| <!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 or more<a name="line.2"></a> |
| <span class="sourceLineNo">003</span> * contributor license agreements. See the NOTICE file distributed with this<a name="line.3"></a> |
| <span class="sourceLineNo">004</span> * work for additional information regarding copyright ownership. The ASF<a name="line.4"></a> |
| <span class="sourceLineNo">005</span> * licenses this file to you under the Apache License, Version 2.0 (the<a name="line.5"></a> |
| <span class="sourceLineNo">006</span> * "License"); you may not use this file except in compliance with the License.<a name="line.6"></a> |
| <span class="sourceLineNo">007</span> * You may obtain a copy of the License at<a name="line.7"></a> |
| <span class="sourceLineNo">008</span> *<a name="line.8"></a> |
| <span class="sourceLineNo">009</span> * http://www.apache.org/licenses/LICENSE-2.0<a name="line.9"></a> |
| <span class="sourceLineNo">010</span> *<a name="line.10"></a> |
| <span class="sourceLineNo">011</span> * Unless required by applicable law or agreed to in writing, software<a name="line.11"></a> |
| <span class="sourceLineNo">012</span> * distributed under the License is distributed on an "AS IS" BASIS, WITHOUT<a name="line.12"></a> |
| <span class="sourceLineNo">013</span> * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the<a name="line.13"></a> |
| <span class="sourceLineNo">014</span> * License for the specific language governing permissions and limitations<a name="line.14"></a> |
| <span class="sourceLineNo">015</span> * under the License.<a name="line.15"></a> |
| <span class="sourceLineNo">016</span> */<a name="line.16"></a> |
| <span class="sourceLineNo">017</span>package org.apache.hadoop.hbase.regionserver;<a name="line.17"></a> |
| <span class="sourceLineNo">018</span><a name="line.18"></a> |
| <span class="sourceLineNo">019</span>import java.io.ByteArrayInputStream;<a name="line.19"></a> |
| <span class="sourceLineNo">020</span>import java.io.ByteArrayOutputStream;<a name="line.20"></a> |
| <span class="sourceLineNo">021</span>import java.io.DataOutputStream;<a name="line.21"></a> |
| <span class="sourceLineNo">022</span>import java.io.IOException;<a name="line.22"></a> |
| <span class="sourceLineNo">023</span>import java.io.InputStream;<a name="line.23"></a> |
| <span class="sourceLineNo">024</span>import java.text.DecimalFormat;<a name="line.24"></a> |
| <span class="sourceLineNo">025</span>import java.util.ArrayList;<a name="line.25"></a> |
| <span class="sourceLineNo">026</span>import java.util.Iterator;<a name="line.26"></a> |
| <span class="sourceLineNo">027</span>import java.util.List;<a name="line.27"></a> |
| <span class="sourceLineNo">028</span>import java.util.Locale;<a name="line.28"></a> |
| <span class="sourceLineNo">029</span><a name="line.29"></a> |
| <span class="sourceLineNo">030</span>import org.apache.hadoop.conf.Configuration;<a name="line.30"></a> |
| <span class="sourceLineNo">031</span>import org.apache.hadoop.fs.FileSystem;<a name="line.31"></a> |
| <span class="sourceLineNo">032</span>import org.apache.hadoop.fs.Path;<a name="line.32"></a> |
| <span class="sourceLineNo">033</span>import org.apache.hadoop.hbase.Cell;<a name="line.33"></a> |
| <span class="sourceLineNo">034</span>import org.apache.hadoop.hbase.HBaseConfiguration;<a name="line.34"></a> |
| <span class="sourceLineNo">035</span>import org.apache.hadoop.hbase.KeyValue;<a name="line.35"></a> |
| <span class="sourceLineNo">036</span>import org.apache.hadoop.hbase.KeyValueUtil;<a name="line.36"></a> |
| <span class="sourceLineNo">037</span>import org.apache.hadoop.hbase.io.compress.Compression;<a name="line.37"></a> |
| <span class="sourceLineNo">038</span>import org.apache.hadoop.hbase.io.compress.Compression.Algorithm;<a name="line.38"></a> |
| <span class="sourceLineNo">039</span>import org.apache.hadoop.hbase.io.encoding.DataBlockEncoder;<a name="line.39"></a> |
| <span class="sourceLineNo">040</span>import org.apache.hadoop.hbase.io.encoding.DataBlockEncoding;<a name="line.40"></a> |
| <span class="sourceLineNo">041</span>import org.apache.hadoop.hbase.io.encoding.EncodedDataBlock;<a name="line.41"></a> |
| <span class="sourceLineNo">042</span>import org.apache.hadoop.hbase.io.hfile.CacheConfig;<a name="line.42"></a> |
| <span class="sourceLineNo">043</span>import org.apache.hadoop.hbase.io.hfile.HFileBlock;<a name="line.43"></a> |
| <span class="sourceLineNo">044</span>import org.apache.hadoop.hbase.io.hfile.HFileContext;<a name="line.44"></a> |
| <span class="sourceLineNo">045</span>import org.apache.hadoop.hbase.io.hfile.HFileContextBuilder;<a name="line.45"></a> |
| <span class="sourceLineNo">046</span>import org.apache.hadoop.hbase.io.hfile.HFileReaderImpl;<a name="line.46"></a> |
| <span class="sourceLineNo">047</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.47"></a> |
| <span class="sourceLineNo">048</span>import org.apache.hadoop.io.WritableUtils;<a name="line.48"></a> |
| <span class="sourceLineNo">049</span>import org.apache.hadoop.io.compress.CompressionOutputStream;<a name="line.49"></a> |
| <span class="sourceLineNo">050</span>import org.apache.hadoop.io.compress.Compressor;<a name="line.50"></a> |
| <span class="sourceLineNo">051</span>import org.apache.hadoop.io.compress.Decompressor;<a name="line.51"></a> |
| <span class="sourceLineNo">052</span>import org.slf4j.Logger;<a name="line.52"></a> |
| <span class="sourceLineNo">053</span>import org.slf4j.LoggerFactory;<a name="line.53"></a> |
| <span class="sourceLineNo">054</span>import org.apache.hbase.thirdparty.org.apache.commons.cli.CommandLine;<a name="line.54"></a> |
| <span class="sourceLineNo">055</span>import org.apache.hbase.thirdparty.org.apache.commons.cli.CommandLineParser;<a name="line.55"></a> |
| <span class="sourceLineNo">056</span>import org.apache.hbase.thirdparty.org.apache.commons.cli.Option;<a name="line.56"></a> |
| <span class="sourceLineNo">057</span>import org.apache.hbase.thirdparty.org.apache.commons.cli.Options;<a name="line.57"></a> |
| <span class="sourceLineNo">058</span>import org.apache.hbase.thirdparty.org.apache.commons.cli.ParseException;<a name="line.58"></a> |
| <span class="sourceLineNo">059</span>import org.apache.hbase.thirdparty.org.apache.commons.cli.PosixParser;<a name="line.59"></a> |
| <span class="sourceLineNo">060</span><a name="line.60"></a> |
| <span class="sourceLineNo">061</span>/**<a name="line.61"></a> |
| <span class="sourceLineNo">062</span> * Tests various algorithms for key compression on an existing HFile. Useful<a name="line.62"></a> |
| <span class="sourceLineNo">063</span> * for testing, debugging and benchmarking.<a name="line.63"></a> |
| <span class="sourceLineNo">064</span> */<a name="line.64"></a> |
| <span class="sourceLineNo">065</span>public class DataBlockEncodingTool {<a name="line.65"></a> |
| <span class="sourceLineNo">066</span> private static final Logger LOG = LoggerFactory.getLogger(<a name="line.66"></a> |
| <span class="sourceLineNo">067</span> DataBlockEncodingTool.class);<a name="line.67"></a> |
| <span class="sourceLineNo">068</span><a name="line.68"></a> |
| <span class="sourceLineNo">069</span> private static final boolean includesMemstoreTS = true;<a name="line.69"></a> |
| <span class="sourceLineNo">070</span><a name="line.70"></a> |
| <span class="sourceLineNo">071</span> /**<a name="line.71"></a> |
| <span class="sourceLineNo">072</span> * How many times to run the benchmark. More times means better data in terms<a name="line.72"></a> |
| <span class="sourceLineNo">073</span> * of statistics but slower execution. Has to be strictly larger than<a name="line.73"></a> |
| <span class="sourceLineNo">074</span> * {@link #DEFAULT_BENCHMARK_N_OMIT}.<a name="line.74"></a> |
| <span class="sourceLineNo">075</span> */<a name="line.75"></a> |
| <span class="sourceLineNo">076</span> private static final int DEFAULT_BENCHMARK_N_TIMES = 12;<a name="line.76"></a> |
| <span class="sourceLineNo">077</span><a name="line.77"></a> |
| <span class="sourceLineNo">078</span> /**<a name="line.78"></a> |
| <span class="sourceLineNo">079</span> * How many first runs should not be included in the benchmark. Done in order<a name="line.79"></a> |
| <span class="sourceLineNo">080</span> * to exclude setup cost.<a name="line.80"></a> |
| <span class="sourceLineNo">081</span> */<a name="line.81"></a> |
| <span class="sourceLineNo">082</span> private static final int DEFAULT_BENCHMARK_N_OMIT = 2;<a name="line.82"></a> |
| <span class="sourceLineNo">083</span><a name="line.83"></a> |
| <span class="sourceLineNo">084</span> /** HFile name to be used in benchmark */<a name="line.84"></a> |
| <span class="sourceLineNo">085</span> private static final String OPT_HFILE_NAME = "f";<a name="line.85"></a> |
| <span class="sourceLineNo">086</span><a name="line.86"></a> |
| <span class="sourceLineNo">087</span> /** Maximum number of key/value pairs to process in a single benchmark run */<a name="line.87"></a> |
| <span class="sourceLineNo">088</span> private static final String OPT_KV_LIMIT = "n";<a name="line.88"></a> |
| <span class="sourceLineNo">089</span><a name="line.89"></a> |
| <span class="sourceLineNo">090</span> /** Whether to run a benchmark to measure read throughput */<a name="line.90"></a> |
| <span class="sourceLineNo">091</span> private static final String OPT_MEASURE_THROUGHPUT = "b";<a name="line.91"></a> |
| <span class="sourceLineNo">092</span><a name="line.92"></a> |
| <span class="sourceLineNo">093</span> /** If this is specified, no correctness testing will be done */<a name="line.93"></a> |
| <span class="sourceLineNo">094</span> private static final String OPT_OMIT_CORRECTNESS_TEST = "c";<a name="line.94"></a> |
| <span class="sourceLineNo">095</span><a name="line.95"></a> |
| <span class="sourceLineNo">096</span> /** What compression algorithm to test */<a name="line.96"></a> |
| <span class="sourceLineNo">097</span> private static final String OPT_COMPRESSION_ALGORITHM = "a";<a name="line.97"></a> |
| <span class="sourceLineNo">098</span><a name="line.98"></a> |
| <span class="sourceLineNo">099</span> /** Number of times to run each benchmark */<a name="line.99"></a> |
| <span class="sourceLineNo">100</span> private static final String OPT_BENCHMARK_N_TIMES = "t";<a name="line.100"></a> |
| <span class="sourceLineNo">101</span><a name="line.101"></a> |
| <span class="sourceLineNo">102</span> /** Number of first runs of every benchmark to omit from statistics */<a name="line.102"></a> |
| <span class="sourceLineNo">103</span> private static final String OPT_BENCHMARK_N_OMIT = "omit";<a name="line.103"></a> |
| <span class="sourceLineNo">104</span><a name="line.104"></a> |
| <span class="sourceLineNo">105</span> /** Compression algorithm to use if not specified on the command line */<a name="line.105"></a> |
| <span class="sourceLineNo">106</span> private static final Algorithm DEFAULT_COMPRESSION =<a name="line.106"></a> |
| <span class="sourceLineNo">107</span> Compression.Algorithm.GZ;<a name="line.107"></a> |
| <span class="sourceLineNo">108</span><a name="line.108"></a> |
| <span class="sourceLineNo">109</span> private static final DecimalFormat DELIMITED_DECIMAL_FORMAT =<a name="line.109"></a> |
| <span class="sourceLineNo">110</span> new DecimalFormat();<a name="line.110"></a> |
| <span class="sourceLineNo">111</span><a name="line.111"></a> |
| <span class="sourceLineNo">112</span> static {<a name="line.112"></a> |
| <span class="sourceLineNo">113</span> DELIMITED_DECIMAL_FORMAT.setGroupingSize(3);<a name="line.113"></a> |
| <span class="sourceLineNo">114</span> }<a name="line.114"></a> |
| <span class="sourceLineNo">115</span><a name="line.115"></a> |
| <span class="sourceLineNo">116</span> private static final String PCT_FORMAT = "%.2f %%";<a name="line.116"></a> |
| <span class="sourceLineNo">117</span> private static final String INT_FORMAT = "%d";<a name="line.117"></a> |
| <span class="sourceLineNo">118</span><a name="line.118"></a> |
| <span class="sourceLineNo">119</span> private static int benchmarkNTimes = DEFAULT_BENCHMARK_N_TIMES;<a name="line.119"></a> |
| <span class="sourceLineNo">120</span> private static int benchmarkNOmit = DEFAULT_BENCHMARK_N_OMIT;<a name="line.120"></a> |
| <span class="sourceLineNo">121</span><a name="line.121"></a> |
| <span class="sourceLineNo">122</span> private List<EncodedDataBlock> codecs = new ArrayList<>();<a name="line.122"></a> |
| <span class="sourceLineNo">123</span> private long totalPrefixLength = 0;<a name="line.123"></a> |
| <span class="sourceLineNo">124</span> private long totalKeyLength = 0;<a name="line.124"></a> |
| <span class="sourceLineNo">125</span> private long totalValueLength = 0;<a name="line.125"></a> |
| <span class="sourceLineNo">126</span> private long totalKeyRedundancyLength = 0;<a name="line.126"></a> |
| <span class="sourceLineNo">127</span> private long totalCFLength = 0;<a name="line.127"></a> |
| <span class="sourceLineNo">128</span><a name="line.128"></a> |
| <span class="sourceLineNo">129</span> private byte[] rawKVs;<a name="line.129"></a> |
| <span class="sourceLineNo">130</span> private boolean useHBaseChecksum = false;<a name="line.130"></a> |
| <span class="sourceLineNo">131</span><a name="line.131"></a> |
| <span class="sourceLineNo">132</span> private final String compressionAlgorithmName;<a name="line.132"></a> |
| <span class="sourceLineNo">133</span> private final Algorithm compressionAlgorithm;<a name="line.133"></a> |
| <span class="sourceLineNo">134</span> private final Compressor compressor;<a name="line.134"></a> |
| <span class="sourceLineNo">135</span> private final Decompressor decompressor;<a name="line.135"></a> |
| <span class="sourceLineNo">136</span><a name="line.136"></a> |
| <span class="sourceLineNo">137</span> // Check if HFile use Tag.<a name="line.137"></a> |
| <span class="sourceLineNo">138</span> private static boolean USE_TAG = false;<a name="line.138"></a> |
| <span class="sourceLineNo">139</span><a name="line.139"></a> |
| <span class="sourceLineNo">140</span> private enum Manipulation {<a name="line.140"></a> |
| <span class="sourceLineNo">141</span> ENCODING,<a name="line.141"></a> |
| <span class="sourceLineNo">142</span> DECODING,<a name="line.142"></a> |
| <span class="sourceLineNo">143</span> COMPRESSION,<a name="line.143"></a> |
| <span class="sourceLineNo">144</span> DECOMPRESSION;<a name="line.144"></a> |
| <span class="sourceLineNo">145</span><a name="line.145"></a> |
| <span class="sourceLineNo">146</span> @Override<a name="line.146"></a> |
| <span class="sourceLineNo">147</span> public String toString() {<a name="line.147"></a> |
| <span class="sourceLineNo">148</span> String s = super.toString();<a name="line.148"></a> |
| <span class="sourceLineNo">149</span> StringBuilder sb = new StringBuilder();<a name="line.149"></a> |
| <span class="sourceLineNo">150</span> sb.append(s.charAt(0));<a name="line.150"></a> |
| <span class="sourceLineNo">151</span> sb.append(s.substring(1).toLowerCase(Locale.ROOT));<a name="line.151"></a> |
| <span class="sourceLineNo">152</span> return sb.toString();<a name="line.152"></a> |
| <span class="sourceLineNo">153</span> }<a name="line.153"></a> |
| <span class="sourceLineNo">154</span> }<a name="line.154"></a> |
| <span class="sourceLineNo">155</span><a name="line.155"></a> |
| <span class="sourceLineNo">156</span> /**<a name="line.156"></a> |
| <span class="sourceLineNo">157</span> * @param compressionAlgorithmName What kind of algorithm should be used<a name="line.157"></a> |
| <span class="sourceLineNo">158</span> * as baseline for comparison (e.g. lzo, gz).<a name="line.158"></a> |
| <span class="sourceLineNo">159</span> */<a name="line.159"></a> |
| <span class="sourceLineNo">160</span> public DataBlockEncodingTool(String compressionAlgorithmName) {<a name="line.160"></a> |
| <span class="sourceLineNo">161</span> this.compressionAlgorithmName = compressionAlgorithmName;<a name="line.161"></a> |
| <span class="sourceLineNo">162</span> this.compressionAlgorithm = Compression.getCompressionAlgorithmByName(<a name="line.162"></a> |
| <span class="sourceLineNo">163</span> compressionAlgorithmName);<a name="line.163"></a> |
| <span class="sourceLineNo">164</span> this.compressor = this.compressionAlgorithm.getCompressor();<a name="line.164"></a> |
| <span class="sourceLineNo">165</span> this.decompressor = this.compressionAlgorithm.getDecompressor();<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> /**<a name="line.168"></a> |
| <span class="sourceLineNo">169</span> * Check statistics for given HFile for different data block encoders.<a name="line.169"></a> |
| <span class="sourceLineNo">170</span> * @param scanner Of file which will be compressed.<a name="line.170"></a> |
| <span class="sourceLineNo">171</span> * @param kvLimit Maximal count of KeyValue which will be processed.<a name="line.171"></a> |
| <span class="sourceLineNo">172</span> * @throws IOException thrown if scanner is invalid<a name="line.172"></a> |
| <span class="sourceLineNo">173</span> */<a name="line.173"></a> |
| <span class="sourceLineNo">174</span> public void checkStatistics(final KeyValueScanner scanner, final int kvLimit)<a name="line.174"></a> |
| <span class="sourceLineNo">175</span> throws IOException {<a name="line.175"></a> |
| <span class="sourceLineNo">176</span> scanner.seek(KeyValue.LOWESTKEY);<a name="line.176"></a> |
| <span class="sourceLineNo">177</span><a name="line.177"></a> |
| <span class="sourceLineNo">178</span> KeyValue currentKV;<a name="line.178"></a> |
| <span class="sourceLineNo">179</span><a name="line.179"></a> |
| <span class="sourceLineNo">180</span> byte[] previousKey = null;<a name="line.180"></a> |
| <span class="sourceLineNo">181</span> byte[] currentKey;<a name="line.181"></a> |
| <span class="sourceLineNo">182</span><a name="line.182"></a> |
| <span class="sourceLineNo">183</span> DataBlockEncoding[] encodings = DataBlockEncoding.values();<a name="line.183"></a> |
| <span class="sourceLineNo">184</span><a name="line.184"></a> |
| <span class="sourceLineNo">185</span> ByteArrayOutputStream uncompressedOutputStream =<a name="line.185"></a> |
| <span class="sourceLineNo">186</span> new ByteArrayOutputStream();<a name="line.186"></a> |
| <span class="sourceLineNo">187</span><a name="line.187"></a> |
| <span class="sourceLineNo">188</span> int j = 0;<a name="line.188"></a> |
| <span class="sourceLineNo">189</span> while ((currentKV = KeyValueUtil.ensureKeyValue(scanner.next())) != null && j < kvLimit) {<a name="line.189"></a> |
| <span class="sourceLineNo">190</span> // Iterates through key/value pairs<a name="line.190"></a> |
| <span class="sourceLineNo">191</span> j++;<a name="line.191"></a> |
| <span class="sourceLineNo">192</span> currentKey = currentKV.getKey();<a name="line.192"></a> |
| <span class="sourceLineNo">193</span> if (previousKey != null) {<a name="line.193"></a> |
| <span class="sourceLineNo">194</span> for (int i = 0; i < previousKey.length && i < currentKey.length &&<a name="line.194"></a> |
| <span class="sourceLineNo">195</span> previousKey[i] == currentKey[i]; ++i) {<a name="line.195"></a> |
| <span class="sourceLineNo">196</span> totalKeyRedundancyLength++;<a name="line.196"></a> |
| <span class="sourceLineNo">197</span> }<a name="line.197"></a> |
| <span class="sourceLineNo">198</span> }<a name="line.198"></a> |
| <span class="sourceLineNo">199</span><a name="line.199"></a> |
| <span class="sourceLineNo">200</span> // Add tagsLen zero to cells don't include tags. Since the process of<a name="line.200"></a> |
| <span class="sourceLineNo">201</span> // scanner converts byte array to KV would abandon tagsLen part if tagsLen<a name="line.201"></a> |
| <span class="sourceLineNo">202</span> // is zero. But we still needs the tagsLen part to check if current cell<a name="line.202"></a> |
| <span class="sourceLineNo">203</span> // include tags. If USE_TAG is true, HFile contains cells with tags,<a name="line.203"></a> |
| <span class="sourceLineNo">204</span> // if the cell tagsLen equals 0, it means other cells may have tags.<a name="line.204"></a> |
| <span class="sourceLineNo">205</span> if (USE_TAG && currentKV.getTagsLength() == 0) {<a name="line.205"></a> |
| <span class="sourceLineNo">206</span> uncompressedOutputStream.write(currentKV.getBuffer(),<a name="line.206"></a> |
| <span class="sourceLineNo">207</span> currentKV.getOffset(), currentKV.getLength());<a name="line.207"></a> |
| <span class="sourceLineNo">208</span> // write tagsLen = 0.<a name="line.208"></a> |
| <span class="sourceLineNo">209</span> uncompressedOutputStream.write(Bytes.toBytes((short) 0));<a name="line.209"></a> |
| <span class="sourceLineNo">210</span> } else {<a name="line.210"></a> |
| <span class="sourceLineNo">211</span> uncompressedOutputStream.write(currentKV.getBuffer(),<a name="line.211"></a> |
| <span class="sourceLineNo">212</span> currentKV.getOffset(), currentKV.getLength());<a name="line.212"></a> |
| <span class="sourceLineNo">213</span> }<a name="line.213"></a> |
| <span class="sourceLineNo">214</span><a name="line.214"></a> |
| <span class="sourceLineNo">215</span> if(includesMemstoreTS) {<a name="line.215"></a> |
| <span class="sourceLineNo">216</span> WritableUtils.writeVLong(<a name="line.216"></a> |
| <span class="sourceLineNo">217</span> new DataOutputStream(uncompressedOutputStream), currentKV.getSequenceId());<a name="line.217"></a> |
| <span class="sourceLineNo">218</span> }<a name="line.218"></a> |
| <span class="sourceLineNo">219</span><a name="line.219"></a> |
| <span class="sourceLineNo">220</span> previousKey = currentKey;<a name="line.220"></a> |
| <span class="sourceLineNo">221</span><a name="line.221"></a> |
| <span class="sourceLineNo">222</span> int kLen = currentKV.getKeyLength();<a name="line.222"></a> |
| <span class="sourceLineNo">223</span> int vLen = currentKV.getValueLength();<a name="line.223"></a> |
| <span class="sourceLineNo">224</span> int cfLen = currentKV.getFamilyLength(currentKV.getFamilyOffset());<a name="line.224"></a> |
| <span class="sourceLineNo">225</span> int restLen = currentKV.getLength() - kLen - vLen;<a name="line.225"></a> |
| <span class="sourceLineNo">226</span><a name="line.226"></a> |
| <span class="sourceLineNo">227</span> totalKeyLength += kLen;<a name="line.227"></a> |
| <span class="sourceLineNo">228</span> totalValueLength += vLen;<a name="line.228"></a> |
| <span class="sourceLineNo">229</span> totalPrefixLength += restLen;<a name="line.229"></a> |
| <span class="sourceLineNo">230</span> totalCFLength += cfLen;<a name="line.230"></a> |
| <span class="sourceLineNo">231</span> }<a name="line.231"></a> |
| <span class="sourceLineNo">232</span><a name="line.232"></a> |
| <span class="sourceLineNo">233</span> rawKVs = uncompressedOutputStream.toByteArray();<a name="line.233"></a> |
| <span class="sourceLineNo">234</span> for (DataBlockEncoding encoding : encodings) {<a name="line.234"></a> |
| <span class="sourceLineNo">235</span> if (encoding == DataBlockEncoding.NONE) {<a name="line.235"></a> |
| <span class="sourceLineNo">236</span> continue;<a name="line.236"></a> |
| <span class="sourceLineNo">237</span> }<a name="line.237"></a> |
| <span class="sourceLineNo">238</span> DataBlockEncoder d = encoding.getEncoder();<a name="line.238"></a> |
| <span class="sourceLineNo">239</span> HFileContext meta = new HFileContextBuilder()<a name="line.239"></a> |
| <span class="sourceLineNo">240</span> .withDataBlockEncoding(encoding)<a name="line.240"></a> |
| <span class="sourceLineNo">241</span> .withCompression(Compression.Algorithm.NONE)<a name="line.241"></a> |
| <span class="sourceLineNo">242</span> .withIncludesMvcc(includesMemstoreTS)<a name="line.242"></a> |
| <span class="sourceLineNo">243</span> .withIncludesTags(USE_TAG).build();<a name="line.243"></a> |
| <span class="sourceLineNo">244</span> codecs.add(new EncodedDataBlock(d, encoding, rawKVs, meta ));<a name="line.244"></a> |
| <span class="sourceLineNo">245</span> }<a name="line.245"></a> |
| <span class="sourceLineNo">246</span> }<a name="line.246"></a> |
| <span class="sourceLineNo">247</span><a name="line.247"></a> |
| <span class="sourceLineNo">248</span> /**<a name="line.248"></a> |
| <span class="sourceLineNo">249</span> * Verify if all data block encoders are working properly.<a name="line.249"></a> |
| <span class="sourceLineNo">250</span> *<a name="line.250"></a> |
| <span class="sourceLineNo">251</span> * @param scanner Of file which was compressed.<a name="line.251"></a> |
| <span class="sourceLineNo">252</span> * @param kvLimit Maximal count of KeyValue which will be processed.<a name="line.252"></a> |
| <span class="sourceLineNo">253</span> * @return true if all data block encoders compressed/decompressed correctly.<a name="line.253"></a> |
| <span class="sourceLineNo">254</span> * @throws IOException thrown if scanner is invalid<a name="line.254"></a> |
| <span class="sourceLineNo">255</span> */<a name="line.255"></a> |
| <span class="sourceLineNo">256</span> public boolean verifyCodecs(final KeyValueScanner scanner, final int kvLimit)<a name="line.256"></a> |
| <span class="sourceLineNo">257</span> throws IOException {<a name="line.257"></a> |
| <span class="sourceLineNo">258</span> KeyValue currentKv;<a name="line.258"></a> |
| <span class="sourceLineNo">259</span><a name="line.259"></a> |
| <span class="sourceLineNo">260</span> scanner.seek(KeyValue.LOWESTKEY);<a name="line.260"></a> |
| <span class="sourceLineNo">261</span> List<Iterator<Cell>> codecIterators = new ArrayList<>();<a name="line.261"></a> |
| <span class="sourceLineNo">262</span> for(EncodedDataBlock codec : codecs) {<a name="line.262"></a> |
| <span class="sourceLineNo">263</span> codecIterators.add(codec.getIterator(HFileBlock.headerSize(useHBaseChecksum)));<a name="line.263"></a> |
| <span class="sourceLineNo">264</span> }<a name="line.264"></a> |
| <span class="sourceLineNo">265</span><a name="line.265"></a> |
| <span class="sourceLineNo">266</span> int j = 0;<a name="line.266"></a> |
| <span class="sourceLineNo">267</span> while ((currentKv = KeyValueUtil.ensureKeyValue(scanner.next())) != null && j < kvLimit) {<a name="line.267"></a> |
| <span class="sourceLineNo">268</span> // Iterates through key/value pairs<a name="line.268"></a> |
| <span class="sourceLineNo">269</span> ++j;<a name="line.269"></a> |
| <span class="sourceLineNo">270</span> for (Iterator<Cell> it : codecIterators) {<a name="line.270"></a> |
| <span class="sourceLineNo">271</span> Cell c = it.next();<a name="line.271"></a> |
| <span class="sourceLineNo">272</span> KeyValue codecKv = KeyValueUtil.ensureKeyValue(c);<a name="line.272"></a> |
| <span class="sourceLineNo">273</span> if (codecKv == null || 0 != Bytes.compareTo(<a name="line.273"></a> |
| <span class="sourceLineNo">274</span> codecKv.getBuffer(), codecKv.getOffset(), codecKv.getLength(),<a name="line.274"></a> |
| <span class="sourceLineNo">275</span> currentKv.getBuffer(), currentKv.getOffset(),<a name="line.275"></a> |
| <span class="sourceLineNo">276</span> currentKv.getLength())) {<a name="line.276"></a> |
| <span class="sourceLineNo">277</span> if (codecKv == null) {<a name="line.277"></a> |
| <span class="sourceLineNo">278</span> LOG.error("There is a bug in codec " + it +<a name="line.278"></a> |
| <span class="sourceLineNo">279</span> " it returned null KeyValue,");<a name="line.279"></a> |
| <span class="sourceLineNo">280</span> } else {<a name="line.280"></a> |
| <span class="sourceLineNo">281</span> int prefix = 0;<a name="line.281"></a> |
| <span class="sourceLineNo">282</span> int limitLength = 2 * Bytes.SIZEOF_INT +<a name="line.282"></a> |
| <span class="sourceLineNo">283</span> Math.min(codecKv.getLength(), currentKv.getLength());<a name="line.283"></a> |
| <span class="sourceLineNo">284</span> while (prefix < limitLength &&<a name="line.284"></a> |
| <span class="sourceLineNo">285</span> codecKv.getBuffer()[prefix + codecKv.getOffset()] ==<a name="line.285"></a> |
| <span class="sourceLineNo">286</span> currentKv.getBuffer()[prefix + currentKv.getOffset()]) {<a name="line.286"></a> |
| <span class="sourceLineNo">287</span> prefix++;<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> LOG.error("There is bug in codec " + it.toString() +<a name="line.290"></a> |
| <span class="sourceLineNo">291</span> "\n on element " + j +<a name="line.291"></a> |
| <span class="sourceLineNo">292</span> "\n codecKv.getKeyLength() " + codecKv.getKeyLength() +<a name="line.292"></a> |
| <span class="sourceLineNo">293</span> "\n codecKv.getValueLength() " + codecKv.getValueLength() +<a name="line.293"></a> |
| <span class="sourceLineNo">294</span> "\n codecKv.getLength() " + codecKv.getLength() +<a name="line.294"></a> |
| <span class="sourceLineNo">295</span> "\n currentKv.getKeyLength() " + currentKv.getKeyLength() +<a name="line.295"></a> |
| <span class="sourceLineNo">296</span> "\n currentKv.getValueLength() " + currentKv.getValueLength() +<a name="line.296"></a> |
| <span class="sourceLineNo">297</span> "\n codecKv.getLength() " + currentKv.getLength() +<a name="line.297"></a> |
| <span class="sourceLineNo">298</span> "\n currentKV rowLength " + currentKv.getRowLength() +<a name="line.298"></a> |
| <span class="sourceLineNo">299</span> " familyName " + currentKv.getFamilyLength() +<a name="line.299"></a> |
| <span class="sourceLineNo">300</span> " qualifier " + currentKv.getQualifierLength() +<a name="line.300"></a> |
| <span class="sourceLineNo">301</span> "\n prefix " + prefix +<a name="line.301"></a> |
| <span class="sourceLineNo">302</span> "\n codecKv '" + Bytes.toStringBinary(codecKv.getBuffer(),<a name="line.302"></a> |
| <span class="sourceLineNo">303</span> codecKv.getOffset(), prefix) + "' diff '" +<a name="line.303"></a> |
| <span class="sourceLineNo">304</span> Bytes.toStringBinary(codecKv.getBuffer(),<a name="line.304"></a> |
| <span class="sourceLineNo">305</span> codecKv.getOffset() + prefix, codecKv.getLength() -<a name="line.305"></a> |
| <span class="sourceLineNo">306</span> prefix) + "'" +<a name="line.306"></a> |
| <span class="sourceLineNo">307</span> "\n currentKv '" + Bytes.toStringBinary(<a name="line.307"></a> |
| <span class="sourceLineNo">308</span> currentKv.getBuffer(),<a name="line.308"></a> |
| <span class="sourceLineNo">309</span> currentKv.getOffset(), prefix) + "' diff '" +<a name="line.309"></a> |
| <span class="sourceLineNo">310</span> Bytes.toStringBinary(currentKv.getBuffer(),<a name="line.310"></a> |
| <span class="sourceLineNo">311</span> currentKv.getOffset() + prefix, currentKv.getLength() -<a name="line.311"></a> |
| <span class="sourceLineNo">312</span> prefix) + "'"<a name="line.312"></a> |
| <span class="sourceLineNo">313</span> );<a name="line.313"></a> |
| <span class="sourceLineNo">314</span> }<a name="line.314"></a> |
| <span class="sourceLineNo">315</span> return false;<a name="line.315"></a> |
| <span class="sourceLineNo">316</span> }<a name="line.316"></a> |
| <span class="sourceLineNo">317</span> }<a name="line.317"></a> |
| <span class="sourceLineNo">318</span> }<a name="line.318"></a> |
| <span class="sourceLineNo">319</span><a name="line.319"></a> |
| <span class="sourceLineNo">320</span> LOG.info("Verification was successful!");<a name="line.320"></a> |
| <span class="sourceLineNo">321</span><a name="line.321"></a> |
| <span class="sourceLineNo">322</span> return true;<a name="line.322"></a> |
| <span class="sourceLineNo">323</span> }<a name="line.323"></a> |
| <span class="sourceLineNo">324</span><a name="line.324"></a> |
| <span class="sourceLineNo">325</span> /**<a name="line.325"></a> |
| <span class="sourceLineNo">326</span> * Benchmark codec's speed.<a name="line.326"></a> |
| <span class="sourceLineNo">327</span> */<a name="line.327"></a> |
| <span class="sourceLineNo">328</span> public void benchmarkCodecs() throws IOException {<a name="line.328"></a> |
| <span class="sourceLineNo">329</span> LOG.info("Starting a throughput benchmark for data block encoding codecs");<a name="line.329"></a> |
| <span class="sourceLineNo">330</span> int prevTotalSize = -1;<a name="line.330"></a> |
| <span class="sourceLineNo">331</span> for (EncodedDataBlock codec : codecs) {<a name="line.331"></a> |
| <span class="sourceLineNo">332</span> prevTotalSize = benchmarkEncoder(prevTotalSize, codec);<a name="line.332"></a> |
| <span class="sourceLineNo">333</span> }<a name="line.333"></a> |
| <span class="sourceLineNo">334</span><a name="line.334"></a> |
| <span class="sourceLineNo">335</span> benchmarkDefaultCompression(prevTotalSize, rawKVs);<a name="line.335"></a> |
| <span class="sourceLineNo">336</span> }<a name="line.336"></a> |
| <span class="sourceLineNo">337</span><a name="line.337"></a> |
| <span class="sourceLineNo">338</span> /**<a name="line.338"></a> |
| <span class="sourceLineNo">339</span> * Benchmark compression/decompression throughput.<a name="line.339"></a> |
| <span class="sourceLineNo">340</span> * @param previousTotalSize Total size used for verification. Use -1 if<a name="line.340"></a> |
| <span class="sourceLineNo">341</span> * unknown.<a name="line.341"></a> |
| <span class="sourceLineNo">342</span> * @param codec Tested encoder.<a name="line.342"></a> |
| <span class="sourceLineNo">343</span> * @return Size of uncompressed data.<a name="line.343"></a> |
| <span class="sourceLineNo">344</span> */<a name="line.344"></a> |
| <span class="sourceLineNo">345</span> private int benchmarkEncoder(int previousTotalSize, EncodedDataBlock codec) {<a name="line.345"></a> |
| <span class="sourceLineNo">346</span> int prevTotalSize = previousTotalSize;<a name="line.346"></a> |
| <span class="sourceLineNo">347</span> int totalSize = 0;<a name="line.347"></a> |
| <span class="sourceLineNo">348</span><a name="line.348"></a> |
| <span class="sourceLineNo">349</span> // decompression time<a name="line.349"></a> |
| <span class="sourceLineNo">350</span> List<Long> durations = new ArrayList<>();<a name="line.350"></a> |
| <span class="sourceLineNo">351</span> for (int itTime = 0; itTime < benchmarkNTimes; ++itTime) {<a name="line.351"></a> |
| <span class="sourceLineNo">352</span> totalSize = 0;<a name="line.352"></a> |
| <span class="sourceLineNo">353</span><a name="line.353"></a> |
| <span class="sourceLineNo">354</span> Iterator<Cell> it;<a name="line.354"></a> |
| <span class="sourceLineNo">355</span><a name="line.355"></a> |
| <span class="sourceLineNo">356</span> it = codec.getIterator(HFileBlock.headerSize(useHBaseChecksum));<a name="line.356"></a> |
| <span class="sourceLineNo">357</span><a name="line.357"></a> |
| <span class="sourceLineNo">358</span> // count only the algorithm time, without memory allocations<a name="line.358"></a> |
| <span class="sourceLineNo">359</span> // (expect first time)<a name="line.359"></a> |
| <span class="sourceLineNo">360</span> final long startTime = System.nanoTime();<a name="line.360"></a> |
| <span class="sourceLineNo">361</span> while (it.hasNext()) {<a name="line.361"></a> |
| <span class="sourceLineNo">362</span> totalSize += KeyValueUtil.ensureKeyValue(it.next()).getLength();<a name="line.362"></a> |
| <span class="sourceLineNo">363</span> }<a name="line.363"></a> |
| <span class="sourceLineNo">364</span> final long finishTime = System.nanoTime();<a name="line.364"></a> |
| <span class="sourceLineNo">365</span> if (itTime >= benchmarkNOmit) {<a name="line.365"></a> |
| <span class="sourceLineNo">366</span> durations.add(finishTime - startTime);<a name="line.366"></a> |
| <span class="sourceLineNo">367</span> }<a name="line.367"></a> |
| <span class="sourceLineNo">368</span><a name="line.368"></a> |
| <span class="sourceLineNo">369</span> if (prevTotalSize != -1 && prevTotalSize != totalSize) {<a name="line.369"></a> |
| <span class="sourceLineNo">370</span> throw new IllegalStateException(String.format(<a name="line.370"></a> |
| <span class="sourceLineNo">371</span> "Algorithm '%s' decoded data to different size", codec.toString()));<a name="line.371"></a> |
| <span class="sourceLineNo">372</span> }<a name="line.372"></a> |
| <span class="sourceLineNo">373</span> prevTotalSize = totalSize;<a name="line.373"></a> |
| <span class="sourceLineNo">374</span> }<a name="line.374"></a> |
| <span class="sourceLineNo">375</span><a name="line.375"></a> |
| <span class="sourceLineNo">376</span> List<Long> encodingDurations = new ArrayList<>();<a name="line.376"></a> |
| <span class="sourceLineNo">377</span> for (int itTime = 0; itTime < benchmarkNTimes; ++itTime) {<a name="line.377"></a> |
| <span class="sourceLineNo">378</span> final long startTime = System.nanoTime();<a name="line.378"></a> |
| <span class="sourceLineNo">379</span> codec.encodeData();<a name="line.379"></a> |
| <span class="sourceLineNo">380</span> final long finishTime = System.nanoTime();<a name="line.380"></a> |
| <span class="sourceLineNo">381</span> if (itTime >= benchmarkNOmit) {<a name="line.381"></a> |
| <span class="sourceLineNo">382</span> encodingDurations.add(finishTime - startTime);<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> System.out.println(codec.toString() + ":");<a name="line.386"></a> |
| <span class="sourceLineNo">387</span> printBenchmarkResult(totalSize, encodingDurations, Manipulation.ENCODING);<a name="line.387"></a> |
| <span class="sourceLineNo">388</span> printBenchmarkResult(totalSize, durations, Manipulation.DECODING);<a name="line.388"></a> |
| <span class="sourceLineNo">389</span> System.out.println();<a name="line.389"></a> |
| <span class="sourceLineNo">390</span><a name="line.390"></a> |
| <span class="sourceLineNo">391</span> return prevTotalSize;<a name="line.391"></a> |
| <span class="sourceLineNo">392</span> }<a name="line.392"></a> |
| <span class="sourceLineNo">393</span><a name="line.393"></a> |
| <span class="sourceLineNo">394</span> private void benchmarkDefaultCompression(int totalSize, byte[] rawBuffer)<a name="line.394"></a> |
| <span class="sourceLineNo">395</span> throws IOException {<a name="line.395"></a> |
| <span class="sourceLineNo">396</span> benchmarkAlgorithm(compressionAlgorithm,<a name="line.396"></a> |
| <span class="sourceLineNo">397</span> compressionAlgorithmName.toUpperCase(Locale.ROOT), rawBuffer, 0, totalSize);<a name="line.397"></a> |
| <span class="sourceLineNo">398</span> }<a name="line.398"></a> |
| <span class="sourceLineNo">399</span><a name="line.399"></a> |
| <span class="sourceLineNo">400</span> /**<a name="line.400"></a> |
| <span class="sourceLineNo">401</span> * Check decompress performance of a given algorithm and print it.<a name="line.401"></a> |
| <span class="sourceLineNo">402</span> * @param algorithm Compression algorithm.<a name="line.402"></a> |
| <span class="sourceLineNo">403</span> * @param name Name of algorithm.<a name="line.403"></a> |
| <span class="sourceLineNo">404</span> * @param buffer Buffer to be compressed.<a name="line.404"></a> |
| <span class="sourceLineNo">405</span> * @param offset Position of the beginning of the data.<a name="line.405"></a> |
| <span class="sourceLineNo">406</span> * @param length Length of data in buffer.<a name="line.406"></a> |
| <span class="sourceLineNo">407</span> * @throws IOException<a name="line.407"></a> |
| <span class="sourceLineNo">408</span> */<a name="line.408"></a> |
| <span class="sourceLineNo">409</span> public void benchmarkAlgorithm(Compression.Algorithm algorithm, String name,<a name="line.409"></a> |
| <span class="sourceLineNo">410</span> byte[] buffer, int offset, int length) throws IOException {<a name="line.410"></a> |
| <span class="sourceLineNo">411</span> System.out.println(name + ":");<a name="line.411"></a> |
| <span class="sourceLineNo">412</span><a name="line.412"></a> |
| <span class="sourceLineNo">413</span> // compress it<a name="line.413"></a> |
| <span class="sourceLineNo">414</span> List<Long> compressDurations = new ArrayList<>();<a name="line.414"></a> |
| <span class="sourceLineNo">415</span> ByteArrayOutputStream compressedStream = new ByteArrayOutputStream();<a name="line.415"></a> |
| <span class="sourceLineNo">416</span> CompressionOutputStream compressingStream =<a name="line.416"></a> |
| <span class="sourceLineNo">417</span> algorithm.createPlainCompressionStream(compressedStream, compressor);<a name="line.417"></a> |
| <span class="sourceLineNo">418</span> try {<a name="line.418"></a> |
| <span class="sourceLineNo">419</span> for (int itTime = 0; itTime < benchmarkNTimes; ++itTime) {<a name="line.419"></a> |
| <span class="sourceLineNo">420</span> final long startTime = System.nanoTime();<a name="line.420"></a> |
| <span class="sourceLineNo">421</span> // The compressedStream should reset before compressingStream resetState since in GZ<a name="line.421"></a> |
| <span class="sourceLineNo">422</span> // resetStatue will write header in the outputstream.<a name="line.422"></a> |
| <span class="sourceLineNo">423</span> compressedStream.reset();<a name="line.423"></a> |
| <span class="sourceLineNo">424</span> compressingStream.resetState();<a name="line.424"></a> |
| <span class="sourceLineNo">425</span> compressingStream.write(buffer, offset, length);<a name="line.425"></a> |
| <span class="sourceLineNo">426</span> compressingStream.flush();<a name="line.426"></a> |
| <span class="sourceLineNo">427</span> compressedStream.toByteArray();<a name="line.427"></a> |
| <span class="sourceLineNo">428</span><a name="line.428"></a> |
| <span class="sourceLineNo">429</span> final long finishTime = System.nanoTime();<a name="line.429"></a> |
| <span class="sourceLineNo">430</span><a name="line.430"></a> |
| <span class="sourceLineNo">431</span> // add time record<a name="line.431"></a> |
| <span class="sourceLineNo">432</span> if (itTime >= benchmarkNOmit) {<a name="line.432"></a> |
| <span class="sourceLineNo">433</span> compressDurations.add(finishTime - startTime);<a name="line.433"></a> |
| <span class="sourceLineNo">434</span> }<a name="line.434"></a> |
| <span class="sourceLineNo">435</span> }<a name="line.435"></a> |
| <span class="sourceLineNo">436</span> } catch (IOException e) {<a name="line.436"></a> |
| <span class="sourceLineNo">437</span> throw new RuntimeException(String.format(<a name="line.437"></a> |
| <span class="sourceLineNo">438</span> "Benchmark, or encoding algorithm '%s' cause some stream problems",<a name="line.438"></a> |
| <span class="sourceLineNo">439</span> name), e);<a name="line.439"></a> |
| <span class="sourceLineNo">440</span> }<a name="line.440"></a> |
| <span class="sourceLineNo">441</span> compressingStream.close();<a name="line.441"></a> |
| <span class="sourceLineNo">442</span> printBenchmarkResult(length, compressDurations, Manipulation.COMPRESSION);<a name="line.442"></a> |
| <span class="sourceLineNo">443</span><a name="line.443"></a> |
| <span class="sourceLineNo">444</span> byte[] compBuffer = compressedStream.toByteArray();<a name="line.444"></a> |
| <span class="sourceLineNo">445</span><a name="line.445"></a> |
| <span class="sourceLineNo">446</span> // uncompress it several times and measure performance<a name="line.446"></a> |
| <span class="sourceLineNo">447</span> List<Long> durations = new ArrayList<>();<a name="line.447"></a> |
| <span class="sourceLineNo">448</span> for (int itTime = 0; itTime < benchmarkNTimes; ++itTime) {<a name="line.448"></a> |
| <span class="sourceLineNo">449</span> final long startTime = System.nanoTime();<a name="line.449"></a> |
| <span class="sourceLineNo">450</span> byte[] newBuf = new byte[length + 1];<a name="line.450"></a> |
| <span class="sourceLineNo">451</span><a name="line.451"></a> |
| <span class="sourceLineNo">452</span> try {<a name="line.452"></a> |
| <span class="sourceLineNo">453</span> ByteArrayInputStream downStream = new ByteArrayInputStream(compBuffer,<a name="line.453"></a> |
| <span class="sourceLineNo">454</span> 0, compBuffer.length);<a name="line.454"></a> |
| <span class="sourceLineNo">455</span> InputStream decompressedStream = algorithm.createDecompressionStream(<a name="line.455"></a> |
| <span class="sourceLineNo">456</span> downStream, decompressor, 0);<a name="line.456"></a> |
| <span class="sourceLineNo">457</span><a name="line.457"></a> |
| <span class="sourceLineNo">458</span> int destOffset = 0;<a name="line.458"></a> |
| <span class="sourceLineNo">459</span> int nextChunk;<a name="line.459"></a> |
| <span class="sourceLineNo">460</span> while ((nextChunk = decompressedStream.available()) > 0) {<a name="line.460"></a> |
| <span class="sourceLineNo">461</span> destOffset += decompressedStream.read(newBuf, destOffset, nextChunk);<a name="line.461"></a> |
| <span class="sourceLineNo">462</span> }<a name="line.462"></a> |
| <span class="sourceLineNo">463</span> decompressedStream.close();<a name="line.463"></a> |
| <span class="sourceLineNo">464</span><a name="line.464"></a> |
| <span class="sourceLineNo">465</span> } catch (IOException e) {<a name="line.465"></a> |
| <span class="sourceLineNo">466</span> throw new RuntimeException(String.format(<a name="line.466"></a> |
| <span class="sourceLineNo">467</span> "Decoding path in '%s' algorithm cause exception ", name), e);<a name="line.467"></a> |
| <span class="sourceLineNo">468</span> }<a name="line.468"></a> |
| <span class="sourceLineNo">469</span><a name="line.469"></a> |
| <span class="sourceLineNo">470</span> final long finishTime = System.nanoTime();<a name="line.470"></a> |
| <span class="sourceLineNo">471</span><a name="line.471"></a> |
| <span class="sourceLineNo">472</span> // check correctness<a name="line.472"></a> |
| <span class="sourceLineNo">473</span> if (0 != Bytes.compareTo(buffer, 0, length, newBuf, 0, length)) {<a name="line.473"></a> |
| <span class="sourceLineNo">474</span> int prefix = 0;<a name="line.474"></a> |
| <span class="sourceLineNo">475</span> for(; prefix < buffer.length && prefix < newBuf.length; ++prefix) {<a name="line.475"></a> |
| <span class="sourceLineNo">476</span> if (buffer[prefix] != newBuf[prefix]) {<a name="line.476"></a> |
| <span class="sourceLineNo">477</span> break;<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> throw new RuntimeException(String.format(<a name="line.480"></a> |
| <span class="sourceLineNo">481</span> "Algorithm '%s' is corrupting the data", name));<a name="line.481"></a> |
| <span class="sourceLineNo">482</span> }<a name="line.482"></a> |
| <span class="sourceLineNo">483</span><a name="line.483"></a> |
| <span class="sourceLineNo">484</span> // add time record<a name="line.484"></a> |
| <span class="sourceLineNo">485</span> if (itTime >= benchmarkNOmit) {<a name="line.485"></a> |
| <span class="sourceLineNo">486</span> durations.add(finishTime - startTime);<a name="line.486"></a> |
| <span class="sourceLineNo">487</span> }<a name="line.487"></a> |
| <span class="sourceLineNo">488</span> }<a name="line.488"></a> |
| <span class="sourceLineNo">489</span> printBenchmarkResult(length, durations, Manipulation.DECOMPRESSION);<a name="line.489"></a> |
| <span class="sourceLineNo">490</span> System.out.println();<a name="line.490"></a> |
| <span class="sourceLineNo">491</span> }<a name="line.491"></a> |
| <span class="sourceLineNo">492</span><a name="line.492"></a> |
| <span class="sourceLineNo">493</span> private static final double BYTES_IN_MB = 1024 * 1024.0;<a name="line.493"></a> |
| <span class="sourceLineNo">494</span> private static final double NS_IN_SEC = 1000.0 * 1000.0 * 1000.0;<a name="line.494"></a> |
| <span class="sourceLineNo">495</span> private static final double MB_SEC_COEF = NS_IN_SEC / BYTES_IN_MB;<a name="line.495"></a> |
| <span class="sourceLineNo">496</span><a name="line.496"></a> |
| <span class="sourceLineNo">497</span> private static void printBenchmarkResult(int totalSize,<a name="line.497"></a> |
| <span class="sourceLineNo">498</span> List<Long> durationsInNanoSec, Manipulation manipulation) {<a name="line.498"></a> |
| <span class="sourceLineNo">499</span> final int n = durationsInNanoSec.size();<a name="line.499"></a> |
| <span class="sourceLineNo">500</span> long meanTime = 0;<a name="line.500"></a> |
| <span class="sourceLineNo">501</span> for (long time : durationsInNanoSec) {<a name="line.501"></a> |
| <span class="sourceLineNo">502</span> meanTime += time;<a name="line.502"></a> |
| <span class="sourceLineNo">503</span> }<a name="line.503"></a> |
| <span class="sourceLineNo">504</span> meanTime /= n;<a name="line.504"></a> |
| <span class="sourceLineNo">505</span><a name="line.505"></a> |
| <span class="sourceLineNo">506</span> double meanMBPerSec = totalSize * MB_SEC_COEF / meanTime;<a name="line.506"></a> |
| <span class="sourceLineNo">507</span> double mbPerSecSTD = 0;<a name="line.507"></a> |
| <span class="sourceLineNo">508</span> if (n > 0) {<a name="line.508"></a> |
| <span class="sourceLineNo">509</span> for (long time : durationsInNanoSec) {<a name="line.509"></a> |
| <span class="sourceLineNo">510</span> double mbPerSec = totalSize * MB_SEC_COEF / time;<a name="line.510"></a> |
| <span class="sourceLineNo">511</span> double dev = mbPerSec - meanMBPerSec;<a name="line.511"></a> |
| <span class="sourceLineNo">512</span> mbPerSecSTD += dev * dev;<a name="line.512"></a> |
| <span class="sourceLineNo">513</span> }<a name="line.513"></a> |
| <span class="sourceLineNo">514</span> mbPerSecSTD = Math.sqrt(mbPerSecSTD / n);<a name="line.514"></a> |
| <span class="sourceLineNo">515</span> }<a name="line.515"></a> |
| <span class="sourceLineNo">516</span><a name="line.516"></a> |
| <span class="sourceLineNo">517</span> outputTuple(manipulation + " performance", "%6.2f MB/s (+/- %.2f MB/s)",<a name="line.517"></a> |
| <span class="sourceLineNo">518</span> meanMBPerSec, mbPerSecSTD);<a name="line.518"></a> |
| <span class="sourceLineNo">519</span> }<a name="line.519"></a> |
| <span class="sourceLineNo">520</span><a name="line.520"></a> |
| <span class="sourceLineNo">521</span> private static void outputTuple(String caption, String format,<a name="line.521"></a> |
| <span class="sourceLineNo">522</span> Object... values) {<a name="line.522"></a> |
| <span class="sourceLineNo">523</span> if (format.startsWith(INT_FORMAT)) {<a name="line.523"></a> |
| <span class="sourceLineNo">524</span> format = "%s" + format.substring(INT_FORMAT.length());<a name="line.524"></a> |
| <span class="sourceLineNo">525</span> values[0] = DELIMITED_DECIMAL_FORMAT.format(values[0]);<a name="line.525"></a> |
| <span class="sourceLineNo">526</span> }<a name="line.526"></a> |
| <span class="sourceLineNo">527</span><a name="line.527"></a> |
| <span class="sourceLineNo">528</span> StringBuilder sb = new StringBuilder();<a name="line.528"></a> |
| <span class="sourceLineNo">529</span> sb.append(" ");<a name="line.529"></a> |
| <span class="sourceLineNo">530</span> sb.append(caption);<a name="line.530"></a> |
| <span class="sourceLineNo">531</span> sb.append(":");<a name="line.531"></a> |
| <span class="sourceLineNo">532</span><a name="line.532"></a> |
| <span class="sourceLineNo">533</span> String v = String.format(format, values);<a name="line.533"></a> |
| <span class="sourceLineNo">534</span> int padding = 60 - sb.length() - v.length();<a name="line.534"></a> |
| <span class="sourceLineNo">535</span> for (int i = 0; i < padding; ++i) {<a name="line.535"></a> |
| <span class="sourceLineNo">536</span> sb.append(' ');<a name="line.536"></a> |
| <span class="sourceLineNo">537</span> }<a name="line.537"></a> |
| <span class="sourceLineNo">538</span> sb.append(v);<a name="line.538"></a> |
| <span class="sourceLineNo">539</span> System.out.println(sb);<a name="line.539"></a> |
| <span class="sourceLineNo">540</span> }<a name="line.540"></a> |
| <span class="sourceLineNo">541</span><a name="line.541"></a> |
| <span class="sourceLineNo">542</span> /**<a name="line.542"></a> |
| <span class="sourceLineNo">543</span> * Display statistics of different compression algorithms.<a name="line.543"></a> |
| <span class="sourceLineNo">544</span> * @throws IOException<a name="line.544"></a> |
| <span class="sourceLineNo">545</span> */<a name="line.545"></a> |
| <span class="sourceLineNo">546</span> public void displayStatistics() throws IOException {<a name="line.546"></a> |
| <span class="sourceLineNo">547</span> final String comprAlgo = compressionAlgorithmName.toUpperCase(Locale.ROOT);<a name="line.547"></a> |
| <span class="sourceLineNo">548</span> long rawBytes = totalKeyLength + totalPrefixLength + totalValueLength;<a name="line.548"></a> |
| <span class="sourceLineNo">549</span><a name="line.549"></a> |
| <span class="sourceLineNo">550</span> System.out.println("Raw data size:");<a name="line.550"></a> |
| <span class="sourceLineNo">551</span> outputTuple("Raw bytes", INT_FORMAT, rawBytes);<a name="line.551"></a> |
| <span class="sourceLineNo">552</span> outputTuplePct("Key bytes", totalKeyLength);<a name="line.552"></a> |
| <span class="sourceLineNo">553</span> outputTuplePct("Value bytes", totalValueLength);<a name="line.553"></a> |
| <span class="sourceLineNo">554</span> outputTuplePct("KV infrastructure", totalPrefixLength);<a name="line.554"></a> |
| <span class="sourceLineNo">555</span> outputTuplePct("CF overhead", totalCFLength);<a name="line.555"></a> |
| <span class="sourceLineNo">556</span> outputTuplePct("Total key redundancy", totalKeyRedundancyLength);<a name="line.556"></a> |
| <span class="sourceLineNo">557</span><a name="line.557"></a> |
| <span class="sourceLineNo">558</span> int compressedSize = EncodedDataBlock.getCompressedSize(<a name="line.558"></a> |
| <span class="sourceLineNo">559</span> compressionAlgorithm, compressor, rawKVs, 0, rawKVs.length);<a name="line.559"></a> |
| <span class="sourceLineNo">560</span> outputTuple(comprAlgo + " only size", INT_FORMAT,<a name="line.560"></a> |
| <span class="sourceLineNo">561</span> compressedSize);<a name="line.561"></a> |
| <span class="sourceLineNo">562</span> outputSavings(comprAlgo + " only", compressedSize, rawBytes);<a name="line.562"></a> |
| <span class="sourceLineNo">563</span> System.out.println();<a name="line.563"></a> |
| <span class="sourceLineNo">564</span><a name="line.564"></a> |
| <span class="sourceLineNo">565</span> for (EncodedDataBlock codec : codecs) {<a name="line.565"></a> |
| <span class="sourceLineNo">566</span> System.out.println(codec.toString());<a name="line.566"></a> |
| <span class="sourceLineNo">567</span> long encodedBytes = codec.getSize();<a name="line.567"></a> |
| <span class="sourceLineNo">568</span> outputTuple("Encoded bytes", INT_FORMAT, encodedBytes);<a name="line.568"></a> |
| <span class="sourceLineNo">569</span> outputSavings("Key encoding", encodedBytes - totalValueLength,<a name="line.569"></a> |
| <span class="sourceLineNo">570</span> rawBytes - totalValueLength);<a name="line.570"></a> |
| <span class="sourceLineNo">571</span> outputSavings("Total encoding", encodedBytes, rawBytes);<a name="line.571"></a> |
| <span class="sourceLineNo">572</span><a name="line.572"></a> |
| <span class="sourceLineNo">573</span> int encodedCompressedSize = codec.getEncodedCompressedSize(<a name="line.573"></a> |
| <span class="sourceLineNo">574</span> compressionAlgorithm, compressor);<a name="line.574"></a> |
| <span class="sourceLineNo">575</span> outputTuple("Encoding + " + comprAlgo + " size", INT_FORMAT,<a name="line.575"></a> |
| <span class="sourceLineNo">576</span> encodedCompressedSize);<a name="line.576"></a> |
| <span class="sourceLineNo">577</span> outputSavings("Encoding + " + comprAlgo, encodedCompressedSize, rawBytes);<a name="line.577"></a> |
| <span class="sourceLineNo">578</span> outputSavings("Encoding with " + comprAlgo, encodedCompressedSize,<a name="line.578"></a> |
| <span class="sourceLineNo">579</span> compressedSize);<a name="line.579"></a> |
| <span class="sourceLineNo">580</span><a name="line.580"></a> |
| <span class="sourceLineNo">581</span> System.out.println();<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><a name="line.584"></a> |
| <span class="sourceLineNo">585</span> private void outputTuplePct(String caption, long size) {<a name="line.585"></a> |
| <span class="sourceLineNo">586</span> outputTuple(caption, INT_FORMAT + " (" + PCT_FORMAT + ")",<a name="line.586"></a> |
| <span class="sourceLineNo">587</span> size, size * 100.0 / rawKVs.length);<a name="line.587"></a> |
| <span class="sourceLineNo">588</span> }<a name="line.588"></a> |
| <span class="sourceLineNo">589</span><a name="line.589"></a> |
| <span class="sourceLineNo">590</span> private void outputSavings(String caption, long part, long whole) {<a name="line.590"></a> |
| <span class="sourceLineNo">591</span> double pct = 100.0 * (1 - 1.0 * part / whole);<a name="line.591"></a> |
| <span class="sourceLineNo">592</span> double times = whole * 1.0 / part;<a name="line.592"></a> |
| <span class="sourceLineNo">593</span> outputTuple(caption + " savings", PCT_FORMAT + " (%.2f x)",<a name="line.593"></a> |
| <span class="sourceLineNo">594</span> pct, times);<a name="line.594"></a> |
| <span class="sourceLineNo">595</span> }<a name="line.595"></a> |
| <span class="sourceLineNo">596</span><a name="line.596"></a> |
| <span class="sourceLineNo">597</span> /**<a name="line.597"></a> |
| <span class="sourceLineNo">598</span> * Test a data block encoder on the given HFile. Output results to console.<a name="line.598"></a> |
| <span class="sourceLineNo">599</span> * @param kvLimit The limit of KeyValue which will be analyzed.<a name="line.599"></a> |
| <span class="sourceLineNo">600</span> * @param hfilePath an HFile path on the file system.<a name="line.600"></a> |
| <span class="sourceLineNo">601</span> * @param compressionName Compression algorithm used for comparison.<a name="line.601"></a> |
| <span class="sourceLineNo">602</span> * @param doBenchmark Run performance benchmarks.<a name="line.602"></a> |
| <span class="sourceLineNo">603</span> * @param doVerify Verify correctness.<a name="line.603"></a> |
| <span class="sourceLineNo">604</span> * @throws IOException When pathName is incorrect.<a name="line.604"></a> |
| <span class="sourceLineNo">605</span> */<a name="line.605"></a> |
| <span class="sourceLineNo">606</span> public static void testCodecs(Configuration conf, int kvLimit,<a name="line.606"></a> |
| <span class="sourceLineNo">607</span> String hfilePath, String compressionName, boolean doBenchmark,<a name="line.607"></a> |
| <span class="sourceLineNo">608</span> boolean doVerify) throws IOException {<a name="line.608"></a> |
| <span class="sourceLineNo">609</span> // create environment<a name="line.609"></a> |
| <span class="sourceLineNo">610</span> Path path = new Path(hfilePath);<a name="line.610"></a> |
| <span class="sourceLineNo">611</span> CacheConfig cacheConf = new CacheConfig(conf);<a name="line.611"></a> |
| <span class="sourceLineNo">612</span> FileSystem fs = FileSystem.get(conf);<a name="line.612"></a> |
| <span class="sourceLineNo">613</span> HStoreFile hsf = new HStoreFile(fs, path, conf, cacheConf, BloomType.NONE, true);<a name="line.613"></a> |
| <span class="sourceLineNo">614</span> hsf.initReader();<a name="line.614"></a> |
| <span class="sourceLineNo">615</span> StoreFileReader reader = hsf.getReader();<a name="line.615"></a> |
| <span class="sourceLineNo">616</span> reader.loadFileInfo();<a name="line.616"></a> |
| <span class="sourceLineNo">617</span> KeyValueScanner scanner = reader.getStoreFileScanner(true, true,<a name="line.617"></a> |
| <span class="sourceLineNo">618</span> false, hsf.getMaxMemStoreTS(), 0, false);<a name="line.618"></a> |
| <span class="sourceLineNo">619</span> USE_TAG = reader.getHFileReader().getFileContext().isIncludesTags();<a name="line.619"></a> |
| <span class="sourceLineNo">620</span> // run the utilities<a name="line.620"></a> |
| <span class="sourceLineNo">621</span> DataBlockEncodingTool comp = new DataBlockEncodingTool(compressionName);<a name="line.621"></a> |
| <span class="sourceLineNo">622</span> int majorVersion = reader.getHFileVersion();<a name="line.622"></a> |
| <span class="sourceLineNo">623</span> comp.useHBaseChecksum = majorVersion > 2 ||<a name="line.623"></a> |
| <span class="sourceLineNo">624</span> (majorVersion == 2 &&<a name="line.624"></a> |
| <span class="sourceLineNo">625</span> reader.getHFileMinorVersion() >= HFileReaderImpl.MINOR_VERSION_WITH_CHECKSUM);<a name="line.625"></a> |
| <span class="sourceLineNo">626</span> comp.checkStatistics(scanner, kvLimit);<a name="line.626"></a> |
| <span class="sourceLineNo">627</span> if (doVerify) {<a name="line.627"></a> |
| <span class="sourceLineNo">628</span> comp.verifyCodecs(scanner, kvLimit);<a name="line.628"></a> |
| <span class="sourceLineNo">629</span> }<a name="line.629"></a> |
| <span class="sourceLineNo">630</span> if (doBenchmark) {<a name="line.630"></a> |
| <span class="sourceLineNo">631</span> comp.benchmarkCodecs();<a name="line.631"></a> |
| <span class="sourceLineNo">632</span> }<a name="line.632"></a> |
| <span class="sourceLineNo">633</span> comp.displayStatistics();<a name="line.633"></a> |
| <span class="sourceLineNo">634</span><a name="line.634"></a> |
| <span class="sourceLineNo">635</span> // cleanup<a name="line.635"></a> |
| <span class="sourceLineNo">636</span> scanner.close();<a name="line.636"></a> |
| <span class="sourceLineNo">637</span> reader.close(cacheConf.shouldEvictOnClose());<a name="line.637"></a> |
| <span class="sourceLineNo">638</span> }<a name="line.638"></a> |
| <span class="sourceLineNo">639</span><a name="line.639"></a> |
| <span class="sourceLineNo">640</span> private static void printUsage(Options options) {<a name="line.640"></a> |
| <span class="sourceLineNo">641</span> System.err.println("Usage:");<a name="line.641"></a> |
| <span class="sourceLineNo">642</span> System.err.println(String.format("./hbase %s <options>",<a name="line.642"></a> |
| <span class="sourceLineNo">643</span> DataBlockEncodingTool.class.getName()));<a name="line.643"></a> |
| <span class="sourceLineNo">644</span> System.err.println("Options:");<a name="line.644"></a> |
| <span class="sourceLineNo">645</span> for (Object it : options.getOptions()) {<a name="line.645"></a> |
| <span class="sourceLineNo">646</span> Option opt = (Option) it;<a name="line.646"></a> |
| <span class="sourceLineNo">647</span> if (opt.hasArg()) {<a name="line.647"></a> |
| <span class="sourceLineNo">648</span> System.err.println(String.format("-%s %s: %s", opt.getOpt(),<a name="line.648"></a> |
| <span class="sourceLineNo">649</span> opt.getArgName(), opt.getDescription()));<a name="line.649"></a> |
| <span class="sourceLineNo">650</span> } else {<a name="line.650"></a> |
| <span class="sourceLineNo">651</span> System.err.println(String.format("-%s: %s", opt.getOpt(),<a name="line.651"></a> |
| <span class="sourceLineNo">652</span> opt.getDescription()));<a name="line.652"></a> |
| <span class="sourceLineNo">653</span> }<a name="line.653"></a> |
| <span class="sourceLineNo">654</span> }<a name="line.654"></a> |
| <span class="sourceLineNo">655</span> }<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> * A command line interface to benchmarks. Parses command-line arguments and<a name="line.658"></a> |
| <span class="sourceLineNo">659</span> * runs the appropriate benchmarks.<a name="line.659"></a> |
| <span class="sourceLineNo">660</span> * @param args Should have length at least 1 and holds the file path to HFile.<a name="line.660"></a> |
| <span class="sourceLineNo">661</span> * @throws IOException If you specified the wrong file.<a name="line.661"></a> |
| <span class="sourceLineNo">662</span> */<a name="line.662"></a> |
| <span class="sourceLineNo">663</span> public static void main(final String[] args) throws IOException {<a name="line.663"></a> |
| <span class="sourceLineNo">664</span> // set up user arguments<a name="line.664"></a> |
| <span class="sourceLineNo">665</span> Options options = new Options();<a name="line.665"></a> |
| <span class="sourceLineNo">666</span> options.addOption(OPT_HFILE_NAME, true, "HFile to analyse (REQUIRED)");<a name="line.666"></a> |
| <span class="sourceLineNo">667</span> options.getOption(OPT_HFILE_NAME).setArgName("FILENAME");<a name="line.667"></a> |
| <span class="sourceLineNo">668</span> options.addOption(OPT_KV_LIMIT, true,<a name="line.668"></a> |
| <span class="sourceLineNo">669</span> "Maximum number of KeyValues to process. A benchmark stops running " +<a name="line.669"></a> |
| <span class="sourceLineNo">670</span> "after iterating over this many KV pairs.");<a name="line.670"></a> |
| <span class="sourceLineNo">671</span> options.getOption(OPT_KV_LIMIT).setArgName("NUMBER");<a name="line.671"></a> |
| <span class="sourceLineNo">672</span> options.addOption(OPT_MEASURE_THROUGHPUT, false,<a name="line.672"></a> |
| <span class="sourceLineNo">673</span> "Measure read throughput");<a name="line.673"></a> |
| <span class="sourceLineNo">674</span> options.addOption(OPT_OMIT_CORRECTNESS_TEST, false,<a name="line.674"></a> |
| <span class="sourceLineNo">675</span> "Omit corectness tests.");<a name="line.675"></a> |
| <span class="sourceLineNo">676</span> options.addOption(OPT_COMPRESSION_ALGORITHM, true,<a name="line.676"></a> |
| <span class="sourceLineNo">677</span> "What kind of compression algorithm use for comparison.");<a name="line.677"></a> |
| <span class="sourceLineNo">678</span> options.addOption(OPT_BENCHMARK_N_TIMES,<a name="line.678"></a> |
| <span class="sourceLineNo">679</span> true, "Number of times to run each benchmark. Default value: " +<a name="line.679"></a> |
| <span class="sourceLineNo">680</span> DEFAULT_BENCHMARK_N_TIMES);<a name="line.680"></a> |
| <span class="sourceLineNo">681</span> options.addOption(OPT_BENCHMARK_N_OMIT, true,<a name="line.681"></a> |
| <span class="sourceLineNo">682</span> "Number of first runs of every benchmark to exclude from "<a name="line.682"></a> |
| <span class="sourceLineNo">683</span> + "statistics (" + DEFAULT_BENCHMARK_N_OMIT<a name="line.683"></a> |
| <span class="sourceLineNo">684</span> + " by default, so that " + "only the last "<a name="line.684"></a> |
| <span class="sourceLineNo">685</span> + (DEFAULT_BENCHMARK_N_TIMES - DEFAULT_BENCHMARK_N_OMIT)<a name="line.685"></a> |
| <span class="sourceLineNo">686</span> + " times are included in statistics.)");<a name="line.686"></a> |
| <span class="sourceLineNo">687</span><a name="line.687"></a> |
| <span class="sourceLineNo">688</span> // parse arguments<a name="line.688"></a> |
| <span class="sourceLineNo">689</span> CommandLineParser parser = new PosixParser();<a name="line.689"></a> |
| <span class="sourceLineNo">690</span> CommandLine cmd = null;<a name="line.690"></a> |
| <span class="sourceLineNo">691</span> try {<a name="line.691"></a> |
| <span class="sourceLineNo">692</span> cmd = parser.parse(options, args);<a name="line.692"></a> |
| <span class="sourceLineNo">693</span> } catch (ParseException e) {<a name="line.693"></a> |
| <span class="sourceLineNo">694</span> System.err.println("Could not parse arguments!");<a name="line.694"></a> |
| <span class="sourceLineNo">695</span> System.exit(-1);<a name="line.695"></a> |
| <span class="sourceLineNo">696</span> return; // avoid warning<a name="line.696"></a> |
| <span class="sourceLineNo">697</span> }<a name="line.697"></a> |
| <span class="sourceLineNo">698</span><a name="line.698"></a> |
| <span class="sourceLineNo">699</span> int kvLimit = Integer.MAX_VALUE;<a name="line.699"></a> |
| <span class="sourceLineNo">700</span> if (cmd.hasOption(OPT_KV_LIMIT)) {<a name="line.700"></a> |
| <span class="sourceLineNo">701</span> kvLimit = Integer.parseInt(cmd.getOptionValue(OPT_KV_LIMIT));<a name="line.701"></a> |
| <span class="sourceLineNo">702</span> if (kvLimit <= 0) {<a name="line.702"></a> |
| <span class="sourceLineNo">703</span> LOG.error("KV_LIMIT should not less than 1.");<a name="line.703"></a> |
| <span class="sourceLineNo">704</span> }<a name="line.704"></a> |
| <span class="sourceLineNo">705</span> }<a name="line.705"></a> |
| <span class="sourceLineNo">706</span><a name="line.706"></a> |
| <span class="sourceLineNo">707</span> // basic argument sanity checks<a name="line.707"></a> |
| <span class="sourceLineNo">708</span> if (!cmd.hasOption(OPT_HFILE_NAME)) {<a name="line.708"></a> |
| <span class="sourceLineNo">709</span> LOG.error("Please specify HFile name using the " + OPT_HFILE_NAME<a name="line.709"></a> |
| <span class="sourceLineNo">710</span> + " option");<a name="line.710"></a> |
| <span class="sourceLineNo">711</span> printUsage(options);<a name="line.711"></a> |
| <span class="sourceLineNo">712</span> System.exit(-1);<a name="line.712"></a> |
| <span class="sourceLineNo">713</span> }<a name="line.713"></a> |
| <span class="sourceLineNo">714</span><a name="line.714"></a> |
| <span class="sourceLineNo">715</span> String pathName = cmd.getOptionValue(OPT_HFILE_NAME);<a name="line.715"></a> |
| <span class="sourceLineNo">716</span> String compressionName = DEFAULT_COMPRESSION.getName();<a name="line.716"></a> |
| <span class="sourceLineNo">717</span> if (cmd.hasOption(OPT_COMPRESSION_ALGORITHM)) {<a name="line.717"></a> |
| <span class="sourceLineNo">718</span> compressionName =<a name="line.718"></a> |
| <span class="sourceLineNo">719</span> cmd.getOptionValue(OPT_COMPRESSION_ALGORITHM).toLowerCase(Locale.ROOT);<a name="line.719"></a> |
| <span class="sourceLineNo">720</span> }<a name="line.720"></a> |
| <span class="sourceLineNo">721</span> boolean doBenchmark = cmd.hasOption(OPT_MEASURE_THROUGHPUT);<a name="line.721"></a> |
| <span class="sourceLineNo">722</span> boolean doVerify = !cmd.hasOption(OPT_OMIT_CORRECTNESS_TEST);<a name="line.722"></a> |
| <span class="sourceLineNo">723</span><a name="line.723"></a> |
| <span class="sourceLineNo">724</span> if (cmd.hasOption(OPT_BENCHMARK_N_TIMES)) {<a name="line.724"></a> |
| <span class="sourceLineNo">725</span> benchmarkNTimes = Integer.valueOf(cmd.getOptionValue(<a name="line.725"></a> |
| <span class="sourceLineNo">726</span> OPT_BENCHMARK_N_TIMES));<a name="line.726"></a> |
| <span class="sourceLineNo">727</span> }<a name="line.727"></a> |
| <span class="sourceLineNo">728</span> if (cmd.hasOption(OPT_BENCHMARK_N_OMIT)) {<a name="line.728"></a> |
| <span class="sourceLineNo">729</span> benchmarkNOmit =<a name="line.729"></a> |
| <span class="sourceLineNo">730</span> Integer.valueOf(cmd.getOptionValue(OPT_BENCHMARK_N_OMIT));<a name="line.730"></a> |
| <span class="sourceLineNo">731</span> }<a name="line.731"></a> |
| <span class="sourceLineNo">732</span> if (benchmarkNTimes < benchmarkNOmit) {<a name="line.732"></a> |
| <span class="sourceLineNo">733</span> LOG.error("The number of times to run each benchmark ("<a name="line.733"></a> |
| <span class="sourceLineNo">734</span> + benchmarkNTimes<a name="line.734"></a> |
| <span class="sourceLineNo">735</span> + ") must be greater than the number of benchmark runs to exclude "<a name="line.735"></a> |
| <span class="sourceLineNo">736</span> + "from statistics (" + benchmarkNOmit + ")");<a name="line.736"></a> |
| <span class="sourceLineNo">737</span> System.exit(1);<a name="line.737"></a> |
| <span class="sourceLineNo">738</span> }<a name="line.738"></a> |
| <span class="sourceLineNo">739</span> LOG.info("Running benchmark " + benchmarkNTimes + " times. " +<a name="line.739"></a> |
| <span class="sourceLineNo">740</span> "Excluding the first " + benchmarkNOmit + " times from statistics.");<a name="line.740"></a> |
| <span class="sourceLineNo">741</span><a name="line.741"></a> |
| <span class="sourceLineNo">742</span> final Configuration conf = HBaseConfiguration.create();<a name="line.742"></a> |
| <span class="sourceLineNo">743</span> testCodecs(conf, kvLimit, pathName, compressionName, doBenchmark, doVerify);<a name="line.743"></a> |
| <span class="sourceLineNo">744</span> }<a name="line.744"></a> |
| <span class="sourceLineNo">745</span><a name="line.745"></a> |
| <span class="sourceLineNo">746</span>}<a name="line.746"></a> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| </pre> |
| </div> |
| </body> |
| </html> |