| <!DOCTYPE HTML> |
| <html lang="en"> |
| <head> |
| <!-- Generated by javadoc (17) --> |
| <title>Source code</title> |
| <meta name="viewport" content="width=device-width, initial-scale=1"> |
| <meta name="description" content="source: package: org.apache.hadoop.hbase.mapreduce, class: WALInputFormat, class: WALSplit"> |
| <meta name="generator" content="javadoc/SourceToHTMLConverter"> |
| <link rel="stylesheet" type="text/css" href="../../../../../../stylesheet.css" title="Style"> |
| </head> |
| <body class="source-page"> |
| <main role="main"> |
| <div class="source-container"> |
| <pre><span class="source-line-no">001</span><span id="line-1">/*</span> |
| <span class="source-line-no">002</span><span id="line-2"> * Licensed to the Apache Software Foundation (ASF) under one</span> |
| <span class="source-line-no">003</span><span id="line-3"> * or more contributor license agreements. See the NOTICE file</span> |
| <span class="source-line-no">004</span><span id="line-4"> * distributed with this work for additional information</span> |
| <span class="source-line-no">005</span><span id="line-5"> * regarding copyright ownership. The ASF licenses this file</span> |
| <span class="source-line-no">006</span><span id="line-6"> * to you under the Apache License, Version 2.0 (the</span> |
| <span class="source-line-no">007</span><span id="line-7"> * "License"); you may not use this file except in compliance</span> |
| <span class="source-line-no">008</span><span id="line-8"> * with the License. You may obtain a copy of the License at</span> |
| <span class="source-line-no">009</span><span id="line-9"> *</span> |
| <span class="source-line-no">010</span><span id="line-10"> * http://www.apache.org/licenses/LICENSE-2.0</span> |
| <span class="source-line-no">011</span><span id="line-11"> *</span> |
| <span class="source-line-no">012</span><span id="line-12"> * Unless required by applicable law or agreed to in writing, software</span> |
| <span class="source-line-no">013</span><span id="line-13"> * distributed under the License is distributed on an "AS IS" BASIS,</span> |
| <span class="source-line-no">014</span><span id="line-14"> * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.</span> |
| <span class="source-line-no">015</span><span id="line-15"> * See the License for the specific language governing permissions and</span> |
| <span class="source-line-no">016</span><span id="line-16"> * limitations under the License.</span> |
| <span class="source-line-no">017</span><span id="line-17"> */</span> |
| <span class="source-line-no">018</span><span id="line-18">package org.apache.hadoop.hbase.mapreduce;</span> |
| <span class="source-line-no">019</span><span id="line-19"></span> |
| <span class="source-line-no">020</span><span id="line-20">import java.io.DataInput;</span> |
| <span class="source-line-no">021</span><span id="line-21">import java.io.DataOutput;</span> |
| <span class="source-line-no">022</span><span id="line-22">import java.io.EOFException;</span> |
| <span class="source-line-no">023</span><span id="line-23">import java.io.FileNotFoundException;</span> |
| <span class="source-line-no">024</span><span id="line-24">import java.io.IOException;</span> |
| <span class="source-line-no">025</span><span id="line-25">import java.time.Instant;</span> |
| <span class="source-line-no">026</span><span id="line-26">import java.util.ArrayList;</span> |
| <span class="source-line-no">027</span><span id="line-27">import java.util.Collections;</span> |
| <span class="source-line-no">028</span><span id="line-28">import java.util.List;</span> |
| <span class="source-line-no">029</span><span id="line-29">import org.apache.hadoop.conf.Configuration;</span> |
| <span class="source-line-no">030</span><span id="line-30">import org.apache.hadoop.fs.FileStatus;</span> |
| <span class="source-line-no">031</span><span id="line-31">import org.apache.hadoop.fs.FileSystem;</span> |
| <span class="source-line-no">032</span><span id="line-32">import org.apache.hadoop.fs.LocatedFileStatus;</span> |
| <span class="source-line-no">033</span><span id="line-33">import org.apache.hadoop.fs.Path;</span> |
| <span class="source-line-no">034</span><span id="line-34">import org.apache.hadoop.fs.RemoteIterator;</span> |
| <span class="source-line-no">035</span><span id="line-35">import org.apache.hadoop.hbase.util.LeaseNotRecoveredException;</span> |
| <span class="source-line-no">036</span><span id="line-36">import org.apache.hadoop.hbase.wal.AbstractFSWALProvider;</span> |
| <span class="source-line-no">037</span><span id="line-37">import org.apache.hadoop.hbase.wal.WAL;</span> |
| <span class="source-line-no">038</span><span id="line-38">import org.apache.hadoop.hbase.wal.WAL.Entry;</span> |
| <span class="source-line-no">039</span><span id="line-39">import org.apache.hadoop.hbase.wal.WALEdit;</span> |
| <span class="source-line-no">040</span><span id="line-40">import org.apache.hadoop.hbase.wal.WALFactory;</span> |
| <span class="source-line-no">041</span><span id="line-41">import org.apache.hadoop.hbase.wal.WALKey;</span> |
| <span class="source-line-no">042</span><span id="line-42">import org.apache.hadoop.hbase.wal.WALStreamReader;</span> |
| <span class="source-line-no">043</span><span id="line-43">import org.apache.hadoop.io.Writable;</span> |
| <span class="source-line-no">044</span><span id="line-44">import org.apache.hadoop.mapreduce.InputFormat;</span> |
| <span class="source-line-no">045</span><span id="line-45">import org.apache.hadoop.mapreduce.InputSplit;</span> |
| <span class="source-line-no">046</span><span id="line-46">import org.apache.hadoop.mapreduce.JobContext;</span> |
| <span class="source-line-no">047</span><span id="line-47">import org.apache.hadoop.mapreduce.RecordReader;</span> |
| <span class="source-line-no">048</span><span id="line-48">import org.apache.hadoop.mapreduce.TaskAttemptContext;</span> |
| <span class="source-line-no">049</span><span id="line-49">import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;</span> |
| <span class="source-line-no">050</span><span id="line-50">import org.apache.hadoop.mapreduce.security.TokenCache;</span> |
| <span class="source-line-no">051</span><span id="line-51">import org.apache.hadoop.util.StringUtils;</span> |
| <span class="source-line-no">052</span><span id="line-52">import org.apache.yetus.audience.InterfaceAudience;</span> |
| <span class="source-line-no">053</span><span id="line-53">import org.slf4j.Logger;</span> |
| <span class="source-line-no">054</span><span id="line-54">import org.slf4j.LoggerFactory;</span> |
| <span class="source-line-no">055</span><span id="line-55"></span> |
| <span class="source-line-no">056</span><span id="line-56">/**</span> |
| <span class="source-line-no">057</span><span id="line-57"> * Simple {@link InputFormat} for {@link org.apache.hadoop.hbase.wal.WAL} files.</span> |
| <span class="source-line-no">058</span><span id="line-58"> */</span> |
| <span class="source-line-no">059</span><span id="line-59">@InterfaceAudience.Public</span> |
| <span class="source-line-no">060</span><span id="line-60">public class WALInputFormat extends InputFormat<WALKey, WALEdit> {</span> |
| <span class="source-line-no">061</span><span id="line-61"> private static final Logger LOG = LoggerFactory.getLogger(WALInputFormat.class);</span> |
| <span class="source-line-no">062</span><span id="line-62"></span> |
| <span class="source-line-no">063</span><span id="line-63"> public static final String START_TIME_KEY = "wal.start.time";</span> |
| <span class="source-line-no">064</span><span id="line-64"> public static final String END_TIME_KEY = "wal.end.time";</span> |
| <span class="source-line-no">065</span><span id="line-65"></span> |
| <span class="source-line-no">066</span><span id="line-66"> /**</span> |
| <span class="source-line-no">067</span><span id="line-67"> * {@link InputSplit} for {@link WAL} files. Each split represent exactly one log file.</span> |
| <span class="source-line-no">068</span><span id="line-68"> */</span> |
| <span class="source-line-no">069</span><span id="line-69"> static class WALSplit extends InputSplit implements Writable {</span> |
| <span class="source-line-no">070</span><span id="line-70"> private String logFileName;</span> |
| <span class="source-line-no">071</span><span id="line-71"> private long fileSize;</span> |
| <span class="source-line-no">072</span><span id="line-72"> private long startTime;</span> |
| <span class="source-line-no">073</span><span id="line-73"> private long endTime;</span> |
| <span class="source-line-no">074</span><span id="line-74"></span> |
| <span class="source-line-no">075</span><span id="line-75"> /** for serialization */</span> |
| <span class="source-line-no">076</span><span id="line-76"> public WALSplit() {</span> |
| <span class="source-line-no">077</span><span id="line-77"> }</span> |
| <span class="source-line-no">078</span><span id="line-78"></span> |
| <span class="source-line-no">079</span><span id="line-79"> /**</span> |
| <span class="source-line-no">080</span><span id="line-80"> * Represent an WALSplit, i.e. a single WAL file. Start- and EndTime are managed by the split,</span> |
| <span class="source-line-no">081</span><span id="line-81"> * so that WAL files can be filtered before WALEdits are passed to the mapper(s).</span> |
| <span class="source-line-no">082</span><span id="line-82"> */</span> |
| <span class="source-line-no">083</span><span id="line-83"> public WALSplit(String logFileName, long fileSize, long startTime, long endTime) {</span> |
| <span class="source-line-no">084</span><span id="line-84"> this.logFileName = logFileName;</span> |
| <span class="source-line-no">085</span><span id="line-85"> this.fileSize = fileSize;</span> |
| <span class="source-line-no">086</span><span id="line-86"> this.startTime = startTime;</span> |
| <span class="source-line-no">087</span><span id="line-87"> this.endTime = endTime;</span> |
| <span class="source-line-no">088</span><span id="line-88"> }</span> |
| <span class="source-line-no">089</span><span id="line-89"></span> |
| <span class="source-line-no">090</span><span id="line-90"> @Override</span> |
| <span class="source-line-no">091</span><span id="line-91"> public long getLength() throws IOException, InterruptedException {</span> |
| <span class="source-line-no">092</span><span id="line-92"> return fileSize;</span> |
| <span class="source-line-no">093</span><span id="line-93"> }</span> |
| <span class="source-line-no">094</span><span id="line-94"></span> |
| <span class="source-line-no">095</span><span id="line-95"> @Override</span> |
| <span class="source-line-no">096</span><span id="line-96"> public String[] getLocations() throws IOException, InterruptedException {</span> |
| <span class="source-line-no">097</span><span id="line-97"> // TODO: Find the data node with the most blocks for this WAL?</span> |
| <span class="source-line-no">098</span><span id="line-98"> return new String[] {};</span> |
| <span class="source-line-no">099</span><span id="line-99"> }</span> |
| <span class="source-line-no">100</span><span id="line-100"></span> |
| <span class="source-line-no">101</span><span id="line-101"> public String getLogFileName() {</span> |
| <span class="source-line-no">102</span><span id="line-102"> return logFileName;</span> |
| <span class="source-line-no">103</span><span id="line-103"> }</span> |
| <span class="source-line-no">104</span><span id="line-104"></span> |
| <span class="source-line-no">105</span><span id="line-105"> public long getStartTime() {</span> |
| <span class="source-line-no">106</span><span id="line-106"> return startTime;</span> |
| <span class="source-line-no">107</span><span id="line-107"> }</span> |
| <span class="source-line-no">108</span><span id="line-108"></span> |
| <span class="source-line-no">109</span><span id="line-109"> public long getEndTime() {</span> |
| <span class="source-line-no">110</span><span id="line-110"> return endTime;</span> |
| <span class="source-line-no">111</span><span id="line-111"> }</span> |
| <span class="source-line-no">112</span><span id="line-112"></span> |
| <span class="source-line-no">113</span><span id="line-113"> @Override</span> |
| <span class="source-line-no">114</span><span id="line-114"> public void readFields(DataInput in) throws IOException {</span> |
| <span class="source-line-no">115</span><span id="line-115"> logFileName = in.readUTF();</span> |
| <span class="source-line-no">116</span><span id="line-116"> fileSize = in.readLong();</span> |
| <span class="source-line-no">117</span><span id="line-117"> startTime = in.readLong();</span> |
| <span class="source-line-no">118</span><span id="line-118"> endTime = in.readLong();</span> |
| <span class="source-line-no">119</span><span id="line-119"> }</span> |
| <span class="source-line-no">120</span><span id="line-120"></span> |
| <span class="source-line-no">121</span><span id="line-121"> @Override</span> |
| <span class="source-line-no">122</span><span id="line-122"> public void write(DataOutput out) throws IOException {</span> |
| <span class="source-line-no">123</span><span id="line-123"> out.writeUTF(logFileName);</span> |
| <span class="source-line-no">124</span><span id="line-124"> out.writeLong(fileSize);</span> |
| <span class="source-line-no">125</span><span id="line-125"> out.writeLong(startTime);</span> |
| <span class="source-line-no">126</span><span id="line-126"> out.writeLong(endTime);</span> |
| <span class="source-line-no">127</span><span id="line-127"> }</span> |
| <span class="source-line-no">128</span><span id="line-128"></span> |
| <span class="source-line-no">129</span><span id="line-129"> @Override</span> |
| <span class="source-line-no">130</span><span id="line-130"> public String toString() {</span> |
| <span class="source-line-no">131</span><span id="line-131"> return logFileName + " (" + startTime + ":" + endTime + ") length:" + fileSize;</span> |
| <span class="source-line-no">132</span><span id="line-132"> }</span> |
| <span class="source-line-no">133</span><span id="line-133"> }</span> |
| <span class="source-line-no">134</span><span id="line-134"></span> |
| <span class="source-line-no">135</span><span id="line-135"> /**</span> |
| <span class="source-line-no">136</span><span id="line-136"> * {@link RecordReader} for an {@link WAL} file. Implementation shared with deprecated</span> |
| <span class="source-line-no">137</span><span id="line-137"> * HLogInputFormat.</span> |
| <span class="source-line-no">138</span><span id="line-138"> */</span> |
| <span class="source-line-no">139</span><span id="line-139"> static abstract class WALRecordReader<K extends WALKey> extends RecordReader<K, WALEdit> {</span> |
| <span class="source-line-no">140</span><span id="line-140"> private WALStreamReader reader = null;</span> |
| <span class="source-line-no">141</span><span id="line-141"> // visible until we can remove the deprecated HLogInputFormat</span> |
| <span class="source-line-no">142</span><span id="line-142"> Entry currentEntry = new Entry();</span> |
| <span class="source-line-no">143</span><span id="line-143"> private long startTime;</span> |
| <span class="source-line-no">144</span><span id="line-144"> private long endTime;</span> |
| <span class="source-line-no">145</span><span id="line-145"> private Configuration conf;</span> |
| <span class="source-line-no">146</span><span id="line-146"> private Path logFile;</span> |
| <span class="source-line-no">147</span><span id="line-147"> private long currentPos;</span> |
| <span class="source-line-no">148</span><span id="line-148"></span> |
| <span class="source-line-no">149</span><span id="line-149"> @edu.umd.cs.findbugs.annotations.SuppressWarnings(value = "DCN_NULLPOINTER_EXCEPTION",</span> |
| <span class="source-line-no">150</span><span id="line-150"> justification = "HDFS-4380")</span> |
| <span class="source-line-no">151</span><span id="line-151"> private WALStreamReader openReader(Path path, long startPosition) throws IOException {</span> |
| <span class="source-line-no">152</span><span id="line-152"> long retryInterval = 2000; // 2 sec</span> |
| <span class="source-line-no">153</span><span id="line-153"> int maxAttempts = 30;</span> |
| <span class="source-line-no">154</span><span id="line-154"> int attempt = 0;</span> |
| <span class="source-line-no">155</span><span id="line-155"> Exception ee = null;</span> |
| <span class="source-line-no">156</span><span id="line-156"> WALStreamReader reader = null;</span> |
| <span class="source-line-no">157</span><span id="line-157"> while (reader == null && attempt++ < maxAttempts) {</span> |
| <span class="source-line-no">158</span><span id="line-158"> try {</span> |
| <span class="source-line-no">159</span><span id="line-159"> // Detect if this is a new file, if so get a new reader else</span> |
| <span class="source-line-no">160</span><span id="line-160"> // reset the current reader so that we see the new data</span> |
| <span class="source-line-no">161</span><span id="line-161"> reader =</span> |
| <span class="source-line-no">162</span><span id="line-162"> WALFactory.createStreamReader(path.getFileSystem(conf), path, conf, startPosition);</span> |
| <span class="source-line-no">163</span><span id="line-163"> return reader;</span> |
| <span class="source-line-no">164</span><span id="line-164"> } catch (LeaseNotRecoveredException lnre) {</span> |
| <span class="source-line-no">165</span><span id="line-165"> // HBASE-15019 the WAL was not closed due to some hiccup.</span> |
| <span class="source-line-no">166</span><span id="line-166"> LOG.warn("Try to recover the WAL lease " + path, lnre);</span> |
| <span class="source-line-no">167</span><span id="line-167"> AbstractFSWALProvider.recoverLease(conf, path);</span> |
| <span class="source-line-no">168</span><span id="line-168"> reader = null;</span> |
| <span class="source-line-no">169</span><span id="line-169"> ee = lnre;</span> |
| <span class="source-line-no">170</span><span id="line-170"> } catch (NullPointerException npe) {</span> |
| <span class="source-line-no">171</span><span id="line-171"> // Workaround for race condition in HDFS-4380</span> |
| <span class="source-line-no">172</span><span id="line-172"> // which throws a NPE if we open a file before any data node has the most recent block</span> |
| <span class="source-line-no">173</span><span id="line-173"> // Just sleep and retry. Will require re-reading compressed WALs for compressionContext.</span> |
| <span class="source-line-no">174</span><span id="line-174"> LOG.warn("Got NPE opening reader, will retry.");</span> |
| <span class="source-line-no">175</span><span id="line-175"> reader = null;</span> |
| <span class="source-line-no">176</span><span id="line-176"> ee = npe;</span> |
| <span class="source-line-no">177</span><span id="line-177"> }</span> |
| <span class="source-line-no">178</span><span id="line-178"> if (reader == null) {</span> |
| <span class="source-line-no">179</span><span id="line-179"> // sleep before next attempt</span> |
| <span class="source-line-no">180</span><span id="line-180"> try {</span> |
| <span class="source-line-no">181</span><span id="line-181"> Thread.sleep(retryInterval);</span> |
| <span class="source-line-no">182</span><span id="line-182"> } catch (InterruptedException e) {</span> |
| <span class="source-line-no">183</span><span id="line-183"> Thread.currentThread().interrupt();</span> |
| <span class="source-line-no">184</span><span id="line-184"> }</span> |
| <span class="source-line-no">185</span><span id="line-185"> }</span> |
| <span class="source-line-no">186</span><span id="line-186"> }</span> |
| <span class="source-line-no">187</span><span id="line-187"> throw new IOException("Could not open reader", ee);</span> |
| <span class="source-line-no">188</span><span id="line-188"> }</span> |
| <span class="source-line-no">189</span><span id="line-189"></span> |
| <span class="source-line-no">190</span><span id="line-190"> @Override</span> |
| <span class="source-line-no">191</span><span id="line-191"> public void initialize(InputSplit split, TaskAttemptContext context)</span> |
| <span class="source-line-no">192</span><span id="line-192"> throws IOException, InterruptedException {</span> |
| <span class="source-line-no">193</span><span id="line-193"> WALSplit hsplit = (WALSplit) split;</span> |
| <span class="source-line-no">194</span><span id="line-194"> logFile = new Path(hsplit.getLogFileName());</span> |
| <span class="source-line-no">195</span><span id="line-195"> conf = context.getConfiguration();</span> |
| <span class="source-line-no">196</span><span id="line-196"> LOG.info("Opening {} for {}", logFile, split);</span> |
| <span class="source-line-no">197</span><span id="line-197"> openReader(logFile);</span> |
| <span class="source-line-no">198</span><span id="line-198"> this.startTime = hsplit.getStartTime();</span> |
| <span class="source-line-no">199</span><span id="line-199"> this.endTime = hsplit.getEndTime();</span> |
| <span class="source-line-no">200</span><span id="line-200"> }</span> |
| <span class="source-line-no">201</span><span id="line-201"></span> |
| <span class="source-line-no">202</span><span id="line-202"> private void openReader(Path path) throws IOException {</span> |
| <span class="source-line-no">203</span><span id="line-203"> closeReader();</span> |
| <span class="source-line-no">204</span><span id="line-204"> reader = openReader(path, currentPos > 0 ? currentPos : -1);</span> |
| <span class="source-line-no">205</span><span id="line-205"> setCurrentPath(path);</span> |
| <span class="source-line-no">206</span><span id="line-206"> }</span> |
| <span class="source-line-no">207</span><span id="line-207"></span> |
| <span class="source-line-no">208</span><span id="line-208"> private void setCurrentPath(Path path) {</span> |
| <span class="source-line-no">209</span><span id="line-209"> this.logFile = path;</span> |
| <span class="source-line-no">210</span><span id="line-210"> }</span> |
| <span class="source-line-no">211</span><span id="line-211"></span> |
| <span class="source-line-no">212</span><span id="line-212"> private void closeReader() throws IOException {</span> |
| <span class="source-line-no">213</span><span id="line-213"> if (reader != null) {</span> |
| <span class="source-line-no">214</span><span id="line-214"> reader.close();</span> |
| <span class="source-line-no">215</span><span id="line-215"> reader = null;</span> |
| <span class="source-line-no">216</span><span id="line-216"> }</span> |
| <span class="source-line-no">217</span><span id="line-217"> }</span> |
| <span class="source-line-no">218</span><span id="line-218"></span> |
| <span class="source-line-no">219</span><span id="line-219"> @Override</span> |
| <span class="source-line-no">220</span><span id="line-220"> public boolean nextKeyValue() throws IOException, InterruptedException {</span> |
| <span class="source-line-no">221</span><span id="line-221"> if (reader == null) {</span> |
| <span class="source-line-no">222</span><span id="line-222"> return false;</span> |
| <span class="source-line-no">223</span><span id="line-223"> }</span> |
| <span class="source-line-no">224</span><span id="line-224"> this.currentPos = reader.getPosition();</span> |
| <span class="source-line-no">225</span><span id="line-225"> Entry temp;</span> |
| <span class="source-line-no">226</span><span id="line-226"> long i = -1;</span> |
| <span class="source-line-no">227</span><span id="line-227"> try {</span> |
| <span class="source-line-no">228</span><span id="line-228"> do {</span> |
| <span class="source-line-no">229</span><span id="line-229"> // skip older entries</span> |
| <span class="source-line-no">230</span><span id="line-230"> try {</span> |
| <span class="source-line-no">231</span><span id="line-231"> temp = reader.next(currentEntry);</span> |
| <span class="source-line-no">232</span><span id="line-232"> i++;</span> |
| <span class="source-line-no">233</span><span id="line-233"> } catch (EOFException x) {</span> |
| <span class="source-line-no">234</span><span id="line-234"> LOG.warn("Corrupted entry detected. Ignoring the rest of the file."</span> |
| <span class="source-line-no">235</span><span id="line-235"> + " (This is normal when a RegionServer crashed.)");</span> |
| <span class="source-line-no">236</span><span id="line-236"> return false;</span> |
| <span class="source-line-no">237</span><span id="line-237"> }</span> |
| <span class="source-line-no">238</span><span id="line-238"> } while (temp != null && temp.getKey().getWriteTime() < startTime);</span> |
| <span class="source-line-no">239</span><span id="line-239"></span> |
| <span class="source-line-no">240</span><span id="line-240"> if (temp == null) {</span> |
| <span class="source-line-no">241</span><span id="line-241"> if (i > 0) {</span> |
| <span class="source-line-no">242</span><span id="line-242"> LOG.info("Skipped " + i + " entries.");</span> |
| <span class="source-line-no">243</span><span id="line-243"> }</span> |
| <span class="source-line-no">244</span><span id="line-244"> LOG.info("Reached end of file.");</span> |
| <span class="source-line-no">245</span><span id="line-245"> return false;</span> |
| <span class="source-line-no">246</span><span id="line-246"> } else if (i > 0) {</span> |
| <span class="source-line-no">247</span><span id="line-247"> LOG.info("Skipped " + i + " entries, until ts: " + temp.getKey().getWriteTime() + ".");</span> |
| <span class="source-line-no">248</span><span id="line-248"> }</span> |
| <span class="source-line-no">249</span><span id="line-249"> boolean res = temp.getKey().getWriteTime() <= endTime;</span> |
| <span class="source-line-no">250</span><span id="line-250"> if (!res) {</span> |
| <span class="source-line-no">251</span><span id="line-251"> LOG.info(</span> |
| <span class="source-line-no">252</span><span id="line-252"> "Reached ts: " + temp.getKey().getWriteTime() + " ignoring the rest of the file.");</span> |
| <span class="source-line-no">253</span><span id="line-253"> }</span> |
| <span class="source-line-no">254</span><span id="line-254"> return res;</span> |
| <span class="source-line-no">255</span><span id="line-255"> } catch (IOException e) {</span> |
| <span class="source-line-no">256</span><span id="line-256"> Path archivedLog = AbstractFSWALProvider.findArchivedLog(logFile, conf);</span> |
| <span class="source-line-no">257</span><span id="line-257"> // archivedLog can be null if unable to locate in archiveDir.</span> |
| <span class="source-line-no">258</span><span id="line-258"> if (archivedLog != null) {</span> |
| <span class="source-line-no">259</span><span id="line-259"> openReader(archivedLog);</span> |
| <span class="source-line-no">260</span><span id="line-260"> // Try call again in recursion</span> |
| <span class="source-line-no">261</span><span id="line-261"> return nextKeyValue();</span> |
| <span class="source-line-no">262</span><span id="line-262"> } else {</span> |
| <span class="source-line-no">263</span><span id="line-263"> throw e;</span> |
| <span class="source-line-no">264</span><span id="line-264"> }</span> |
| <span class="source-line-no">265</span><span id="line-265"> }</span> |
| <span class="source-line-no">266</span><span id="line-266"> }</span> |
| <span class="source-line-no">267</span><span id="line-267"></span> |
| <span class="source-line-no">268</span><span id="line-268"> @Override</span> |
| <span class="source-line-no">269</span><span id="line-269"> public WALEdit getCurrentValue() throws IOException, InterruptedException {</span> |
| <span class="source-line-no">270</span><span id="line-270"> return currentEntry.getEdit();</span> |
| <span class="source-line-no">271</span><span id="line-271"> }</span> |
| <span class="source-line-no">272</span><span id="line-272"></span> |
| <span class="source-line-no">273</span><span id="line-273"> @Override</span> |
| <span class="source-line-no">274</span><span id="line-274"> public float getProgress() throws IOException, InterruptedException {</span> |
| <span class="source-line-no">275</span><span id="line-275"> // N/A depends on total number of entries, which is unknown</span> |
| <span class="source-line-no">276</span><span id="line-276"> return 0;</span> |
| <span class="source-line-no">277</span><span id="line-277"> }</span> |
| <span class="source-line-no">278</span><span id="line-278"></span> |
| <span class="source-line-no">279</span><span id="line-279"> @Override</span> |
| <span class="source-line-no">280</span><span id="line-280"> public void close() throws IOException {</span> |
| <span class="source-line-no">281</span><span id="line-281"> LOG.info("Closing reader");</span> |
| <span class="source-line-no">282</span><span id="line-282"> if (reader != null) {</span> |
| <span class="source-line-no">283</span><span id="line-283"> this.reader.close();</span> |
| <span class="source-line-no">284</span><span id="line-284"> }</span> |
| <span class="source-line-no">285</span><span id="line-285"> }</span> |
| <span class="source-line-no">286</span><span id="line-286"> }</span> |
| <span class="source-line-no">287</span><span id="line-287"></span> |
| <span class="source-line-no">288</span><span id="line-288"> /**</span> |
| <span class="source-line-no">289</span><span id="line-289"> * handler for non-deprecated WALKey version. fold into WALRecordReader once we no longer need to</span> |
| <span class="source-line-no">290</span><span id="line-290"> * support HLogInputFormat.</span> |
| <span class="source-line-no">291</span><span id="line-291"> */</span> |
| <span class="source-line-no">292</span><span id="line-292"> static class WALKeyRecordReader extends WALRecordReader<WALKey> {</span> |
| <span class="source-line-no">293</span><span id="line-293"> @Override</span> |
| <span class="source-line-no">294</span><span id="line-294"> public WALKey getCurrentKey() throws IOException, InterruptedException {</span> |
| <span class="source-line-no">295</span><span id="line-295"> return currentEntry.getKey();</span> |
| <span class="source-line-no">296</span><span id="line-296"> }</span> |
| <span class="source-line-no">297</span><span id="line-297"> }</span> |
| <span class="source-line-no">298</span><span id="line-298"></span> |
| <span class="source-line-no">299</span><span id="line-299"> @Override</span> |
| <span class="source-line-no">300</span><span id="line-300"> public List<InputSplit> getSplits(JobContext context) throws IOException, InterruptedException {</span> |
| <span class="source-line-no">301</span><span id="line-301"> return getSplits(context, START_TIME_KEY, END_TIME_KEY);</span> |
| <span class="source-line-no">302</span><span id="line-302"> }</span> |
| <span class="source-line-no">303</span><span id="line-303"></span> |
| <span class="source-line-no">304</span><span id="line-304"> /**</span> |
| <span class="source-line-no">305</span><span id="line-305"> * implementation shared with deprecated HLogInputFormat</span> |
| <span class="source-line-no">306</span><span id="line-306"> */</span> |
| <span class="source-line-no">307</span><span id="line-307"> List<InputSplit> getSplits(final JobContext context, final String startKey, final String endKey)</span> |
| <span class="source-line-no">308</span><span id="line-308"> throws IOException, InterruptedException {</span> |
| <span class="source-line-no">309</span><span id="line-309"> Configuration conf = context.getConfiguration();</span> |
| <span class="source-line-no">310</span><span id="line-310"> boolean ignoreMissing = conf.getBoolean(WALPlayer.IGNORE_MISSING_FILES, false);</span> |
| <span class="source-line-no">311</span><span id="line-311"> Path[] inputPaths = getInputPaths(conf);</span> |
| <span class="source-line-no">312</span><span id="line-312"> // get delegation token for the filesystem</span> |
| <span class="source-line-no">313</span><span id="line-313"> TokenCache.obtainTokensForNamenodes(context.getCredentials(), inputPaths, conf);</span> |
| <span class="source-line-no">314</span><span id="line-314"> long startTime = conf.getLong(startKey, Long.MIN_VALUE);</span> |
| <span class="source-line-no">315</span><span id="line-315"> long endTime = conf.getLong(endKey, Long.MAX_VALUE);</span> |
| <span class="source-line-no">316</span><span id="line-316"></span> |
| <span class="source-line-no">317</span><span id="line-317"> List<FileStatus> allFiles = new ArrayList<FileStatus>();</span> |
| <span class="source-line-no">318</span><span id="line-318"> for (Path inputPath : inputPaths) {</span> |
| <span class="source-line-no">319</span><span id="line-319"> FileSystem fs = inputPath.getFileSystem(conf);</span> |
| <span class="source-line-no">320</span><span id="line-320"> try {</span> |
| <span class="source-line-no">321</span><span id="line-321"> List<FileStatus> files = getFiles(fs, inputPath, startTime, endTime);</span> |
| <span class="source-line-no">322</span><span id="line-322"> allFiles.addAll(files);</span> |
| <span class="source-line-no">323</span><span id="line-323"> } catch (FileNotFoundException e) {</span> |
| <span class="source-line-no">324</span><span id="line-324"> if (ignoreMissing) {</span> |
| <span class="source-line-no">325</span><span id="line-325"> LOG.warn("File " + inputPath + " is missing. Skipping it.");</span> |
| <span class="source-line-no">326</span><span id="line-326"> continue;</span> |
| <span class="source-line-no">327</span><span id="line-327"> }</span> |
| <span class="source-line-no">328</span><span id="line-328"> throw e;</span> |
| <span class="source-line-no">329</span><span id="line-329"> }</span> |
| <span class="source-line-no">330</span><span id="line-330"> }</span> |
| <span class="source-line-no">331</span><span id="line-331"> List<InputSplit> splits = new ArrayList<InputSplit>(allFiles.size());</span> |
| <span class="source-line-no">332</span><span id="line-332"> for (FileStatus file : allFiles) {</span> |
| <span class="source-line-no">333</span><span id="line-333"> splits.add(new WALSplit(file.getPath().toString(), file.getLen(), startTime, endTime));</span> |
| <span class="source-line-no">334</span><span id="line-334"> }</span> |
| <span class="source-line-no">335</span><span id="line-335"> return splits;</span> |
| <span class="source-line-no">336</span><span id="line-336"> }</span> |
| <span class="source-line-no">337</span><span id="line-337"></span> |
| <span class="source-line-no">338</span><span id="line-338"> private Path[] getInputPaths(Configuration conf) {</span> |
| <span class="source-line-no">339</span><span id="line-339"> String inpDirs = conf.get(FileInputFormat.INPUT_DIR);</span> |
| <span class="source-line-no">340</span><span id="line-340"> return StringUtils</span> |
| <span class="source-line-no">341</span><span id="line-341"> .stringToPath(inpDirs.split(conf.get(WALPlayer.INPUT_FILES_SEPARATOR_KEY, ",")));</span> |
| <span class="source-line-no">342</span><span id="line-342"> }</span> |
| <span class="source-line-no">343</span><span id="line-343"></span> |
| <span class="source-line-no">344</span><span id="line-344"> /**</span> |
| <span class="source-line-no">345</span><span id="line-345"> * @param startTime If file looks like it has a timestamp in its name, we'll check if newer or</span> |
| <span class="source-line-no">346</span><span id="line-346"> * equal to this value else we will filter out the file. If name does not seem to</span> |
| <span class="source-line-no">347</span><span id="line-347"> * have a timestamp, we will just return it w/o filtering.</span> |
| <span class="source-line-no">348</span><span id="line-348"> * @param endTime If file looks like it has a timestamp in its name, we'll check if older or</span> |
| <span class="source-line-no">349</span><span id="line-349"> * equal to this value else we will filter out the file. If name does not seem to</span> |
| <span class="source-line-no">350</span><span id="line-350"> * have a timestamp, we will just return it w/o filtering.</span> |
| <span class="source-line-no">351</span><span id="line-351"> */</span> |
| <span class="source-line-no">352</span><span id="line-352"> private List<FileStatus> getFiles(FileSystem fs, Path dir, long startTime, long endTime)</span> |
| <span class="source-line-no">353</span><span id="line-353"> throws IOException {</span> |
| <span class="source-line-no">354</span><span id="line-354"> List<FileStatus> result = new ArrayList<>();</span> |
| <span class="source-line-no">355</span><span id="line-355"> LOG.debug("Scanning " + dir.toString() + " for WAL files");</span> |
| <span class="source-line-no">356</span><span id="line-356"> RemoteIterator<LocatedFileStatus> iter = fs.listLocatedStatus(dir);</span> |
| <span class="source-line-no">357</span><span id="line-357"> if (!iter.hasNext()) {</span> |
| <span class="source-line-no">358</span><span id="line-358"> return Collections.emptyList();</span> |
| <span class="source-line-no">359</span><span id="line-359"> }</span> |
| <span class="source-line-no">360</span><span id="line-360"> while (iter.hasNext()) {</span> |
| <span class="source-line-no">361</span><span id="line-361"> LocatedFileStatus file = iter.next();</span> |
| <span class="source-line-no">362</span><span id="line-362"> if (file.isDirectory()) {</span> |
| <span class="source-line-no">363</span><span id="line-363"> // Recurse into sub directories</span> |
| <span class="source-line-no">364</span><span id="line-364"> result.addAll(getFiles(fs, file.getPath(), startTime, endTime));</span> |
| <span class="source-line-no">365</span><span id="line-365"> } else {</span> |
| <span class="source-line-no">366</span><span id="line-366"> addFile(result, file, startTime, endTime);</span> |
| <span class="source-line-no">367</span><span id="line-367"> }</span> |
| <span class="source-line-no">368</span><span id="line-368"> }</span> |
| <span class="source-line-no">369</span><span id="line-369"> // TODO: These results should be sorted? Results could be content of recovered.edits directory</span> |
| <span class="source-line-no">370</span><span id="line-370"> // -- null padded increasing numeric -- or a WAL file w/ timestamp suffix or timestamp and</span> |
| <span class="source-line-no">371</span><span id="line-371"> // then meta suffix. See AbstractFSWALProvider#WALStartTimeComparator</span> |
| <span class="source-line-no">372</span><span id="line-372"> return result;</span> |
| <span class="source-line-no">373</span><span id="line-373"> }</span> |
| <span class="source-line-no">374</span><span id="line-374"></span> |
| <span class="source-line-no">375</span><span id="line-375"> static void addFile(List<FileStatus> result, LocatedFileStatus lfs, long startTime,</span> |
| <span class="source-line-no">376</span><span id="line-376"> long endTime) {</span> |
| <span class="source-line-no">377</span><span id="line-377"> long timestamp = AbstractFSWALProvider.getTimestamp(lfs.getPath().getName());</span> |
| <span class="source-line-no">378</span><span id="line-378"> if (timestamp > 0) {</span> |
| <span class="source-line-no">379</span><span id="line-379"> // Looks like a valid timestamp.</span> |
| <span class="source-line-no">380</span><span id="line-380"> if (timestamp <= endTime && timestamp >= startTime) {</span> |
| <span class="source-line-no">381</span><span id="line-381"> LOG.info("Found {}", lfs.getPath());</span> |
| <span class="source-line-no">382</span><span id="line-382"> result.add(lfs);</span> |
| <span class="source-line-no">383</span><span id="line-383"> } else {</span> |
| <span class="source-line-no">384</span><span id="line-384"> LOG.info("Skipped {}, outside range [{}/{} - {}/{}]", lfs.getPath(), startTime,</span> |
| <span class="source-line-no">385</span><span id="line-385"> Instant.ofEpochMilli(startTime), endTime, Instant.ofEpochMilli(endTime));</span> |
| <span class="source-line-no">386</span><span id="line-386"> }</span> |
| <span class="source-line-no">387</span><span id="line-387"> } else {</span> |
| <span class="source-line-no">388</span><span id="line-388"> // If no timestamp, add it regardless.</span> |
| <span class="source-line-no">389</span><span id="line-389"> LOG.info("Found (no-timestamp!) {}", lfs);</span> |
| <span class="source-line-no">390</span><span id="line-390"> result.add(lfs);</span> |
| <span class="source-line-no">391</span><span id="line-391"> }</span> |
| <span class="source-line-no">392</span><span id="line-392"> }</span> |
| <span class="source-line-no">393</span><span id="line-393"></span> |
| <span class="source-line-no">394</span><span id="line-394"> @Override</span> |
| <span class="source-line-no">395</span><span id="line-395"> public RecordReader<WALKey, WALEdit> createRecordReader(InputSplit split,</span> |
| <span class="source-line-no">396</span><span id="line-396"> TaskAttemptContext context) throws IOException, InterruptedException {</span> |
| <span class="source-line-no">397</span><span id="line-397"> return new WALKeyRecordReader();</span> |
| <span class="source-line-no">398</span><span id="line-398"> }</span> |
| <span class="source-line-no">399</span><span id="line-399">}</span> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| </pre> |
| </div> |
| </main> |
| </body> |
| </html> |