.49"></a> +<span class="sourceLineNo">050</span><a name="line.50"></a> +<span class="sourceLineNo">051</span>import org.apache.commons.lang.mutable.MutableInt;<a name="line.51"></a> +<span class="sourceLineNo">052</span>import org.apache.commons.logging.Log;<a name="line.52"></a> +<span class="sourceLineNo">053</span>import org.apache.commons.logging.LogFactory;<a name="line.53"></a> +<span class="sourceLineNo">054</span>import org.apache.hadoop.conf.Configuration;<a name="line.54"></a> +<span class="sourceLineNo">055</span>import org.apache.hadoop.conf.Configured;<a name="line.55"></a> +<span class="sourceLineNo">056</span>import org.apache.hadoop.fs.FileStatus;<a name="line.56"></a> +<span class="sourceLineNo">057</span>import org.apache.hadoop.fs.FileSystem;<a name="line.57"></a> +<span class="sourceLineNo">058</span>import org.apache.hadoop.fs.Path;<a name="line.58"></a> +<span class="sourceLineNo">059</span>import org.apache.hadoop.fs.permission.FsPermission;<a name="line.59"></a> +<span class="sourceLineNo">060</span>import org.apache.hadoop.hbase.HBaseConfiguration;<a name="line.60"></a> +<span class="sourceLineNo">061</span>import org.apache.hadoop.hbase.HConstants;<a name="line.61"></a> +<span class="sourceLineNo">062</span>import org.apache.hadoop.hbase.TableName;<a name="line.62"></a> +<span class="sourceLineNo">063</span>import org.apache.hadoop.hbase.TableNotFoundException;<a name="line.63"></a> +<span class="sourceLineNo">064</span>import org.apache.hadoop.hbase.classification.InterfaceAudience;<a name="line.64"></a> +<span class="sourceLineNo">065</span>import org.apache.hadoop.hbase.client.Admin;<a name="line.65"></a> +<span class="sourceLineNo">066</span>import org.apache.hadoop.hbase.client.ClientServiceCallable;<a name="line.66"></a> +<span class="sourceLineNo">067</span>import org.apache.hadoop.hbase.client.ColumnFamilyDescriptor;<a name="line.67"></a> +<span class="sourceLineNo">068</span>import org.apache.hadoop.hbase.client.ColumnFamilyDescriptorBuilder;<a name="line.68"></a> +<span class="sourceLineNo">069</span>import org.apache.hadoop.hbase.client.Connection;<a name="line.69"></a> +<span class="sourceLineNo">070</span>import org.apache.hadoop.hbase.client.ConnectionFactory;<a name="line.70"></a> +<span class="sourceLineNo">071</span>import org.apache.hadoop.hbase.client.RegionLocator;<a name="line.71"></a> +<span class="sourceLineNo">072</span>import org.apache.hadoop.hbase.client.RpcRetryingCallerFactory;<a name="line.72"></a> +<span class="sourceLineNo">073</span>import org.apache.hadoop.hbase.client.SecureBulkLoadClient;<a name="line.73"></a> +<span class="sourceLineNo">074</span>import org.apache.hadoop.hbase.client.Table;<a name="line.74"></a> +<span class="sourceLineNo">075</span>import org.apache.hadoop.hbase.client.TableDescriptorBuilder;<a name="line.75"></a> +<span class="sourceLineNo">076</span>import org.apache.hadoop.hbase.coprocessor.CoprocessorHost;<a name="line.76"></a> +<span class="sourceLineNo">077</span>import org.apache.hadoop.hbase.io.HFileLink;<a name="line.77"></a> +<span class="sourceLineNo">078</span>import org.apache.hadoop.hbase.io.HalfStoreFileReader;<a name="line.78"></a> +<span class="sourceLineNo">079</span>import org.apache.hadoop.hbase.io.Reference;<a name="line.79"></a> +<span class="sourceLineNo">080</span>import org.apache.hadoop.hbase.io.compress.Compression.Algorithm;<a name="line.80"></a> +<span class="sourceLineNo">081</span>import org.apache.hadoop.hbase.io.hfile.CacheConfig;<a name="line.81"></a> +<span class="sourceLineNo">082</span>import org.apache.hadoop.hbase.io.hfile.HFile;<a name="line.82"></a> +<span class="sourceLineNo">083</span>import org.apache.hadoop.hbase.io.hfile.HFileContext;<a name="line.83"></a> +<span class="sourceLineNo">084</span>import org.apache.hadoop.hbase.io.hfile.HFileContextBuilder;<a name="line.84"></a> +<span class="sourceLineNo">085</span>import org.apache.hadoop.hbase.io.hfile.HFileDataBlockEncoder;<a name="line.85"></a> +<span class="sourceLineNo">086</span>import org.apache.hadoop.hbase.io.hfile.HFileScanner;<a name="line.86"></a> +<span class="sourceLineNo">087</span>import org.apache.hadoop.hbase.ipc.RpcControllerFactory;<a name="line.87"></a> +<span class="sourceLineNo">088</span>import org.apache.hadoop.hbase.regionserver.BloomType;<a name="line.88"></a> +<span class="sourceLineNo">089</span>import org.apache.hadoop.hbase.regionserver.HStore;<a name="line.89"></a> +<span class="sourceLineNo">090</span>import org.apache.hadoop.hbase.regionserver.StoreFileInfo;<a name="line.90"></a> +<span class="sourceLineNo">091</span>import org.apache.hadoop.hbase.regionserver.StoreFileWriter;<a name="line.91"></a> +<span class="sourceLineNo">092</span>import org.apache.hadoop.hbase.security.UserProvider;<a name="line.92"></a> +<span class="sourceLineNo">093</span>import org.apache.hadoop.hbase.security.token.FsDelegationToken;<a name="line.93"></a> +<span class="sourceLineNo">094</span>import org.apache.hadoop.hbase.shaded.com.google.common.annotations.VisibleForTesting;<a name="line.94"></a> +<span class="sourceLineNo">095</span>import org.apache.hadoop.hbase.shaded.com.google.common.collect.HashMultimap;<a name="line.95"></a> +<span class="sourceLineNo">096</span>import org.apache.hadoop.hbase.shaded.com.google.common.collect.Multimap;<a name="line.96"></a> +<span class="sourceLineNo">097</span>import org.apache.hadoop.hbase.shaded.com.google.common.collect.Multimaps;<a name="line.97"></a> +<span class="sourceLineNo">098</span>import org.apache.hadoop.hbase.shaded.com.google.common.util.concurrent.ThreadFactoryBuilder;<a name="line.98"></a> +<span class="sourceLineNo">099</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.99"></a> +<span class="sourceLineNo">100</span>import org.apache.hadoop.hbase.util.FSHDFSUtils;<a name="line.100"></a> +<span class="sourceLineNo">101</span>import org.apache.hadoop.hbase.util.Pair;<a name="line.101"></a> +<span class="sourceLineNo">102</span>import org.apache.hadoop.util.Tool;<a name="line.102"></a> +<span class="sourceLineNo">103</span>import org.apache.hadoop.util.ToolRunner;<a name="line.103"></a> +<span class="sourceLineNo">104</span><a name="line.104"></a> +<span class="sourceLineNo">105</span>/**<a name="line.105"></a> +<span class="sourceLineNo">106</span> * Tool to load the output of HFileOutputFormat into an existing table.<a name="line.106"></a> +<span class="sourceLineNo">107</span> */<a name="line.107"></a> +<span class="sourceLineNo">108</span>@InterfaceAudience.Public<a name="line.108"></a> +<span class="sourceLineNo">109</span>public class LoadIncrementalHFiles extends Configured implements Tool {<a name="line.109"></a> +<span class="sourceLineNo">110</span><a name="line.110"></a> +<span class="sourceLineNo">111</span> private static final Log LOG = LogFactory.getLog(LoadIncrementalHFiles.class);<a name="line.111"></a> +<span class="sourceLineNo">112</span><a name="line.112"></a> +<span class="sourceLineNo">113</span> public static final String NAME = "completebulkload";<a name="line.113"></a> +<span class="sourceLineNo">114</span> static final String RETRY_ON_IO_EXCEPTION = "hbase.bulkload.retries.retryOnIOException";<a name="line.114"></a> +<span class="sourceLineNo">115</span> public static final String MAX_FILES_PER_REGION_PER_FAMILY =<a name="line.115"></a> +<span class="sourceLineNo">116</span> "hbase.mapreduce.bulkload.max.hfiles.perRegion.perFamily";<a name="line.116"></a> +<span class="sourceLineNo">117</span> private static final String ASSIGN_SEQ_IDS = "hbase.mapreduce.bulkload.assign.sequenceNumbers";<a name="line.117"></a> +<span class="sourceLineNo">118</span> public final static String CREATE_TABLE_CONF_KEY = "create.table";<a name="line.118"></a> +<span class="sourceLineNo">119</span> public final static String IGNORE_UNMATCHED_CF_CONF_KEY = "ignore.unmatched.families";<a name="line.119"></a> +<span class="sourceLineNo">120</span> public final static String ALWAYS_COPY_FILES = "always.copy.files";<a name="line.120"></a> +<span class="sourceLineNo">121</span><a name="line.121"></a> +<span class="sourceLineNo">122</span> // We use a '.' prefix which is ignored when walking directory trees<a name="line.122"></a> +<span class="sourceLineNo">123</span> // above. It is invalid family name.<a name="line.123"></a> +<span class="sourceLineNo">124</span> static final String TMP_DIR = ".tmp";<a name="line.124"></a> +<span class="sourceLineNo">125</span><a name="line.125"></a> +<span class="sourceLineNo">126</span> private final int maxFilesPerRegionPerFamily;<a name="line.126"></a> +<span class="sourceLineNo">127</span> private final boolean assignSeqIds;<a name="line.127"></a> +<span class="sourceLineNo">128</span><a name="line.128"></a> +<span class="sourceLineNo">129</span> // Source delegation token<a name="line.129"></a> +<span class="sourceLineNo">130</span> private final FsDelegationToken fsDelegationToken;<a name="line.130"></a> +<span class="sourceLineNo">131</span> private final UserProvider userProvider;<a name="line.131"></a> +<span class="sourceLineNo">132</span> private final int nrThreads;<a name="line.132"></a> +<span class="sourceLineNo">133</span> private final RpcControllerFactory rpcControllerFactory;<a name="line.133"></a> +<span class="sourceLineNo">134</span><a name="line.134"></a> +<span class="sourceLineNo">135</span> private String bulkToken;<a name="line.135"></a> +<span class="sourceLineNo">136</span><a name="line.136"></a> +<span class="sourceLineNo">137</span> /**<a name="line.137"></a> +<span class="sourceLineNo">138</span> * Represents an HFile waiting to be loaded. An queue is used in this class in order to support<a name="line.138"></a> +<span class="sourceLineNo">139</span> * the case where a region has split during the process of the load. When this happens, the HFile<a name="line.139"></a> +<span class="sourceLineNo">140</span> * is split into two physical parts across the new region boundary, and each part is added back<a name="line.140"></a> +<span class="sourceLineNo">141</span> * into the queue. The import process finishes when the queue is empty.<a name="line.141"></a> +<span class="sourceLineNo">142</span> */<a name="line.142"></a> +<span class="sourceLineNo">143</span> @InterfaceAudience.Public<a name="line.143"></a> +<span class="sourceLineNo">144</span> public static class LoadQueueItem {<a name="line.144"></a> +<span class="sourceLineNo">145</span> private final byte[] family;<a name="line.145"></a> +<span class="sourceLineNo">146</span> private final Path hfilePath;<a name="line.146"></a> +<span class="sourceLineNo">147</span><a name="line.147"></a> +<span class="sourceLineNo">148</span> public LoadQueueItem(byte[] family, Path hfilePath) {<a name="line.148"></a> +<span class="sourceLineNo">149</span> this.family = family;<a name="line.149"></a> +<span class="sourceLineNo">150</span> this.hfilePath = hfilePath;<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> @Override<a name="line.153"></a> +<span class="sourceLineNo">154</span> public String toString() {<a name="line.154"></a> +<span class="sourceLineNo">155</span> return "family:" + Bytes.toString(family) + " path:" + hfilePath.toString();<a name="line.155"></a> +<span class="sourceLineNo">156</span> }<a name="line.156"></a> +<span class="sourceLineNo">157</span><a name="line.157"></a> +<span class="sourceLineNo">158</span> public byte[] getFamily() {<a name="line.158"></a> +<span class="sourceLineNo">159</span> return family;<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> public Path getFilePath() {<a name="line.162"></a> +<span class="sourceLineNo">163</span> return hfilePath;<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><a name="line.166"></a> +<span class="sourceLineNo">167</span> public LoadIncrementalHFiles(Configuration conf) {<a name="line.167"></a> +<span class="sourceLineNo">168</span> // make a copy, just to be sure we're not overriding someone else's config<a name="line.168"></a> +<span class="sourceLineNo">169</span> super(HBaseConfiguration.create(conf));<a name="line.169"></a> +<span class="sourceLineNo">170</span> conf = getConf();<a name="line.170"></a> +<span class="sourceLineNo">171</span> // disable blockcache for tool invocation, see HBASE-10500<a name="line.171"></a> +<span class="sourceLineNo">172</span> conf.setFloat(HConstants.HFILE_BLOCK_CACHE_SIZE_KEY, 0);<a name="line.172"></a> +<span class="sourceLineNo">173</span> userProvider = UserProvider.instantiate(conf);<a name="line.173"></a> +<span class="sourceLineNo">174</span> fsDelegationToken = new FsDelegationToken(userProvider, "renewer");<a name="line.174"></a> +<span class="sourceLineNo">175</span> assignSeqIds = conf.getBoolean(ASSIGN_SEQ_IDS, true);<a name="line.175"></a> +<span class="sourceLineNo">176</span> maxFilesPerRegionPerFamily = conf.getInt(MAX_FILES_PER_REGION_PER_FAMILY, 32);<a name="line.176"></a> +<span class="sourceLineNo">177</span> nrThreads = conf.getInt("hbase.loadincremental.threads.max",<a name="line.177"></a> +<span class="sourceLineNo">178</span> Runtime.getRuntime().availableProcessors());<a name="line.178"></a> +<span class="sourceLineNo">179</span> rpcControllerFactory = new RpcControllerFactory(conf);<a name="line.179"></a> +<span class="sourceLineNo">180</span> }<a name="line.180"></a> +<span class="sourceLineNo">181</span><a name="line.181"></a> +<span class="sourceLineNo">182</span> private void usage() {<a name="line.182"></a> +<span class="sourceLineNo">183</span> System.err.println("usage: " + NAME + " /path/to/hfileoutputformat-output tablename" + "\n -D" +<a name="line.183"></a> +<span class="sourceLineNo">184</span> CREATE_TABLE_CONF_KEY + "=no - can be used to avoid creation of table by this tool\n" +<a name="line.184"></a> +<span class="sourceLineNo">185</span> " Note: if you set this to 'no', then the target table must already exist in HBase\n -D" +<a name="line.185"></a> +<span class="sourceLineNo">186</span> IGNORE_UNMATCHED_CF_CONF_KEY + "=yes - can be used to ignore unmatched column families\n" +<a name="line.186"></a> +<span class="sourceLineNo">187</span> "\n");<a name="line.187"></a> +<span class="sourceLineNo">188</span> }<a name="line.188"></a> +<span class="sourceLineNo">189</span><a name="line.189"></a> +<span class="sourceLineNo">190</span> /**<a name="line.190"></a> +<span class="sourceLineNo">191</span> * Prepare a collection of {@link LoadQueueItem} from list of source hfiles contained in the<a name="line.191"></a> +<span class="sourceLineNo">192</span> * passed directory and validates whether the prepared queue has all the valid table column<a name="line.192"></a> +<span class="sourceLineNo">193</span> * families in it.<a name="line.193"></a> +<span class="sourceLineNo">194</span> * @param hfilesDir directory containing list of hfiles to be loaded into the table<a name="line.194"></a> +<span class="sourceLineNo">195</span> * @param table table to which hfiles should be loaded<a name="line.195"></a> +<span class="sourceLineNo">196</span> * @param queue queue which needs to be loaded into the table<a name="line.196"></a> +<span class="sourceLineNo">197</span> * @param validateHFile if true hfiles will be validated for its format<a name="line.197"></a> +<span class="sourceLineNo">198</span> * @throws IOException If any I/O or network error occurred<a name="line.198"></a> +<span class="sourceLineNo">199</span> */<a name="line.199"></a> +<span class="sourceLineNo">200</span> public void prepareHFileQueue(Path hfilesDir, Table table, Deque<LoadQueueItem> queue,<a name="line.200"></a> +<span class="sourceLineNo">201</span> boolean validateHFile) throws IOException {<a name="line.201"></a> +<span class="sourceLineNo">202</span> prepareHFileQueue(hfilesDir, table, queue, validateHFile, false);<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> /**<a name="line.205"></a> +<span class="sourceLineNo">206</span> * Prepare a collection of {@link LoadQueueItem} from list of source hfiles contained in the<a name="line.206"></a> +<span class="sourceLineNo">207</span> * passed directory and validates whether the prepared queue has all the valid table column<a name="line.207"></a> +<span class="sourceLineNo">208</span> * families in it.<a name="line.208"></a> +<span class="sourceLineNo">209</span> * @param hfilesDir directory containing list of hfiles to be loaded into the table<a name="line.209"></a> +<span class="sourceLineNo">210</span> * @param table table to which hfiles should be loaded<a name="line.210"></a> +<span class="sourceLineNo">211</span> * @param queue queue which needs to be loaded into the table<a name="line.211"></a> +<span class="sourceLineNo">212</span> * @param validateHFile if true hfiles will be validated for its format<a name="line.212"></a> +<span class="sourceLineNo">213</span> * @param silence true to ignore unmatched column families<a name="line.213"></a> +<span class="sourceLineNo">214</span> * @throws IOException If any I/O or network error occurred<a name="line.214"></a> +<span class="sourceLineNo">215</span> */<a name="line.215"></a> +<span class="sourceLineNo">216</span> public void prepareHFileQueue(Path hfilesDir, Table table, Deque<LoadQueueItem> queue,<a name="line.216"></a> +<span class="sourceLineNo">217</span> boolean validateHFile, boolean silence) throws IOException {<a name="line.217"></a> +<span class="sourceLineNo">218</span> discoverLoadQueue(queue, hfilesDir, validateHFile);<a name="line.218"></a> +<span class="sourceLineNo">219</span> validateFamiliesInHFiles(table, queue, silence);<a name="line.219"></a> +<span class="sourceLineNo">220</span> }<a name="line.220"></a> +<span class="sourceLineNo">221</span><a name="line.221"></a> +<span class="sourceLineNo">222</span> /**<a name="line.222"></a> +<span class="sourceLineNo">223</span> * Prepare a collection of {@link LoadQueueItem} from list of source hfiles contained in the<a name="line.223"></a> +<span class="sourceLineNo">224</span> * passed directory and validates whether the prepared queue has all the valid table column<a name="line.224"></a> +<span class="sourceLineNo">225</span> * families in it.<a name="line.225"></a> +<span class="sourceLineNo">226</span> * @param map map of family to List of hfiles<a name="line.226"></a> +<span class="sourceLineNo">227</span> * @param table table to which hfiles should be loaded<a name="line.227"></a> +<span class="sourceLineNo">228</span> * @param queue queue which needs to be loaded into the table<a name="line.228"></a> +<span class="sourceLineNo">229</span> * @param silence true to ignore unmatched column families<a name="line.229"></a> +<span class="sourceLineNo">230</span> * @throws IOException If any I/O or network error occurred<a name="line.230"></a> +<span class="sourceLineNo">231</span> */<a name="line.231"></a> +<span class="sourceLineNo">232</span> public void prepareHFileQueue(Map<byte[], List<Path>> map, Table table,<a name="line.232"></a> +<span class="sourceLineNo">233</span> Deque<LoadQueueItem> queue, boolean silence) throws IOException {<a name="line.233"></a> +<span class="sourceLineNo">234</span> populateLoadQueue(queue, map);<a name="line.234"></a> +<span class="sourceLineNo">235</span> validateFamiliesInHFiles(table, queue, silence);<a name="line.235"></a> +<span class="sourceLineNo">236</span> }<a name="line.236"></a> +<span class="sourceLineNo">237</span><a name="line.237"></a> +<span class="sourceLineNo">238</span> /**<a name="line.238"></a> +<span class="sourceLineNo">239</span> * Perform a bulk load of the given directory into the given pre-existing table. This method is<a name="line.239"></a> +<span class="sourceLineNo">240</span> * not threadsafe.<a name="line.240"></a> +<span class="sourceLineNo">241</span> * @param hfofDir the directory that was provided as the output path of a job using<a name="line.241"></a> +<span class="sourceLineNo">242</span> * HFileOutputFormat<a name="line.242"></a> +<span class="sourceLineNo">243</span> * @param admin the Admin<a name="line.243"></a> +<span class="sourceLineNo">244</span> * @param table the table to load into<a name="line.244"></a> +<span class="sourceLineNo">245</span> * @param regionLocator region locator<a name="line.245"></a> +<span class="sourceLineNo">246</span> * @throws TableNotFoundException if table does not yet exist<a name="line.246"></a> +<span class="sourceLineNo">247</span> */<a name="line.247"></a> +<span class="sourceLineNo">248</span> public Map<LoadQueueItem, ByteBuffer> doBulkLoad(Path hfofDir, final Admin admin, Table table,<a name="line.248"></a> +<span class="sourceLineNo">249</span> RegionLocator regionLocator) throws TableNotFoundException, IOException {<a name="line.249"></a> +<span class="sourceLineNo">250</span> return doBulkLoad(hfofDir, admin, table, regionLocator, false, false);<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> /**<a name="line.253"></a> +<span class="sourceLineNo">254</span> * Perform a bulk load of the given directory into the given pre-existing table. This method is<a name="line.254"></a> +<span class="sourceLineNo">255</span> * not threadsafe.<a name="line.255"></a> +<span class="sourceLineNo">256</span> * @param map map of family to List of hfiles<a name="line.256"></a> +<span class="sourceLineNo">257</span> * @param admin the Admin<a name="line.257"></a> +<span class="sourceLineNo">258</span> * @param table the table to load into<a name="line.258"></a> +<span class="sourceLineNo">259</span> * @param regionLocator region locator<a name="line.259"></a> +<span class="sourceLineNo">260</span> * @param silence true to ignore unmatched column families<a name="line.260"></a> +<span class="sourceLineNo">261</span> * @param copyFile always copy hfiles if true<a name="line.261"></a> +<span class="sourceLineNo">262</span> * @throws TableNotFoundException if table does not yet exist<a name="line.262"></a> +<span class="sourceLineNo">263</span> */<a name="line.263"></a> +<span class="sourceLineNo">264</span> public Map<LoadQueueItem, ByteBuffer> doBulkLoad(Map<byte[], List<Path>> map, final Admin admin,<a name="line.264"></a> +<span class="sourceLineNo">265</span> Table table, RegionLocator regionLocator, boolean silence, boolean copyFile)<a name="line.265"></a> +<span class="sourceLineNo">266</span> throws TableNotFoundException, IOException {<a name="line.266"></a> +<span class="sourceLineNo">267</span> if (!admin.isTableAvailable(regionLocator.getName())) {<a name="line.267"></a> +<span class="sourceLineNo">268</span> throw new TableNotFoundException("Table " + table.getName() + " is not currently available.");<a name="line.268"></a> +<span class="sourceLineNo">269</span> }<a name="line.269"></a> +<span class="sourceLineNo">270</span> // LQI queue does not need to be threadsafe -- all operations on this queue<a name="line.270"></a> +<span class="sourceLineNo">271</span> // happen in this thread<a name="line.271"></a> +<span class="sourceLineNo">272</span> Deque<LoadQueueItem> queue = new ArrayDeque<>();<a name="line.272"></a> +<span class="sourceLineNo">273</span> ExecutorService pool = null;<a name="line.273"></a> +<span class="sourceLineNo">274</span> SecureBulkLoadClient secureClient = null;<a name="line.274"></a> +<span class="sourceLineNo">275</span> try {<a name="line.275"></a> +<span class="sourceLineNo">276</span> prepareHFileQueue(map, table, queue, silence);<a name="line.276"></a> +<span class="sourceLineNo">277</span> if (queue.isEmpty()) {<a name="line.277"></a> +<span class="sourceLineNo">278</span> LOG.warn("Bulk load operation did not get any files to load");<a name="line.278"></a> +<span class="sourceLineNo">279</span> return Collections.emptyMap();<a name="line.279"></a> +<span class="sourceLineNo">280</span> }<a name="line.280"></a> +<span class="sourceLineNo">281</span> pool = createExecutorService();<a name="line.281"></a> +<span class="sourceLineNo">282</span> secureClient = new SecureBulkLoadClient(table.getConfiguration(), table);<a name="line.282"></a> +<span class="sourceLineNo">283</span> return performBulkLoad(admin, table, regionLocator, queue, pool, secureClient, copyFile);<a name="line.283"></a> +<span class="sourceLineNo">284</span> } finally {<a name="line.284"></a> +<span class="sourceLineNo">285</span> cleanup(admin, queue, pool, secureClient);<a name="line.285"></a> +<span class="sourceLineNo">286</span> }<a name="line.286"></a> +<span class="sourceLineNo">287</span> }<a name="line.287"></a> +<span class="sourceLineNo">288</span><a name="line.288"></a> +<span class="sourceLineNo">289</span> /**<a name="line.289"></a> +<span class="sourceLineNo">290</span> * Perform a bulk load of the given directory into the given pre-existing table. This method is<a name="line.290"></a> +<span class="sourceLineNo">291</span> * not threadsafe.<a name="line.291"></a> +<span class="sourceLineNo">292</span> * @param hfofDir the directory that was provided as the output path of a job using<a name="line.292"></a> +<span class="sourceLineNo">293</span> * HFileOutputFormat<a name="line.293"></a> +<span class="sourceLineNo">294</span> * @param admin the Admin<a name="line.294"></a> +<span class="sourceLineNo">295</span> * @param table the table to load into<a name="line.295"></a> +<span class="sourceLineNo">296</span> * @param regionLocator region locator<a name="line.296"></a> +<span class="sourceLineNo">297</span> * @param silence true to ignore unmatched column families<a name="line.297"></a> +<span class="sourceLineNo">298</span> * @param copyFile always copy hfiles if true<a name="line.298"></a> +<span class="sourceLineNo">299</span> * @throws TableNotFoundException if table does not yet exist<a name="line.299"></a> +<span class="sourceLineNo">300</span> */<a name="line.300"></a> +<span class="sourceLineNo">301</span> public Map<LoadQueueItem, ByteBuffer> doBulkLoad(Path hfofDir, final Admin admin, Table table,<a name="line.301"></a> +<span class="sourceLineNo">302</span> RegionLocator regionLocator, boolean silence, boolean copyFile)<a name="line.302"></a> +<span class="sourceLineNo">303</span> throws TableNotFoundException, IOException {<a name="line.303"></a> +<span class="sourceLineNo">304</span> if (!admin.isTableAvailable(regionLocator.getName())) {<a name="line.304"></a> +<span class="sourceLineNo">305</span> throw new TableNotFoundException("Table " + table.getName() + " is not currently available.");<a name="line.305"></a> +<span class="sourceLineNo">306</span> }<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> * Checking hfile format is a time-consuming operation, we should have an option to skip this<a name="line.309"></a> +<span class="sourceLineNo">310</span> * step when bulkloading millions of HFiles. See HBASE-13985.<a name="line.310"></a> +<span class="sourceLineNo">311</span> */<a name="line.311"></a> +<span class="sourceLineNo">312</span> boolean validateHFile = getConf().getBoolean("hbase.loadincremental.validate.hfile", true);<a name="line.312"></a> +<span class="sourceLineNo">313</span> if (!validateHFile) {<a name="line.313"></a> +<span class="sourceLineNo">314</span> LOG.warn("You are skipping HFiles validation, it might cause some data loss if files " +<a name="line.314"></a> +<span class="sourceLineNo">315</span> "are not correct. If you fail to read data from your table after using this " +<a name="line.315"></a> +<span class="sourceLineNo">316</span> "option, consider removing the files and bulkload again without this option. " +<a name="line.316"></a> +<span class="sourceLineNo">317</span> "See HBASE-13985");<a name="line.317"></a> +<span class="sourceLineNo">318</span> }<a name="line.318"></a> +<span class="sourceLineNo">319</span> // LQI queue does not need to be threadsafe -- all operations on this queue<a name="line.319"></a> +<span class="sourceLineNo">320</span> // happen in this thread<a name="line.320"></a> +<span class="sourceLineNo">321</span> Deque<LoadQueueItem> queue = new ArrayDeque<>();<a name="line.321"></a> +<span class="sourceLineNo">322</span> ExecutorService pool = null;<a name="line.322"></a> +<span class="sourceLineNo">323</span> SecureBulkLoadClient secureClient = null;<a name="line.323"></a> +<span class="sourceLineNo">324</span> try {<a name="line.324"></a> +<span class="sourceLineNo">325</span> prepareHFileQueue(hfofDir, table, queue, validateHFile, silence);<a name="line.325"></a> +<span class="sourceLineNo">326</span><a name="line.326"></a> +<span class="sourceLineNo">327</span> if (queue.isEmpty()) {<a name="line.327"></a> +<span class="sourceLineNo">328</span> LOG.warn(<a name="line.328"></a> +<span class="sourceLineNo">329</span> "Bulk load operation did not find any files to load in " + "directory " + hfofDir != null<a name="line.329"></a> +<span class="sourceLineNo">330</span> ? hfofDir.toUri()<a name="line.330"></a> +<span class="sourceLineNo">331</span> : "" + ". Does it contain files in " +<a name="line.331"></a> +<span class="sourceLineNo">332</span> "subdirectories that correspond to column family names?");<a name="line.332"></a> +<span class="sourceLineNo">333</span> return Collections.emptyMap();<a name="line.333"></a> +<span class="sourceLineNo">334</span> }<a name="line.334"></a> +<span class="sourceLineNo">335</span> pool = createExecutorService();<a name="line.335"></a> +<span class="sourceLineNo">336</span> secureClient = new SecureBulkLoadClient(table.getConfiguration(), table);<a name="line.336"></a> +<span class="sourceLineNo">337</span> return performBulkLoad(admin, table, regionLocator, queue, pool, secureClient, copyFile);<a name="line.337"></a> +<span class="sourceLineNo">338</span> } finally {<a name="line.338"></a> +<span class="sourceLineNo">339</span> cleanup(admin, queue, pool, secureClient);<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><a name="line.342"></a> +<span class="sourceLineNo">343</span> /**<a name="line.343"></a> +<span class="sourceLineNo">344</span> * Used by the replication sink to load the hfiles from the source cluster. It does the following,<a name="line.344"></a> +<span class="sourceLineNo">345</span> * <ol><a name="line.345"></a> +<span class="sourceLineNo">346</span> * <li>LoadIncrementalHFiles#groupOrSplitPhase(Table, ExecutorService, Deque, Pair)}</li><a name="line.346"></a> +<span class="sourceLineNo">347</span> * <li>LoadIncrementalHFiles#bulkLoadPhase(Table, Connection, ExecutorService, Deque, Multimap)<a name="line.347"></a> +<span class="sourceLineNo">348</span> * </li><a name="line.348"></a> +<span class="sourceLineNo">349</span> * </ol><a name="line.349"></a> +<span class="sourceLineNo">350</span> * @param table Table to which these hfiles should be loaded to<a name="line.350"></a> +<span class="sourceLineNo">351</span> * @param conn Connection to use<a name="line.351"></a> +<span class="sourceLineNo">352</span> * @param queue {@link LoadQueueItem} has hfiles yet to be loaded<a name="line.352"></a> +<span class="sourceLineNo">353</span> * @param startEndKeys starting and ending row keys of the region<a name="line.353"></a> +<span class="sourceLineNo">354</span> */<a name="line.354"></a> +<span class="sourceLineNo">355</span> public void loadHFileQueue(Table table, Connection conn, Deque<LoadQueueItem> queue,<a name="line.355"></a> +<span class="sourceLineNo">356</span> Pair<byte[][], byte[][]> startEndKeys) throws IOException {<a name="line.356"></a> +<span class="sourceLineNo">357</span> loadHFileQueue(table, conn, queue, startEndKeys, false);<a name="line.357"></a> +<span class="sourceLineNo">358</span> }<a name="line.358"></a> +<span class="sourceLineNo">359</span><a name="line.359"></a> +<span class="sourceLineNo">360</span> /**<a name="line.360"></a> +<span class="sourceLineNo">361</span> * Used by the replication sink to load the hfiles from the source cluster. It does the following,<a name="line.361"></a> +<span class="sourceLineNo">362</span> * <ol><a name="line.362"></a> +<span class="sourceLineNo">363</span> * <li>LoadIncrementalHFiles#groupOrSplitPhase(Table, ExecutorService, Deque, Pair)}</li><a name="line.363"></a> +<span class="sourceLineNo">364</span> * <li>LoadIncrementalHFiles#bulkLoadPhase(Table, Connection, ExecutorService, Deque, Multimap)<a name="line.364"></a> +<span class="sourceLineNo">365</span> * </li><a name="line.365"></a> +<span class="sourceLineNo">366</span> * </ol><a name="line.366"></a> +<span class="sourceLineNo">367</span> * @param table Table to which these hfiles should be loaded to<a name="line.367"></a> +<span class="sourceLineNo">368</span> * @param conn Connection to use<a name="line.368"></a> +<span class="sourceLineNo">369</span> * @param queue {@link LoadQueueItem} has hfiles yet to be loaded<a name="line.369"></a> +<span class="sourceLineNo">370</span> * @param startEndKeys starting and ending row keys of the region<a name="line.370"></a> +<span class="sourceLineNo">371</span> */<a name="line.371"></a> +<span class="sourceLineNo">372</span> public void loadHFileQueue(Table table, Connection conn, Deque<LoadQueueItem> queue,<a name="line.372"></a> +<span class="sourceLineNo">373</span> Pair<byte[][], byte[][]> startEndKeys, boolean copyFile) throws IOException {<a name="line.373"></a> +<span class="sourceLineNo">374</span> ExecutorService pool = null;<a name="line.374"></a> +<span class="sourceLineNo">375</span> try {<a name="line.375"></a> +<span class="sourceLineNo">376</span> pool = createExecutorService();<a name="line.376"></a> +<span class="sourceLineNo">377</span> Multimap<ByteBuffer, LoadQueueItem> regionGroups =<a name="line.377"></a> +<span class="sourceLineNo">378</span> groupOrSplitPhase(table, pool, queue, startEndKeys).getFirst();<a name="line.378"></a> +<span class="sourceLineNo">379</span> bulkLoadPhase(table, conn, pool, queue, regionGroups, copyFile, null);<a name="line.379"></a> +<span class="sourceLineNo">380</span> } finally {<a name="line.380"></a> +<span class="sourceLineNo">381</span> if (pool != null) {<a name="line.381"></a> +<span class="sourceLineNo">382</span> pool.shutdown();<a name="line.382"></a> +<span class="sourceLineNo">383</span> }<a name="line.383"></a> +<span class="sourceLineNo">384</span> }<a name="line.384"></a> +<span class="sourceLineNo">385</span> }<a name="line.385"></a> +<span class="sourceLineNo">386</span><a name="line.386"></a> +<span class="sourceLineNo">387</span> private Map<LoadQueueItem, ByteBuffer> performBulkLoad(Admin admin, Table table,<a name="line.387"></a> +<span class="sourceLineNo">388</span> RegionLocator regionLocator, Deque<LoadQueueItem> queue, ExecutorService pool,<a name="line.388"></a> +<span class="sourceLineNo">389</span> SecureBulkLoadClient secureClient, boolean copyFile) throws IOException {<a name="line.389"></a> +<span class="sourceLineNo">390</span> int count = 0;<a name="line.390"></a> +<span class="sourceLineNo">391</span><a name="line.391"></a> +<span class="sourceLineNo">392</span> if (isSecureBulkLoadEndpointAvailable()) {<a name="line.392"></a> +<span class="sourceLineNo">393</span> LOG.warn("SecureBulkLoadEndpoint is deprecated. It will be removed in future releases.");<a name="line.393"></a> +<span class="sourceLineNo">394</span> LOG.warn("Secure bulk load has been integrated into HBase core.");<a name="line.394"></a> +<span class="sourceLineNo">395</span> }<a name="line.395"></a> +<span class="sourceLineNo">396</span><a name="line.396"></a> +<span class="sourceLineNo">397</span> fsDelegationToken.acquireDelegationToken(queue.peek().getFilePath().getFileSystem(getConf()));<a name="line.397"></a> +<span class="sourceLineNo">398</span> bulkToken = secureClient.prepareBulkLoad(admin.getConnection());<a name="line.398"></a> +<span class="sourceLineNo">399</span> Pair<Multimap<ByteBuffer, LoadQueueItem>, Set<String>> pair = null;<a name="line.399"></a> +<span class="sourceLineNo">400</span><a name="line.400"></a> +<span class="sourceLineNo">401</span> Map<LoadQueueItem, ByteBuffer> item2RegionMap = new HashMap<>();<a name="line.401"></a> +<span class="sourceLineNo">402</span> // Assumes that region splits can happen while this occurs.<a name="line.402"></a> +<span class="sourceLineNo">403</span> while (!queue.isEmpty()) {<a name="line.403"></a> +<span class="sourceLineNo">404</span> // need to reload split keys each iteration.<a name="line.404"></a> +<span class="sourceLineNo">405</span> final Pair<byte[][], byte[][]> startEndKeys = regionLocator.getStartEndKeys();<a name="line.405"></a> +<span class="sourceLineNo">406</span> if (count != 0) {<a name="line.406"></a> +<span class="sourceLineNo">407</span> LOG.info("Split occurred while grouping HFiles, retry attempt " + +count + " with " +<a name="line.407"></a> +<span class="sourceLineNo">408</span> queue.size() + " files remaining to group or split");<a name="line.408"></a> +<span class="sourceLineNo">409</span> }<a name="line.409"></a> +<span class="sourceLineNo">410</span><a name="line.410"></a> +<span class="sourceLineNo">411</span> int maxRetries = getConf().getInt(HConstants.BULKLOAD_MAX_RETRIES_NUMBER, 10);<a name="line.411"></a> +<span class="sourceLineNo">412</span> maxRetries = Math.max(maxRetries, startEndKeys.getFirst().length + 1);<a name="line.412"></a> +<span class="sourceLineNo">413</span> if (maxRetries != 0 && count >= maxRetries) {<a name="line.413"></a> +<span class="sourceLineNo">414</span> throw new IOException(<a name="line.414"></a> +<span class="sourceLineNo">415</span> "Retry attempted " + count + " times without completing, bailing out");<a name="line.415"></a> +<span class="sourceLineNo">416</span> }<a name="line.416"></a> +<span class="sourceLineNo">417</span> count++;<a name="line.417"></a> +<span class="sourceLineNo">418</span><a name="line.418"></a> +<span class="sourceLineNo">419</span> // Using ByteBuffer for byte[] equality semantics<a name="line.419"></a> +<span class="sourceLineNo">420</span> pair = groupOrSplitPhase(table, pool, queue, startEndKeys);<a name="line.420"></a> +<span class="sourceLineNo">421</span> Multimap<ByteBuffer, LoadQueueItem> regionGroups = pair.getFirst();<a name="line.421"></a> +<span class="sourceLineNo">422</span><a name="line.422"></a> +<span class="sourceLineNo">423</span> if (!checkHFilesCountPerRegionPerFamily(regionGroups)) {<a name="line.423"></a> +<span class="sourceLineNo">424</span> // Error is logged inside checkHFilesCountPerRegionPerFamily.<a name="line.424"></a> +<span class="sourceLineNo">425</span> throw new IOException("Trying to load more than " + maxFilesPerRegionPerFamily +<a name="line.425"></a> +<span class="sourceLineNo">426</span> " hfiles to one family of one region");<a name="line.426"></a> +<span class="sourceLineNo">427</span> }<a name="line.427"></a> +<span class="sourceLineNo">428</span><a name="line.428"></a> +<span class="sourceLineNo">429</span> bulkLoadPhase(table, admin.getConnection(), pool, queue, regionGroups, copyFile,<a name="line.429"></a> +<span class="sourceLineNo">430</span> item2RegionMap);<a name="line.430"></a> +<span class="sourceLineNo">431</span><a name="line.431"></a> +<span class="sourceLineNo">432</span> // NOTE: The next iteration's split / group could happen in parallel to<a name="line.432"></a> +<span class="sourceLineNo">433</span> // atomic bulkloads assuming that there are splits and no merges, and<a name="line.433"></a> +<span class="sourceLineNo">434</span> // that we can atomically pull out the groups we want to retry.<a name="line.434"></a> +<span class="sourceLineNo">435</span> }<a name="line.435"></a> +<span class="sourceLineNo">436</span><a name="line.436"></a> +<span class="sourceLineNo">437</span> if (!queue.isEmpty()) {<a name="line.437"></a> +<span class="sourceLineNo">438</span> throw new RuntimeException("Bulk load aborted with some files not yet loaded." +<a name="line.438"></a> +<span class="sourceLineNo">439</span> "Please check log for more details.");<a name="line.439"></a> +<span class="sourceLineNo">440</span> }<a name="line.440"></a> +<span class="sourceLineNo">441</span> return item2RegionMap;<a name="line.441"></a> +<span class="sourceLineNo">442</span> }<a name="line.442"></a> +<span class="sourceLineNo">443</span><a name="line.443"></a> +<span class="sourceLineNo">444</span> /**<a name="line.444"></a> +<span class="sourceLineNo">445</span> * This takes the LQI's grouped by likely regions and attempts to bulk load them. Any failures are<a name="line.445"></a> +<span class="sourceLineNo">446</span> * re-queued for another pass with the groupOrSplitPhase.<a name="line.446"></a> +<span class="sourceLineNo">447</span> * <p><a name="line.447"></a> +<span class="sourceLineNo">448</span> * protected for testing.<a name="line.448"></a> +<span class="sourceLineNo">449</span> */<a name="line.449"></a> +<span class="sourceLineNo">450</span> @VisibleForTesting<a name="line.450"></a> +<span class="sourceLineNo">451</span> protected void bulkLoadPhase(Table table, Connection conn, ExecutorService pool,<a name="line.451"></a> +<span class="sourceLineNo">452</span> Deque<LoadQueueItem> queue, Multimap<ByteBuffer, LoadQueueItem> regionGroups,<a name="line.452"></a> +<span class="sourceLineNo">453</span> boolean copyFile, Map<LoadQueueItem, ByteBuffer> item2RegionMap) throws IOException {<a name="line.453"></a> +<span class="sourceLineNo">454</span> // atomically bulk load the groups.<a name="line.454"></a> +<span class="sourceLineNo">455</span> Set<Future<List<LoadQueueItem>>> loadingFutures = new HashSet<>();<a name="line.455"></a> +<span class="sourceLineNo">456</span> for (Entry<ByteBuffer, ? extends Collection<LoadQueueItem>> e : regionGroups.asMap()<a name="line.456"></a> +<span class="sourceLineNo">457</span> .entrySet()) {<a name="line.457"></a> +<span class="sourceLineNo">458</span> byte[] first = e.getKey().array();<a name="line.458"></a> +<span class="sourceLineNo">459</span> Collection<LoadQueueItem> lqis = e.getValue();<a name="line.459"></a> +<span class="sourceLineNo">460</span><a name="line.460"></a> +<span class="sourceLineNo">461</span> ClientServiceCallable<byte[]> serviceCallable =<a name="line.461"></a> +<span class="sourceLineNo">462</span> buildClientServiceCallable(conn, table.getName(), first, lqis, copyFile);<a name="line.462"></a> +<span class="sourceLineNo">463</span><a name="line.463"></a> +<span class="sourceLineNo">464</span> Callable<List<LoadQueueItem>> call = new Callable<List<LoadQueueItem>>() {<a name="line.464"></a> +<span class="sourceLineNo">465</span> @Override<a name="line.465"></a> +<span class="sourceLineNo">466</span> public List<LoadQueueItem> call() throws Exception {<a name="line.466"></a> +<span class="sourceLineNo">467</span> List<LoadQueueItem> toRetry =<a name="line.467"></a> +<span class="sourceLineNo">468</span> tryAtomicRegionLoad(serviceCallable, table.getName(), first, lqis);<a name="line.468"></a> +<span class="sourceLineNo">469</span> return toRetry;<a name="line.469"></a> +<span class="sourceLineNo">470</span> }<a name="line.470"></a> +<span class="sourceLineNo">471</span> };<a name="line.471"></a> +<span class="sourceLineNo">472</span> if (item2RegionMap != null) {<a name="line.472"></a> +<span class="sourceLineNo">473</span> for (LoadQueueItem lqi : lqis) {<a name="line.473"></a> +<span class="sourceLineNo">474</span> item2RegionMap.put(lqi, e.getKey());<a name="line.474"></a> +<span class="sourceLineNo">475</span> }<a name="line.475"></a> +<span class="sourceLineNo">476</span> }<a name="line.476"></a> +<span class="sourceLineNo">477</span> loadingFutures.add(pool.submit(call));<a name="line.477"></a> +<span class="sourceLineNo">478</span> }<a name="line.478"></a> +<span class="sourceLineNo">479</span><a name="line.479"></a> +<span class="sourceLineNo">480</span> // get all the results.<a name="line.480"></a> +<span class="sourceLineNo">481</span> for (Future<List<LoadQueueItem>> future : loadingFutures) {<a name="line.481"></a> +<span class="sourceLineNo">482</span> try {<a name="line.482"></a> +<span class="sourceLineNo">483</span> List<LoadQueueItem> toRetry = future.get();<a name="line.483"></a> +<span class="sourceLineNo">484</span><a name="line.484"></a> +<span class="sourceLineNo">485</span> if (item2RegionMap != null) {<a name="line.485"></a> +<span class="sourceLineNo">486</span> for (LoadQueueItem lqi : toRetry) {<a name="line.486"></a> +<span class="sourceLineNo">487</span> item2RegionMap.remove(lqi);<a name="line.487"></a> +<span class="sourceLineNo">488</span> }<a name="line.488"></a> +<span class="sourceLineNo">489</span> }<a name="line.489"></a> +<span class="sourceLineNo">490</span> // LQIs that are requeued to be regrouped.<a name="line.490"></a> +<span class="sourceLineNo">491</span> queue.addAll(toRetry);<a name="line.491"></a> +<span class="sourceLineNo">492</span><a name="line.492"></a> +<span class="sourceLineNo">493</span> } catch (ExecutionException e1) {<a name="line.493"></a> +<span class="sourceLineNo">494</span> Throwable t = e1.getCause();<a name="line.494"></a> +<span class="sourceLineNo">495</span> if (t instanceof IOException) {<a name="line.495"></a> +<span class="sourceLineNo">496</span> // At this point something unrecoverable has happened.<a name="line.496"></a> +<span class="sourceLineNo">497</span> // TODO Implement bulk load recovery<a name="line.497"></a> +<span class="sourceLineNo">498</span> throw new IOException("BulkLoad encountered an unrecoverable problem", t);<a name="line.498"></a> +<span class="sourceLineNo">499</span> }<a name="line.499"></a> +<span class="sourceLineNo">500</span> LOG.error("Unexpected execution exception during bulk load", e1);<a name="line.500"></a> +<span class="sourceLineNo">501</span> throw new IllegalStateException(t);<a name="line.501"></a> +<span class="sourceLineNo">502</span> } catch (InterruptedException e1) {<a name="line.502"></a> +<span class="sourceLineNo">503</span> LOG.error("Unexpected interrupted exception during bulk load", e1);<a name="line.503"></a> +<span class="sourceLineNo">504</span> throw (InterruptedIOException) new InterruptedIOException().initCause(e1);<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> }<a name="line.507"></a> +<span class="sourceLineNo">508</span><a name="line.508"></a> +<span class="sourceLineNo">509</span> @VisibleForTesting<a name="line.509"></a> +<span class="sourceLineNo">510</span> protected ClientServiceCallable<byte[]> buildClientServiceCallable(Connection conn,<a name="line.510"></a> +<span class="sourceLineNo">511</span> TableName tableName, byte[] first, Collection<LoadQueueItem> lqis, boolean copyFile) {<a name="line.511"></a> +<span class="sourceLineNo">512</span> List<Pair<byte[], String>> famPaths =<a name="line.512"></a> +<span class="sourceLineNo">513</span> lqis.stream().map(lqi -> Pair.newPair(lqi.getFamily(), lqi.getFilePath().toString()))<a name="line.513"></a> +<span class="sourceLineNo">514</span> .collect(Collectors.toList());<a name="line.514"></a> +<span class="sourceLineNo">515</span> return new ClientServiceCallable<byte[]>(conn, tableName, first,<a name="line.515"></a> +<span class="sourceLineNo">516</span> rpcControllerFactory.newController(), HConstants.PRIORITY_UNSET) {<a name="line.516"></a> +<span class="sourceLineNo">517</span> @Override<a name="line.517"></a> +<span class="sourceLineNo">518</span> protected byte[] rpcCall() throws Exception {<a name="line.518"></a> +<span class="sourceLineNo">519</span> SecureBulkLoadClient secureClient = null;<a name="line.519"></a> +<span class="sourceLineNo">520</span> boolean success = false;<a name="line.520"></a> +<span class="sourceLineNo">521</span> try {<a name="line.521"></a> +<span class="sourceLineNo">522</span> if (LOG.isDebugEnabled()) {<a name="line.522"></a> +<span class="sourceLineNo">523</span> LOG.debug("Going to connect to server " + getLocation() + " for row " +<a name="line.523"></a> +<span class="sourceLineNo">524</span> Bytes.toStringBinary(getRow()) + " with hfile group " +<a name="line.524"></a> +<span class="sourceLineNo">525</span> LoadIncrementalHFiles.this.toString(famPaths));<a name="line.525"></a> +<span class="sourceLineNo">526</span> }<a name="line.526"></a> +<span class="sourceLineNo">527</span> byte[] regionName = getLocation().getRegionInfo().getRegionName();<a name="line.527"></a> +<span class="sourceLineNo">528</span> try (Table table = conn.getTable(getTableName())) {<a name="line.528"></a> +<span class="sourceLineNo">529</span> secureClient = new SecureBulkLoadClient(getConf(), table);<a name="line.529"></a> +<span class="sourceLineNo">530</span> success = secureClient.secureBulkLoadHFiles(getStub(), famPaths, regionName,<a name="line.530"></a> +<span class="sourceLineNo">531</span> assignSeqIds, fsDelegationToken.getUserToken(), bulkToken, copyFile);<a name="line.531"></a> +<span class="sourceLineNo">532</span> }<a name="line.532"></a> +<span class="sourceLineNo">533</span> return success ? regionName : null;<a name="line.533"></a> +<span class="sourceLineNo">534</span> } finally {<a name="line.534"></a> +<span class="sourceLineNo">535</span> // Best effort copying of files that might not have been imported<a name="line.535"></a> +<span class="sourceLineNo">536</span> // from the staging directory back to original location<a name="line.536"></a> +<span class="sourceLineNo">537</span> // in user directory<a name="line.537"></a> +<span class="sourceLineNo">538</span> if (secureClient != null && !success) {<a name="line.538"></a> +<span class="sourceLineNo">539</span> FileSystem targetFs = FileSystem.get(getConf());<a name="line.539"></a> +<span class="sourceLineNo">540</span> FileSystem sourceFs = lqis.iterator().next().getFilePath().getFileSystem(getConf());<a name="line.540"></a> +<span class="sourceLineNo">541</span> // Check to see if the source and target filesystems are the same<a name="line.541"></a> +<span class="sourceLineNo">542</span> // If they are the same filesystem, we will try move the files back<a name="line.542"></a> +<span class="sourceLineNo">543</span> // because previously we moved them to the staging directory.<a name="line.543"></a> +<span class="sourceLineNo">544</span> if (FSHDFSUtils.isSameHdfs(getConf(), sourceFs, targetFs)) {<a name="line.544"></a> +<span class="sourceLineNo">545</span> for (Pair<byte[], String> el : famPaths) {<a name="line.545"></a> +<span class="sourceLineNo">546</span> Path hfileStagingPath = null;<a name="line.546"></a> +<span class="sourceLineNo">547</span> Path hfileOrigPath = new Path(el.getSecond());<a name="line.547"></a> +<span class="sourceLineNo">548</span> try {<a name="line.548"></a> +<span class="sourceLineNo">549</span> hfileStagingPath = new Path(new Path(bulkToken, Bytes.toString(el.getFirst())),<a name="line.549"></a> +<span class="sourceLineNo">550</span> hfileOrigPath.getName());<a name="line.550"></a> +<span class="sourceLineNo">551</span> if (targetFs.rename(hfileStagingPath, hfileOrigPath)) {<a name="line.551"></a> +<span class="sourceLineNo">552</span> LOG.debug("Moved back file " + hfileOrigPath + " from " + hfileStagingPath);<a name="line.552"></a> +<span class="sourceLineNo">553</span> } else if (targetFs.exists(hfileStagingPath)) {<a name="line.553"></a> +<span class="sourceLineNo">554</span> LOG.debug(<a name="line.554"></a> +<span class="sourceLineNo">555</span> "Unable to move back file " + hfileOrigPath + " from " + hfileStagingPath);<a name="line.555"></a> +<span class="sourceLineNo">556</span> }<a name="line.556"></a> +<span class="sourceLineNo">557</span> } catch (Exception ex) {<a name="line.557"></a> +<span class="sourceLineNo">558</span> LOG.debug(<a name="line.558"></a> +<span class="sourceLineNo">559</span> "Unable to move back file " + hfileOrigPath + " from " + hfileStagingPath, ex);<a name="line.559"></a> +<span class="sourceLineNo">560</span> }<a name="line.560"></a> +<span class="sourceLineNo">561</span> }<a name="line.561"></a> +<span class="sourceLineNo">562</span> }<a name="line.562"></a> +<span class="sourceLineNo">563</span> }<a name="line.563"></a> +<span class="sourceLineNo">564</span> }<a name="line.564"></a> +<span class="sourceLineNo">565</span> }<a name="line.565"></a> +<span class="sourceLineNo">566</span> };<a name="line.566"></a> +<span class="sourceLineNo">567</span> }<a name="line.567"></a> +<span class="sourceLineNo">568</span><a name="line.568"></a> +<span class="sourceLineNo">569</span> private boolean checkHFilesCountPerRegionPerFamily(<a name="line.569"></a> +<span class="sourceLineNo">570</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups) {<a name="line.570"></a> +<span class="sourceLineNo">571</span> for (Map.Entry<ByteBuffer, Collection<LoadQueueItem>> e : regionGroups.asMap().entrySet()) {<a name="line.571"></a> +<span class="sourceLineNo">572</span> Map<byte[], MutableInt> filesMap = new TreeMap<>(Bytes.BYTES_COMPARATOR);<a name="line.572"></a> +<span class="sourceLineNo">573</span> for (LoadQueueItem lqi : e.getValue()) {<a name="line.573"></a> +<span class="sourceLineNo">574</span> MutableInt count = filesMap.computeIfAbsent(lqi.getFamily(), k -> new MutableInt());<a name="line.574"></a> +<span class="sourceLineNo">575</span> count.increment();<a name="line.575"></a> +<span class="sourceLineNo">576</span> if (count.intValue() > maxFilesPerRegionPerFamily) {<a name="line.576"></a> +<span class="sourceLineNo">577</span> LOG.error("Trying to load more than " + maxFilesPerRegionPerFamily +<a name="line.577"></a> +<span class="sourceLineNo">578</span> " hfiles to family " + Bytes.toStringBinary(lqi.getFamily()) +<a name="line.578"></a> +<span class="sourceLineNo">579</span> " of region with start key " + Bytes.toStringBinary(e.getKey()));<a name="line.579"></a> +<span class="sourceLineNo">580</span> return false;<a name="line.580"></a> +<span class="sourceLineNo">581</span> }<a name="line.581"></a> +<span class="sourceLineNo">582</span> }<a name="line.582"></a> +<span class="sourceLineNo">583</span> }<a name="line.583"></a> +<span class="sourceLineNo">584</span> return true;<a name="line.584"></a> +<span class="sourceLineNo">585</span> }<a name="line.585"></a> +<span class="sourceLineNo">586</span><a name="line.586"></a> +<span class="sourceLineNo">587</span> /**<a name="line.587"></a> +<span class="sourceLineNo">588</span> * @param table the table to load into<a name="line.588"></a> +<span class="sourceLineNo">589</span> * @param pool the ExecutorService<a name="line.589"></a> +<span class="sourceLineNo">590</span> * @param queue the queue for LoadQueueItem<a name="line.590"></a> +<span class="sourceLineNo">591</span> * @param startEndKeys start and end keys<a name="line.591"></a> +<span class="sourceLineNo">592</span> * @return A map that groups LQI by likely bulk load region targets and Set of missing hfiles.<a name="line.592"></a> +<span class="sourceLineNo">593</span> */<a name="line.593"></a> +<span class="sourceLineNo">594</span> private Pair<Multimap<ByteBuffer, LoadQueueItem>, Set<String>> groupOrSplitPhase(<a name="line.594"></a> +<span class="sourceLineNo">595</span> final Table table, ExecutorService pool, Deque<LoadQueueItem> queue,<a name="line.595"></a> +<span class="sourceLineNo">596</span> final Pair<byte[][], byte[][]> startEndKeys) throws IOException {<a name="line.596"></a> +<span class="sourceLineNo">597</span> // <region start key, LQI> need synchronized only within this scope of this<a name="line.597"></a> +<span class="sourceLineNo">598</span> // phase because of the puts that happen in futures.<a name="line.598"></a> +<span class="sourceLineNo">599</span> Multimap<ByteBuffer, LoadQueueItem> rgs = HashMultimap.create();<a name="line.599"></a> +<span class="sourceLineNo">600</span> final Multimap<ByteBuffer, LoadQueueItem> regionGroups = Multimaps.synchronizedMultimap(rgs);<a name="line.600"></a> +<span class="sourceLineNo">601</span> Set<String> missingHFiles = new HashSet<>();<a name="line.601"></a> +<span class="sourceLineNo">602</span> Pair<Multimap<ByteBuffer, LoadQueueItem>, Set<String>> pair =<a name="line.602"></a> +<span class="sourceLineNo">603</span> new Pair<>(regionGroups, missingHFiles);<a name="line.603"></a> +<span class="sourceLineNo">604</span><a name="line.604"></a> +<span class="sourceLineNo">605</span> // drain LQIs and figure out bulk load groups<a name="line.605"></a> +<span class="sourceLineNo">606</span> Set<Future<Pair<List<LoadQueueItem>, String>>> splittingFutures = new HashSet<>();<a name="line.606"></a> +<span class="sourceLineNo">607</span> while (!queue.isEmpty()) {<a name="line.607"></a> +<span class="sourceLineNo">608</span> final LoadQueueItem item = queue.remove();<a name="line.608"></a> +<span class="sourceLineNo">609</span><a name="line.609"></a> +<span class="sourceLineNo">610</span> final Callable<Pair<List<LoadQueueItem>, String>> call =<a name="line.610"></a> +<span class="sourceLineNo">611</span> new Callable<Pair<List<LoadQueueItem>, String>>() {<a name="line.611"></a> +<span class="sourceLineNo">612</span> @Override<a name="line.612"></a> +<span class="sourceLineNo">613</span> public Pair<List<LoadQueueItem>, String> call() throws Exception {<a name="line.613"></a> +<span class="sourceLineNo">614</span> Pair<List<LoadQueueItem>, String> splits =<a name="line.614"></a> +<span class="sourceLineNo">615</span> groupOrSplit(regionGroups, item, table, startEndKeys);<a name="line.615"></a> +<span class="sourceLineNo">616</span> return splits;<a name="line.616"></a> +<span class="sourceLineNo">617</span> }<a name="line.617"></a> +<span class="sourceLineNo">618</span> };<a name="line.618"></a> +<span class="sourceLineNo">619</span> splittingFutures.add(pool.submit(call));<a name="line.619"></a> +<span class="sourceLineNo">620</span> }<a name="line.620"></a> +<span class="sourceLineNo">621</span> // get all the results. All grouping and splitting must finish before<a name="line.621"></a> +<span class="sourceLineNo">622</span> // we can attempt the atomic loads.<a name="line.622"></a> +<span class="sourceLineNo">623</span> for (Future<Pair<List<LoadQueueItem>, String>> lqis : splittingFutures) {<a name="line.623"></a> +<span class="sourceLineNo">624</span> try {<a name="line.624"></a> +<span class="sourceLineNo">625</span> Pair<List<LoadQueueItem>, String> splits = lqis.get();<a name="line.625"></a> +<span class="sourceLineNo">626</span> if (splits != null) {<a name="line.626"></a> +<span class="sourceLineNo">627</span> if (splits.getFirst() != null) {<a name="line.627"></a> +<span class="sourceLineNo">628</span> queue.addAll(splits.getFirst());<a name="line.628"></a> +<span class="sourceLineNo">629</span> } else {<a name="line.629"></a> +<span class="sourceLineNo">630</span> missingHFiles.add(splits.getSecond());<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> } catch (ExecutionException e1) {<a name="line.633"></a> +<span class="sourceLineNo">634</span> Throwable t = e1.getCause();<a name="line.634"></a> +<span class="sourceLineNo">635</span> if (t instanceof IOException) {<a name="line.635"></a> +<span class="sourceLineNo">636</span> LOG.error("IOException during splitting", e1);<a name="line.636"></a> +<span class="sourceLineNo">637</span> throw (IOException) t; // would have been thrown if not parallelized,<a name="line.637"></a> +<span class="sourceLineNo">638</span> }<a name="line.638"></a> +<span class="sourceLineNo">639</span> LOG.error("Unexpected execution exception during splitting", e1);<a name="line.639"></a> +<span class="sourceLineNo">640</span> throw new IllegalStateException(t);<a name="line.640"></a> +<span class="sourceLineNo">641</span> } catch (InterruptedException e1) {<a name="line.641"></a> +<span class="sourceLineNo">642</span> LOG.error("Unexpected interrupted exception during splitting", e1);<a name="line.642"></a> +<span class="sourceLineNo">643</span> throw (InterruptedIOException) new InterruptedIOException().initCause(e1);<a name="line.643"></a> +<span class="sourceLineNo">644</span> }<a name="line.644"></a> +<span class="sourceLineNo">645</span> }<a name="line.645"></a> +<span class="sourceLineNo">646</span> return pair;<a name="line.646"></a> +<span class="sourceLineNo">647</span> }<a name="line.647"></a> +<span class="sourceLineNo">648</span><a name="line.648"></a> +<span class="sourceLineNo">649</span> private List<LoadQueueItem> splitStoreFile(final LoadQueueItem item, final Table table,<a name="line.649"></a> +<span class="sourceLineNo">650</span> byte[] startKey, byte[] splitKey) throws IOException {<a name="line.650"></a> +<span class="sourceLineNo">651</span> Path hfilePath = item.getFilePath();<a name="line.651"></a> +<span class="sourceLineNo">652</span> byte[] family = item.getFamily();<a name="line.652"></a> +<span class="sourceLineNo">653</span> Path tmpDir = hfilePath.getParent();<a name="line.653"></a> +<span class="sourceLineNo">654</span> if (!tmpDir.getName().equals(TMP_DIR)) {<a name="line.654"></a> +<span class="sourceLineNo">655</span> tmpDir = new Path(tmpDir, TMP_DIR);<a name="line.655"></a> +<span class="sourceLineNo">656</span> }<a name="line.656"></a> +<span class="sourceLineNo">657</span><a name="line.657"></a> +<span class="sourceLineNo">658</span> LOG.info("HFile at " + hfilePath + " no longer fits inside a single " + "region. Splitting...");<a name="line.658"></a> +<span class="sourceLineNo">659</span><a name="line.659"></a> +<span class="sourceLineNo">660</span> String uniqueName = getUniqueName();<a name="line.660"></a> +<span class="sourceLineNo">661</span> ColumnFamilyDescriptor familyDesc = table.getDescriptor().getColumnFamily(family);<a name="line.661"></a> +<span class="sourceLineNo">662</span><a name="line.662"></a> +<span class="sourceLineNo">663</span> Path botOut = new Path(tmpDir, uniqueName + ".bottom");<a name="line.663"></a> +<span class="sourceLineNo">664</span> Path topOut = new Path(tmpDir, uniqueName + ".top");<a name="line.664"></a> +<span class="sourceLineNo">665</span> splitStoreFile(getConf(), hfilePath, familyDesc, splitKey, botOut, topOut);<a name="line.665"></a> +<span class="sourceLineNo">666</span><a name="line.666"></a> +<span class="sourceLineNo">667</span> FileSystem fs = tmpDir.getFileSystem(getConf());<a name="line.667"></a> +<span class="sourceLineNo">668</span> fs.setPermission(tmpDir, FsPermission.valueOf("-rwxrwxrwx"));<a name="line.668"></a> +<span class="sourceLineNo">669</span> fs.setPermission(botOut, FsPermission.valueOf("-rwxrwxrwx"));<a name="line.669"></a> +<span class="sourceLineNo">670</span> fs.setPermission(topOut, FsPermission.valueOf("-rwxrwxrwx"));<a name="line.670"></a> +<span class="sourceLineNo">671</span><a name="line.671"></a> +<span class="sourceLineNo">672</span> // Add these back at the *front* of the queue, so there's a lower<a name="line.672"></a> +<span class="sourceLineNo">673</span> // chance that the region will just split again before we get there.<a name="line.673"></a> +<span class="sourceLineNo">674</span> List<LoadQueueItem> lqis = new ArrayList<>(2);<a name="line.674"></a> +<span class="sourceLineNo">675</span> lqis.add(new LoadQueueItem(family, botOut));<a name="line.675"></a> +<span class="sourceLineNo">676</span> lqis.add(new LoadQueueItem(family, topOut));<a name="line.676"></a> +<span class="sourceLineNo">677</span><a name="line.677"></a> +<span class="sourceLineNo">678</span> // If the current item is already the result of previous splits,<a name="line.678"></a> +<span class="sourceLineNo">679</span> // we don't need it anymore. Clean up to save space.<a name="line.679"></a> +<span class="sourceLineNo">680</span> // It is not part of the original input files.<a name="line.680"></a> +<span class="sourceLineNo">681</span> try {<a name="line.681"></a> +<span class="sourceLineNo">682</span> if (tmpDir.getName().equals(TMP_DIR)) {<a name="line.682"></a> +<span class="sourceLineNo">683</span> fs.delete(hfilePath, false);<a name="line.683"></a> +<span class="sourceLineNo">684</span> }<a name="line.684"></a> +<span class="sourceLineNo">685</span> } catch (IOException e) {<a name="line.685"></a> +<span class="sourceLineNo">686</span> LOG.warn("Unable to delete temporary split file " + hfilePath);<a name="line.686"></a> +<span class="sourceLineNo">687</span> }<a name="line.687"></a> +<span class="sourceLineNo">688</span> LOG.info("Successfully split into new HFiles " + botOut + " and " + topOut);<a name="line.688"></a> +<span class="sourceLineNo">689</span> return lqis;<a name="line.689"></a> +<span class="sourceLineNo">690</span> }<a name="line.690"></a> +<span class="sourceLineNo">691</span><a name="line.691"></a> +<span class="sourceLineNo">692</span> /**<a name="line.692"></a> +<span class="sourceLineNo">693</span> * Attempt to assign the given load queue item into its target region group. If the hfile boundary<a name="line.693"></a> +<span class="sourceLineNo">694</span> * no longer fits into a region, physically splits the hfile such that the new bottom half will<a name="line.694"></a> +<span class="sourceLineNo">695</span> * fit and returns the list of LQI's corresponding to the resultant hfiles.<a name="line.695"></a> +<span class="sourceLineNo">696</span> * <p><a name="line.696"></a> +<span class="sourceLineNo">697</span> * protected for testing<a name="line.697"></a> +<span class="sourceLineNo">698</span> * @throws IOException if an IO failure is encountered<a name="line.698"></a> +<span class="sourceLineNo">699</span> */<a name="line.699"></a> +<span class="sourceLineNo">700</span> @VisibleForTesting<a name="line.700"></a> +<span class="sourceLineNo">701</span> protected Pair<List<LoadQueueItem>, String> groupOrSplit(<a name="line.701"></a> +<span class="sourceLineNo">702</span> Multimap<ByteBuffer, LoadQueueItem> regionGroups, final LoadQueueItem item, final Table table,<a name="line.702"></a> +<span class="sourceLineNo">703</span> final Pair<byte[][], byte[][]> startEndKeys) throws IOException {<a name="line.703"></a> +<span class="sourceLineNo">704</span> Path hfilePath = item.getFilePath();<a name="line.704"></a> +<span class="sourceLineNo">705</span> byte[] first, last;<a name="line.705"></a> +<span class="sourceLineNo">706</span> try (HFile.Reader hfr = HFile.createReader(hfilePath.getFileSystem(getConf()), hfilePath,<a name="line.706"></a> +<span class="sourceLineNo">707</span> new CacheConfig(getConf()), true, getConf())) {<a name="line.707"></a> +<span class="sourceLineNo">708</span> hfr.loadFileInfo();<a name="line.708"></a> +<span class="sourceLineNo">709</span> first = hfr.getFirstRowKey();<a name="line.709"></a> +<span class="sourceLineNo">710</span> last = hfr.getLastRowKey();<a name="line.710"></a> +<span class="sourceLineNo">711</span> } catch (FileNotFoundException fnfe) {<a name="line.711"></a> +<span class="sourceLineNo">712</span> LOG.debug("encountered", fnfe);<a name="line.712"></a> +<span class="sourceLineNo">713</span> return new Pair<>(null, hfilePath.getName());<a name="line.713"></a> +<span class="sourceLineNo">714</span> }<a name="line.714"></a> +<span class="sourceLineNo">715</span><a name="line.715"></a> +<span class="sourceLineNo">716</span> LOG.info("Trying to load hfile=" + hfilePath + " first=" + Bytes.toStringBinary(first) +<a name="line.716"></a> +<span class="sourceLineNo">717</span> " last=" + Bytes.toStringBinary(last));<a name="line.717"></a> +<span class="sourceLineNo">718</span> if (first == null || last == null) {<a name="line.718"></a> +<span class="sourceLineNo">719</span> assert first == null && last == null;<a name="line.719"></a> +<span class="sourceLineNo">720</span> // TODO what if this is due to a bad HFile?<a name="line.720"></a> +<span class="sourceLineNo">721</span> LOG.info("hfile " + hfilePath + " has no entries, skipping");<a name="line.721"></a> +<span class="sourceLineNo">722</span> return null;<a name="line.722"></a> +<span class="sourceLineNo">723</span> }<a name="line.723"></a> +<span class="sourceLineNo">724</span> if (Bytes.compareTo(first, last) > 0) {<a name="line.724"></a> +<span class="sourceLineNo">725</span> throw new IllegalArgumentException(<a name="line.725"></a> +<span class="sourceLineNo">726</span> "Invalid range: " + Bytes.toStringBinary(first) + " > " + Bytes.toStringBinary(last));<a name="line.726"></a> +<span class="sourceLineNo">727</span> }<a name="line.727"></a> +<span class="sourceLineNo">728</span> int idx = Arrays.binarySearch(startEndKeys.getFirst(), first, Bytes.BYTES_COMPARATOR);<a name="line.728"></a> +<span class="sourceLineNo">729</span> if (idx < 0) {<a name="line.729"></a> +<span class="sourceLineNo">730</span> // not on boundary, returns -(insertion index). Calculate region it<a name="line.730"></a> +<span class="sourceLineNo">731</span> // would be in.<a name="line.731"></a> +<span class="sourceLineNo">732</span> idx = -(idx + 1) - 1;<a name="line.732"></a> +<span class="sourceLineNo">733</span> }<a name="line.733"></a> +<span class="sourceLineNo">734</span> int indexForCallable = idx;<a name="line.734"></a> +<span class="sourceLineNo">735</span><a name="line.735"></a> +<span class="sourceLineNo">736</span> /**<a name="line.736"></a> +<span class="sourceLineNo">737</span> * we can consider there is a region hole in following conditions. 1) if idx < 0,then first<a name="line.737"></a> +<span class="sourceLineNo">738</span> * region info is lost. 2) if the endkey of a region is not equal to the startkey of the next<a name="line.738"></a> +<span class="sourceLineNo">739</span> * region. 3) if the endkey of the last region is not empty.<a name="line.739"></a> +<span class="sourceLineNo">740</span> */<a name="line.740"></a> +<span class="sourceLineNo">741</span> if (indexForCallable < 0) {<a name="line.741"></a> +<span class="sourceLineNo">742</span> throw new IOException("The first region info for table " + table.getName() +<a name="line.742"></a> +<span class="sourceLineNo">743</span> " can't be found in hbase:meta.Please use hbck tool to fix it first.");<a name="line.743"></a> +<span class="sourceLineNo">744</span> } else if ((indexForCallable == startEndKeys.getFirst().length - 1) &&<a name="line.744"></a> +<span class="sourceLineNo">745</span> !Bytes.equals(startEndKeys.getSecond()[indexForCallable], HConstants.EMPTY_BYTE_ARRAY)) {<a name="line.745"></a> +<span class="sourceLineNo">746</span> throw new IOException("The last region info for table " + table.getName() +<a name="line.746"></a> +<span class="sourceLineNo">747</span> " can't be found in hbase:meta.Please use hbck tool to fix it first.");<a name="line.747"></a> +<span class="sourceLineNo">748</span> } else if (indexForCallable + 1 < startEndKeys.getFirst().length &&<a name="line.748"></a> +<span class="sourceLineNo">749</span> !(Bytes.compareTo(startEndKeys.getSecond()[indexForCallable],<a name="line.749"></a> +<span class="sourceLineNo">750</span> startEndKeys.getFirst()[indexForCallable + 1]) == 0)) {<a name="line.750"></a> +<span class="sourceLineNo">751</span> throw new IOException("The endkey of one region for table " + table.getName() +<a name="line.751"></a> +<span class="sourceLineNo">752</span> " is not equal to the startkey of the next region in hbase:meta." +<a name="line.752"></a> +<span class="sourceLineNo">753</span> "Please use hbck tool to fix it first.");<a name="line.753"></a> +<span class="sourceLineNo">754</span> }<a name="line.754"></a> +<span class="sourceLineNo">755</span><a name="line.755"></a> +<span class="sourceLineNo">756</span> boolean lastKeyInRange = Bytes.compareTo(last, startEndKeys.getSecond()[idx]) < 0 ||<a name="line.756"></a> +<span class="sourceLineNo">757</span> Bytes.equals(startEndKeys.getSecond()[idx], HConstants.EMPTY_BYTE_ARRAY);<a name="line.757"></a> +<span class="sourceLineNo">758</span> if (!lastKeyInRange) {<a name="line.758"></a> +<span class="sourceLineNo">759</span> List<LoadQueueItem> lqis = splitStoreFile(item, table,<a name="line.759"></a> +<span class="sourceLineNo">760</span> startEndKeys.getFirst()[indexForCallable], startEndKeys.getSecond()[indexForCallable]);<a name="line.760"></a> +<span class="sourceLineNo">761</span> return new Pair<>(lqis, null);<a name="line.761"></a> +<span class="sourceLineNo">762</span> }<a name="line.762"></a> +<span class="sourceLineNo">763</span><a name="line.763"></a> +<span class="sourceLineNo">764</span> // group regions.<a name="line.764"></a> +<span class="sourceLineNo">765</span> regionGroups.put(ByteBuffer.wrap(startEndKeys.getFirst()[idx]), item);<a name="line.765"></a> +<span class="sourceLineNo">766</span> return null;<a name="line.766"></a> +<span class="sourceLineNo">767</span> }<a name="line.767"></a> +<span class="sourceLineNo">768</span><a name="line.768"></a> +<span class="sourceLineNo">769</span> /**<a name="line.769"></a> +<span class="sourceLineNo">770</span> * Attempts to do an atomic load of many hfiles into a region. If it fails, it returns a list of<a name="line.770"></a> +<span class="sourceLineNo">771</span> * hfiles that need to be retried. If it is successful it will return an empty list.<a name="line.771"></a> +<span class="sourceLineNo">772</span> * <p><a name="line.772"></a> +<span class="sourceLineNo">773</span> * NOTE: To maintain row atomicity guarantees, region server callable should succeed atomically<a name="line.773"></a> +<span class="sourceLineNo">774</span> * and fails atomically.<a name="line.774"></a> +<span class="sourceLineNo">775</span> * <p><a name="line.775"></a> +<span class="sourceLineNo">776</span> * Protected for testing.<a name="line.776"></a> +<span class="sourceLineNo">777</span> * @return empty list if success, list of items to retry on recoverable failure<a name="line.777"></a> +<span class="sourceLineNo">778</span> */<a name="line.778"></a> +<span class="sourceLineNo">779</span> @VisibleForTesting<a name="line.779"></a> +<span class="sourceLineNo">780</span> protected List<LoadQueueItem> tryAtomicRegionLoad(ClientServiceCallable<byte[]> serviceCallable,<a name="line.780"></a> +<span class="sourceLineNo">781</span> final TableName tableName, final byte[] first, final Collection<LoadQueueItem> lqis)<a name="line.781"></a> +<span class="sourceLineNo">782</span> throws IOException {<a name="line.782"></a> +<span class="sourceLineNo">783</span> try {<a name="line.783"></a> +<span class="sourceLineNo">784</span> List<LoadQueueItem> toRetry = new ArrayList<>();<a name="line.784"></a> +<span class="sourceLineNo">785</span> Configuration conf = getConf();<a name="line.785"></a> +<span class="sourceLineNo">786</span> byte[] region = RpcRetryingCallerFactory.instantiate(conf, null).<byte[]> newCaller()<a name="line.786"></a> +<span class="sourceLineNo">787</span> .callWithRetries(serviceCallable, Integer.MAX_VALUE);<a name="line.787"></a> +<span class="sourceLineNo">788</span> if (region == null) {<a name="line.788"></a> +<span class="sourceLineNo">789</span> LOG.warn("Attempt to bulk load region containing " + Bytes.toStringBinary(first) +<a name="line.789"></a> +<span class="sourceLineNo">790</span> " into table " + tableName + " with files " + lqis +<a name="line.790"></a> +<span class="sourceLineNo">791</span> " failed. This is recoverable and they will be retried.");<a name="line.791"></a> +<span class="sourceLineNo">792</span> toRetry.addAll(lqis); // return lqi's to retry<a name="line.792"></a> +<span class="sourceLineNo">793</span> }<a name="line.793"></a> +<span class="sourceLineNo">794</span> // success<a name="line.794"></a> +<span class="sourceLineNo">795</span> return toRetry;<a name="line.795"></a> +<span class="sourceLineNo">796</span> } catch (IOException e) {<a name="line.796"></a> +<span class="sourceLineNo">797</span> LOG.error("Encountered unrecoverable error from region server, additional details: " +<a name="line.797"></a> +<span class="sourceLineNo">798</span> serviceCallable.getExceptionMessageAdditionalDetail(),<a name="line.798"></a> +<span class="sourceLineNo">799</span> e);<a name="line.799"></a> +<span class="sourceLineNo">800</span> throw e;<a name="line.800"></a> +<span class="sourceLineNo">801</span> }<a name="line.801"></a> +<span class="sourceLineNo">802</span> }<a name="line.802"></a> +<span class="sourceLineNo">803</span><a name="line.803"></a> +<span class="sourceLineNo">804</span> /**<a name="line.804"></a> +<span class="sourceLineNo">805</span> * If the table is created for the first time, then "completebulkload" reads the files twice. More<a name="line.805"></a> +<span class="sourceLineNo">806</span> * modifications necessary if we want to avoid doing it.<a name="line.806"></a> +<span class="sourceLineNo">807</span> */<a name="line.807"></a> +<span class="sourceLineNo">808</span> private void createTable(TableName tableName, String dirPath, Admin admin) throws IOException {<a name="line.808"></a> +<span class="sourceLineNo">809</span> final Path hfofDir = new Path(dirPath);<a name="line.809"></a> +<span class="sourceLineNo">810</span> final FileSystem fs = hfofDir.getFileSystem(getConf());<a name="line.810"></a> +<span class="sourceLineNo">811</span><a name="line.811"></a> +<span class="sourceLineNo">812</span> // Add column families<a name="line.812"></a> +<span class="sourceLineNo">813</span> // Build a set of keys<a name="line.813"></a> +<span class="sourceLineNo">814</span> List<ColumnFamilyDescriptorBuilder> familyBuilders = new ArrayList<>();<a name="line.814"></a> +<span class="sourceLineNo">815</span> SortedMap<byte[], Integer> map = new TreeMap<>(Bytes.BYTES_COMPARATOR);<a name="line.815"></a> +<span class="sourceLineNo">816</span> visitBulkHFiles(fs, hfofDir, new BulkHFileVisitor<ColumnFamilyDescriptorBuilder>() {<a name="line.816"></a> +<span class="sourceLineNo">817</span> @Override<a name="line.817"></a> +<span class="sourceLineNo">818</span> public ColumnFamilyDescriptorBuilder bulkFamily(byte[] familyName) {<a name="line.818"></a> +<span class="sourceLineNo">819</span> ColumnFamilyDescriptorBuilder builder =<a name="line.819"></a> +<span class="sourceLineNo">820</span> ColumnFamilyDescriptorBuilder.newBuilder(familyName);<a name="line.820"></a> +<span class="sourceLineNo">821</span> familyBuilders.add(builder);<a name="line.821"></a> +<span class="sourceLineNo">822</span> return builder;<a name="line.822"></a> +<span class="sourceLineNo">823</span> }<a name="line.823"></a> +<span class="sourceLineNo">824</span><a name="line.824"></a> +<span class="sourceLineNo">825</span> @Override<a name="line.825"></a> +<span class="sourceLineNo">826</span> public void bulkHFile(ColumnFamilyDescriptorBuilder builder, FileStatus hfileStatus)<a name="line.826"></a> +<span class="sourceLineNo">827</span> throws IOException {<a name="line.827"></a> +<span class="sourceLineNo">828</span> Path hfile = hfileStatus.getPath();<a name="line.828"></a> +<span class="sourceLineNo">829</span> try (HFile.Reader reader =<a name="line.829"></a>
<TRUNCATED>
