| <!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.regionserver.compactions;<a name="line.18"></a> |
| <span class="sourceLineNo">019</span><a name="line.19"></a> |
| <span class="sourceLineNo">020</span>import java.io.IOException;<a name="line.20"></a> |
| <span class="sourceLineNo">021</span>import java.util.ArrayList;<a name="line.21"></a> |
| <span class="sourceLineNo">022</span>import java.util.List;<a name="line.22"></a> |
| <span class="sourceLineNo">023</span>import org.apache.hadoop.conf.Configuration;<a name="line.23"></a> |
| <span class="sourceLineNo">024</span>import org.apache.hadoop.hbase.regionserver.HStoreFile;<a name="line.24"></a> |
| <span class="sourceLineNo">025</span>import org.apache.hadoop.hbase.regionserver.StoreConfigInformation;<a name="line.25"></a> |
| <span class="sourceLineNo">026</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.26"></a> |
| <span class="sourceLineNo">027</span>import org.slf4j.Logger;<a name="line.27"></a> |
| <span class="sourceLineNo">028</span>import org.slf4j.LoggerFactory;<a name="line.28"></a> |
| <span class="sourceLineNo">029</span><a name="line.29"></a> |
| <span class="sourceLineNo">030</span>/**<a name="line.30"></a> |
| <span class="sourceLineNo">031</span> * Class to pick which files if any to compact together. This class will search all possibilities<a name="line.31"></a> |
| <span class="sourceLineNo">032</span> * for different and if it gets stuck it will choose the smallest set of files to compact.<a name="line.32"></a> |
| <span class="sourceLineNo">033</span> */<a name="line.33"></a> |
| <span class="sourceLineNo">034</span>@InterfaceAudience.Private<a name="line.34"></a> |
| <span class="sourceLineNo">035</span>public class ExploringCompactionPolicy extends RatioBasedCompactionPolicy {<a name="line.35"></a> |
| <span class="sourceLineNo">036</span> private static final Logger LOG = LoggerFactory.getLogger(ExploringCompactionPolicy.class);<a name="line.36"></a> |
| <span class="sourceLineNo">037</span><a name="line.37"></a> |
| <span class="sourceLineNo">038</span> /**<a name="line.38"></a> |
| <span class="sourceLineNo">039</span> * Constructor for ExploringCompactionPolicy.<a name="line.39"></a> |
| <span class="sourceLineNo">040</span> * @param conf The configuration object<a name="line.40"></a> |
| <span class="sourceLineNo">041</span> * @param storeConfigInfo An object to provide info about the store.<a name="line.41"></a> |
| <span class="sourceLineNo">042</span> */<a name="line.42"></a> |
| <span class="sourceLineNo">043</span> public ExploringCompactionPolicy(final Configuration conf,<a name="line.43"></a> |
| <span class="sourceLineNo">044</span> final StoreConfigInformation storeConfigInfo) {<a name="line.44"></a> |
| <span class="sourceLineNo">045</span> super(conf, storeConfigInfo);<a name="line.45"></a> |
| <span class="sourceLineNo">046</span> }<a name="line.46"></a> |
| <span class="sourceLineNo">047</span><a name="line.47"></a> |
| <span class="sourceLineNo">048</span> @Override<a name="line.48"></a> |
| <span class="sourceLineNo">049</span> protected final ArrayList<HStoreFile> applyCompactionPolicy(ArrayList<HStoreFile> candidates,<a name="line.49"></a> |
| <span class="sourceLineNo">050</span> boolean mayUseOffPeak, boolean mightBeStuck) throws IOException {<a name="line.50"></a> |
| <span class="sourceLineNo">051</span> return new ArrayList<>(applyCompactionPolicy(candidates, mightBeStuck, mayUseOffPeak,<a name="line.51"></a> |
| <span class="sourceLineNo">052</span> comConf.getMinFilesToCompact(), comConf.getMaxFilesToCompact()));<a name="line.52"></a> |
| <span class="sourceLineNo">053</span> }<a name="line.53"></a> |
| <span class="sourceLineNo">054</span><a name="line.54"></a> |
| <span class="sourceLineNo">055</span> public List<HStoreFile> applyCompactionPolicy(List<HStoreFile> candidates, boolean mightBeStuck,<a name="line.55"></a> |
| <span class="sourceLineNo">056</span> boolean mayUseOffPeak, int minFiles, int maxFiles) {<a name="line.56"></a> |
| <span class="sourceLineNo">057</span> final double currentRatio =<a name="line.57"></a> |
| <span class="sourceLineNo">058</span> mayUseOffPeak ? comConf.getCompactionRatioOffPeak() : comConf.getCompactionRatio();<a name="line.58"></a> |
| <span class="sourceLineNo">059</span><a name="line.59"></a> |
| <span class="sourceLineNo">060</span> // Start off choosing nothing.<a name="line.60"></a> |
| <span class="sourceLineNo">061</span> List<HStoreFile> bestSelection = new ArrayList<>(0);<a name="line.61"></a> |
| <span class="sourceLineNo">062</span> List<HStoreFile> smallest = mightBeStuck ? new ArrayList<>(0) : null;<a name="line.62"></a> |
| <span class="sourceLineNo">063</span> long bestSize = 0;<a name="line.63"></a> |
| <span class="sourceLineNo">064</span> long smallestSize = Long.MAX_VALUE;<a name="line.64"></a> |
| <span class="sourceLineNo">065</span><a name="line.65"></a> |
| <span class="sourceLineNo">066</span> int opts = 0, optsInRatio = 0, bestStart = -1; // for debug logging<a name="line.66"></a> |
| <span class="sourceLineNo">067</span> // Consider every starting place.<a name="line.67"></a> |
| <span class="sourceLineNo">068</span> for (int start = 0; start < candidates.size(); start++) {<a name="line.68"></a> |
| <span class="sourceLineNo">069</span> // Consider every different sub list permutation in between start and end with min files.<a name="line.69"></a> |
| <span class="sourceLineNo">070</span> for (int currentEnd = start + minFiles - 1; currentEnd < candidates.size(); currentEnd++) {<a name="line.70"></a> |
| <span class="sourceLineNo">071</span> List<HStoreFile> potentialMatchFiles = candidates.subList(start, currentEnd + 1);<a name="line.71"></a> |
| <span class="sourceLineNo">072</span><a name="line.72"></a> |
| <span class="sourceLineNo">073</span> // Sanity checks<a name="line.73"></a> |
| <span class="sourceLineNo">074</span> if (potentialMatchFiles.size() < minFiles) {<a name="line.74"></a> |
| <span class="sourceLineNo">075</span> continue;<a name="line.75"></a> |
| <span class="sourceLineNo">076</span> }<a name="line.76"></a> |
| <span class="sourceLineNo">077</span> if (potentialMatchFiles.size() > maxFiles) {<a name="line.77"></a> |
| <span class="sourceLineNo">078</span> continue;<a name="line.78"></a> |
| <span class="sourceLineNo">079</span> }<a name="line.79"></a> |
| <span class="sourceLineNo">080</span><a name="line.80"></a> |
| <span class="sourceLineNo">081</span> // Compute the total size of files that will<a name="line.81"></a> |
| <span class="sourceLineNo">082</span> // have to be read if this set of files is compacted.<a name="line.82"></a> |
| <span class="sourceLineNo">083</span> long size = getTotalStoreSize(potentialMatchFiles);<a name="line.83"></a> |
| <span class="sourceLineNo">084</span><a name="line.84"></a> |
| <span class="sourceLineNo">085</span> // Store the smallest set of files. This stored set of files will be used<a name="line.85"></a> |
| <span class="sourceLineNo">086</span> // if it looks like the algorithm is stuck.<a name="line.86"></a> |
| <span class="sourceLineNo">087</span> if (mightBeStuck && size < smallestSize) {<a name="line.87"></a> |
| <span class="sourceLineNo">088</span> smallest = potentialMatchFiles;<a name="line.88"></a> |
| <span class="sourceLineNo">089</span> smallestSize = size;<a name="line.89"></a> |
| <span class="sourceLineNo">090</span> }<a name="line.90"></a> |
| <span class="sourceLineNo">091</span><a name="line.91"></a> |
| <span class="sourceLineNo">092</span> if (size > comConf.getMaxCompactSize(mayUseOffPeak)) {<a name="line.92"></a> |
| <span class="sourceLineNo">093</span> continue;<a name="line.93"></a> |
| <span class="sourceLineNo">094</span> }<a name="line.94"></a> |
| <span class="sourceLineNo">095</span><a name="line.95"></a> |
| <span class="sourceLineNo">096</span> ++opts;<a name="line.96"></a> |
| <span class="sourceLineNo">097</span> if (<a name="line.97"></a> |
| <span class="sourceLineNo">098</span> size >= comConf.getMinCompactSize() && !filesInRatio(potentialMatchFiles, currentRatio)<a name="line.98"></a> |
| <span class="sourceLineNo">099</span> ) {<a name="line.99"></a> |
| <span class="sourceLineNo">100</span> continue;<a name="line.100"></a> |
| <span class="sourceLineNo">101</span> }<a name="line.101"></a> |
| <span class="sourceLineNo">102</span><a name="line.102"></a> |
| <span class="sourceLineNo">103</span> ++optsInRatio;<a name="line.103"></a> |
| <span class="sourceLineNo">104</span> if (isBetterSelection(bestSelection, bestSize, potentialMatchFiles, size, mightBeStuck)) {<a name="line.104"></a> |
| <span class="sourceLineNo">105</span> bestSelection = potentialMatchFiles;<a name="line.105"></a> |
| <span class="sourceLineNo">106</span> bestSize = size;<a name="line.106"></a> |
| <span class="sourceLineNo">107</span> bestStart = start;<a name="line.107"></a> |
| <span class="sourceLineNo">108</span> }<a name="line.108"></a> |
| <span class="sourceLineNo">109</span> }<a name="line.109"></a> |
| <span class="sourceLineNo">110</span> }<a name="line.110"></a> |
| <span class="sourceLineNo">111</span> if (bestSelection.isEmpty() && mightBeStuck) {<a name="line.111"></a> |
| <span class="sourceLineNo">112</span> LOG.debug("Exploring compaction algorithm has selected " + smallest.size() + " files of size "<a name="line.112"></a> |
| <span class="sourceLineNo">113</span> + smallestSize + " because the store might be stuck");<a name="line.113"></a> |
| <span class="sourceLineNo">114</span> return new ArrayList<>(smallest);<a name="line.114"></a> |
| <span class="sourceLineNo">115</span> }<a name="line.115"></a> |
| <span class="sourceLineNo">116</span> LOG.debug(<a name="line.116"></a> |
| <span class="sourceLineNo">117</span> "Exploring compaction algorithm has selected {} files of size {} starting at "<a name="line.117"></a> |
| <span class="sourceLineNo">118</span> + "candidate #{} after considering {} permutations with {} in ratio",<a name="line.118"></a> |
| <span class="sourceLineNo">119</span> bestSelection.size(), bestSize, bestStart, opts, optsInRatio);<a name="line.119"></a> |
| <span class="sourceLineNo">120</span> return new ArrayList<>(bestSelection);<a name="line.120"></a> |
| <span class="sourceLineNo">121</span> }<a name="line.121"></a> |
| <span class="sourceLineNo">122</span><a name="line.122"></a> |
| <span class="sourceLineNo">123</span> /**<a name="line.123"></a> |
| <span class="sourceLineNo">124</span> * Select at least one file in the candidates list to compact, through choosing files from the<a name="line.124"></a> |
| <span class="sourceLineNo">125</span> * head to the index that the accumulation length larger the max compaction size. This method is a<a name="line.125"></a> |
| <span class="sourceLineNo">126</span> * supplementary of the selectSimpleCompaction() method, aims to make sure at least one file can<a name="line.126"></a> |
| <span class="sourceLineNo">127</span> * be selected to compact, for compactions like L0 files, which need to compact all files and as<a name="line.127"></a> |
| <span class="sourceLineNo">128</span> * soon as possible.<a name="line.128"></a> |
| <span class="sourceLineNo">129</span> */<a name="line.129"></a> |
| <span class="sourceLineNo">130</span> public List<HStoreFile> selectCompactFiles(final List<HStoreFile> candidates, int maxFiles,<a name="line.130"></a> |
| <span class="sourceLineNo">131</span> boolean isOffpeak) {<a name="line.131"></a> |
| <span class="sourceLineNo">132</span> long selectedSize = 0L;<a name="line.132"></a> |
| <span class="sourceLineNo">133</span> for (int end = 0; end < Math.min(candidates.size(), maxFiles); end++) {<a name="line.133"></a> |
| <span class="sourceLineNo">134</span> selectedSize += candidates.get(end).getReader().length();<a name="line.134"></a> |
| <span class="sourceLineNo">135</span> if (selectedSize >= comConf.getMaxCompactSize(isOffpeak)) {<a name="line.135"></a> |
| <span class="sourceLineNo">136</span> return candidates.subList(0, end + 1);<a name="line.136"></a> |
| <span class="sourceLineNo">137</span> }<a name="line.137"></a> |
| <span class="sourceLineNo">138</span> }<a name="line.138"></a> |
| <span class="sourceLineNo">139</span> return candidates;<a name="line.139"></a> |
| <span class="sourceLineNo">140</span> }<a name="line.140"></a> |
| <span class="sourceLineNo">141</span><a name="line.141"></a> |
| <span class="sourceLineNo">142</span> private boolean isBetterSelection(List<HStoreFile> bestSelection, long bestSize,<a name="line.142"></a> |
| <span class="sourceLineNo">143</span> List<HStoreFile> selection, long size, boolean mightBeStuck) {<a name="line.143"></a> |
| <span class="sourceLineNo">144</span> if (mightBeStuck && bestSize > 0 && size > 0) {<a name="line.144"></a> |
| <span class="sourceLineNo">145</span> // Keep the selection that removes most files for least size. That penaltizes adding<a name="line.145"></a> |
| <span class="sourceLineNo">146</span> // large files to compaction, but not small files, so we don't become totally inefficient<a name="line.146"></a> |
| <span class="sourceLineNo">147</span> // (might want to tweak that in future). Also, given the current order of looking at<a name="line.147"></a> |
| <span class="sourceLineNo">148</span> // permutations, prefer earlier files and smaller selection if the difference is small.<a name="line.148"></a> |
| <span class="sourceLineNo">149</span> final double REPLACE_IF_BETTER_BY = 1.05;<a name="line.149"></a> |
| <span class="sourceLineNo">150</span> double thresholdQuality = ((double) bestSelection.size() / bestSize) * REPLACE_IF_BETTER_BY;<a name="line.150"></a> |
| <span class="sourceLineNo">151</span> return thresholdQuality < ((double) selection.size() / size);<a name="line.151"></a> |
| <span class="sourceLineNo">152</span> }<a name="line.152"></a> |
| <span class="sourceLineNo">153</span> // Keep if this gets rid of more files. Or the same number of files for less io.<a name="line.153"></a> |
| <span class="sourceLineNo">154</span> return selection.size() > bestSelection.size()<a name="line.154"></a> |
| <span class="sourceLineNo">155</span> || (selection.size() == bestSelection.size() && size < bestSize);<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> /**<a name="line.158"></a> |
| <span class="sourceLineNo">159</span> * Find the total size of a list of store files.<a name="line.159"></a> |
| <span class="sourceLineNo">160</span> * @param potentialMatchFiles StoreFile list.<a name="line.160"></a> |
| <span class="sourceLineNo">161</span> * @return Sum of StoreFile.getReader().length();<a name="line.161"></a> |
| <span class="sourceLineNo">162</span> */<a name="line.162"></a> |
| <span class="sourceLineNo">163</span> private long getTotalStoreSize(List<HStoreFile> potentialMatchFiles) {<a name="line.163"></a> |
| <span class="sourceLineNo">164</span> return potentialMatchFiles.stream().mapToLong(sf -> sf.getReader().length()).sum();<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> /**<a name="line.167"></a> |
| <span class="sourceLineNo">168</span> * Check that all files satisfy the constraint<a name="line.168"></a> |
| <span class="sourceLineNo">169</span> *<a name="line.169"></a> |
| <span class="sourceLineNo">170</span> * <pre><a name="line.170"></a> |
| <span class="sourceLineNo">171</span> * FileSize(i) <= ( Sum(0,N,FileSize(_)) - FileSize(i)) * Ratio.<a name="line.171"></a> |
| <span class="sourceLineNo">172</span> * </pre><a name="line.172"></a> |
| <span class="sourceLineNo">173</span> *<a name="line.173"></a> |
| <span class="sourceLineNo">174</span> * @param files List of store files to consider as a compaction candidate.<a name="line.174"></a> |
| <span class="sourceLineNo">175</span> * @param currentRatio The ratio to use.<a name="line.175"></a> |
| <span class="sourceLineNo">176</span> * @return a boolean if these files satisfy the ratio constraints.<a name="line.176"></a> |
| <span class="sourceLineNo">177</span> */<a name="line.177"></a> |
| <span class="sourceLineNo">178</span> private boolean filesInRatio(List<HStoreFile> files, double currentRatio) {<a name="line.178"></a> |
| <span class="sourceLineNo">179</span> if (files.size() < 2) {<a name="line.179"></a> |
| <span class="sourceLineNo">180</span> return true;<a name="line.180"></a> |
| <span class="sourceLineNo">181</span> }<a name="line.181"></a> |
| <span class="sourceLineNo">182</span><a name="line.182"></a> |
| <span class="sourceLineNo">183</span> long totalFileSize = getTotalStoreSize(files);<a name="line.183"></a> |
| <span class="sourceLineNo">184</span><a name="line.184"></a> |
| <span class="sourceLineNo">185</span> for (HStoreFile file : files) {<a name="line.185"></a> |
| <span class="sourceLineNo">186</span> long singleFileSize = file.getReader().length();<a name="line.186"></a> |
| <span class="sourceLineNo">187</span> long sumAllOtherFileSizes = totalFileSize - singleFileSize;<a name="line.187"></a> |
| <span class="sourceLineNo">188</span><a name="line.188"></a> |
| <span class="sourceLineNo">189</span> if (singleFileSize > sumAllOtherFileSizes * currentRatio) {<a name="line.189"></a> |
| <span class="sourceLineNo">190</span> return false;<a name="line.190"></a> |
| <span class="sourceLineNo">191</span> }<a name="line.191"></a> |
| <span class="sourceLineNo">192</span> }<a name="line.192"></a> |
| <span class="sourceLineNo">193</span> return true;<a name="line.193"></a> |
| <span class="sourceLineNo">194</span> }<a name="line.194"></a> |
| <span class="sourceLineNo">195</span>}<a name="line.195"></a> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| </pre> |
| </div> |
| </body> |
| </html> |