| <!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.io.hfile;<a name="line.18"></a> |
| <span class="sourceLineNo">019</span><a name="line.19"></a> |
| <span class="sourceLineNo">020</span>import static org.apache.hadoop.hbase.io.ByteBuffAllocator.HEAP;<a name="line.20"></a> |
| <span class="sourceLineNo">021</span>import static org.apache.hadoop.hbase.io.hfile.BlockCompressedSizePredicator.BLOCK_COMPRESSED_SIZE_PREDICATOR;<a name="line.21"></a> |
| <span class="sourceLineNo">022</span>import static org.apache.hadoop.hbase.io.hfile.trace.HFileContextAttributesBuilderConsumer.CONTEXT_KEY;<a name="line.22"></a> |
| <span class="sourceLineNo">023</span><a name="line.23"></a> |
| <span class="sourceLineNo">024</span>import io.opentelemetry.api.common.Attributes;<a name="line.24"></a> |
| <span class="sourceLineNo">025</span>import io.opentelemetry.api.common.AttributesBuilder;<a name="line.25"></a> |
| <span class="sourceLineNo">026</span>import io.opentelemetry.api.trace.Span;<a name="line.26"></a> |
| <span class="sourceLineNo">027</span>import io.opentelemetry.context.Context;<a name="line.27"></a> |
| <span class="sourceLineNo">028</span>import io.opentelemetry.context.Scope;<a name="line.28"></a> |
| <span class="sourceLineNo">029</span>import java.io.DataInputStream;<a name="line.29"></a> |
| <span class="sourceLineNo">030</span>import java.io.DataOutput;<a name="line.30"></a> |
| <span class="sourceLineNo">031</span>import java.io.DataOutputStream;<a name="line.31"></a> |
| <span class="sourceLineNo">032</span>import java.io.IOException;<a name="line.32"></a> |
| <span class="sourceLineNo">033</span>import java.nio.ByteBuffer;<a name="line.33"></a> |
| <span class="sourceLineNo">034</span>import java.util.ArrayList;<a name="line.34"></a> |
| <span class="sourceLineNo">035</span>import java.util.List;<a name="line.35"></a> |
| <span class="sourceLineNo">036</span>import java.util.Optional;<a name="line.36"></a> |
| <span class="sourceLineNo">037</span>import java.util.concurrent.atomic.AtomicReference;<a name="line.37"></a> |
| <span class="sourceLineNo">038</span>import java.util.concurrent.locks.Lock;<a name="line.38"></a> |
| <span class="sourceLineNo">039</span>import java.util.concurrent.locks.ReentrantLock;<a name="line.39"></a> |
| <span class="sourceLineNo">040</span>import org.apache.hadoop.conf.Configuration;<a name="line.40"></a> |
| <span class="sourceLineNo">041</span>import org.apache.hadoop.fs.FSDataInputStream;<a name="line.41"></a> |
| <span class="sourceLineNo">042</span>import org.apache.hadoop.fs.FSDataOutputStream;<a name="line.42"></a> |
| <span class="sourceLineNo">043</span>import org.apache.hadoop.hbase.Cell;<a name="line.43"></a> |
| <span class="sourceLineNo">044</span>import org.apache.hadoop.hbase.HConstants;<a name="line.44"></a> |
| <span class="sourceLineNo">045</span>import org.apache.hadoop.hbase.fs.HFileSystem;<a name="line.45"></a> |
| <span class="sourceLineNo">046</span>import org.apache.hadoop.hbase.io.ByteArrayOutputStream;<a name="line.46"></a> |
| <span class="sourceLineNo">047</span>import org.apache.hadoop.hbase.io.ByteBuffAllocator;<a name="line.47"></a> |
| <span class="sourceLineNo">048</span>import org.apache.hadoop.hbase.io.ByteBuffInputStream;<a name="line.48"></a> |
| <span class="sourceLineNo">049</span>import org.apache.hadoop.hbase.io.ByteBufferWriterDataOutputStream;<a name="line.49"></a> |
| <span class="sourceLineNo">050</span>import org.apache.hadoop.hbase.io.FSDataInputStreamWrapper;<a name="line.50"></a> |
| <span class="sourceLineNo">051</span>import org.apache.hadoop.hbase.io.encoding.DataBlockEncoding;<a name="line.51"></a> |
| <span class="sourceLineNo">052</span>import org.apache.hadoop.hbase.io.encoding.EncodingState;<a name="line.52"></a> |
| <span class="sourceLineNo">053</span>import org.apache.hadoop.hbase.io.encoding.HFileBlockDecodingContext;<a name="line.53"></a> |
| <span class="sourceLineNo">054</span>import org.apache.hadoop.hbase.io.encoding.HFileBlockDefaultDecodingContext;<a name="line.54"></a> |
| <span class="sourceLineNo">055</span>import org.apache.hadoop.hbase.io.encoding.HFileBlockDefaultEncodingContext;<a name="line.55"></a> |
| <span class="sourceLineNo">056</span>import org.apache.hadoop.hbase.io.encoding.HFileBlockEncodingContext;<a name="line.56"></a> |
| <span class="sourceLineNo">057</span>import org.apache.hadoop.hbase.io.hfile.trace.HFileContextAttributesBuilderConsumer;<a name="line.57"></a> |
| <span class="sourceLineNo">058</span>import org.apache.hadoop.hbase.io.util.BlockIOUtils;<a name="line.58"></a> |
| <span class="sourceLineNo">059</span>import org.apache.hadoop.hbase.nio.ByteBuff;<a name="line.59"></a> |
| <span class="sourceLineNo">060</span>import org.apache.hadoop.hbase.nio.MultiByteBuff;<a name="line.60"></a> |
| <span class="sourceLineNo">061</span>import org.apache.hadoop.hbase.nio.SingleByteBuff;<a name="line.61"></a> |
| <span class="sourceLineNo">062</span>import org.apache.hadoop.hbase.regionserver.ShipperListener;<a name="line.62"></a> |
| <span class="sourceLineNo">063</span>import org.apache.hadoop.hbase.trace.HBaseSemanticAttributes.ReadType;<a name="line.63"></a> |
| <span class="sourceLineNo">064</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.64"></a> |
| <span class="sourceLineNo">065</span>import org.apache.hadoop.hbase.util.ChecksumType;<a name="line.65"></a> |
| <span class="sourceLineNo">066</span>import org.apache.hadoop.hbase.util.ClassSize;<a name="line.66"></a> |
| <span class="sourceLineNo">067</span>import org.apache.hadoop.hbase.util.EnvironmentEdgeManager;<a name="line.67"></a> |
| <span class="sourceLineNo">068</span>import org.apache.hadoop.util.ReflectionUtils;<a name="line.68"></a> |
| <span class="sourceLineNo">069</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.69"></a> |
| <span class="sourceLineNo">070</span>import org.slf4j.Logger;<a name="line.70"></a> |
| <span class="sourceLineNo">071</span>import org.slf4j.LoggerFactory;<a name="line.71"></a> |
| <span class="sourceLineNo">072</span><a name="line.72"></a> |
| <span class="sourceLineNo">073</span>import org.apache.hbase.thirdparty.com.google.common.base.Preconditions;<a name="line.73"></a> |
| <span class="sourceLineNo">074</span><a name="line.74"></a> |
| <span class="sourceLineNo">075</span>/**<a name="line.75"></a> |
| <span class="sourceLineNo">076</span> * Cacheable Blocks of an {@link HFile} version 2 file. Version 2 was introduced in hbase-0.92.0.<a name="line.76"></a> |
| <span class="sourceLineNo">077</span> * <p><a name="line.77"></a> |
| <span class="sourceLineNo">078</span> * Version 1 was the original file block. Version 2 was introduced when we changed the hbase file<a name="line.78"></a> |
| <span class="sourceLineNo">079</span> * format to support multi-level block indexes and compound bloom filters (HBASE-3857). Support for<a name="line.79"></a> |
| <span class="sourceLineNo">080</span> * Version 1 was removed in hbase-1.3.0.<a name="line.80"></a> |
| <span class="sourceLineNo">081</span> * <h3>HFileBlock: Version 2</h3> In version 2, a block is structured as follows:<a name="line.81"></a> |
| <span class="sourceLineNo">082</span> * <ul><a name="line.82"></a> |
| <span class="sourceLineNo">083</span> * <li><b>Header:</b> See Writer#putHeader() for where header is written; header total size is<a name="line.83"></a> |
| <span class="sourceLineNo">084</span> * HFILEBLOCK_HEADER_SIZE<a name="line.84"></a> |
| <span class="sourceLineNo">085</span> * <ul><a name="line.85"></a> |
| <span class="sourceLineNo">086</span> * <li>0. blockType: Magic record identifying the {@link BlockType} (8 bytes): e.g.<a name="line.86"></a> |
| <span class="sourceLineNo">087</span> * <code>DATABLK*</code><a name="line.87"></a> |
| <span class="sourceLineNo">088</span> * <li>1. onDiskSizeWithoutHeader: Compressed -- a.k.a 'on disk' -- block size, excluding header,<a name="line.88"></a> |
| <span class="sourceLineNo">089</span> * but including tailing checksum bytes (4 bytes)<a name="line.89"></a> |
| <span class="sourceLineNo">090</span> * <li>2. uncompressedSizeWithoutHeader: Uncompressed block size, excluding header, and excluding<a name="line.90"></a> |
| <span class="sourceLineNo">091</span> * checksum bytes (4 bytes)<a name="line.91"></a> |
| <span class="sourceLineNo">092</span> * <li>3. prevBlockOffset: The offset of the previous block of the same type (8 bytes). This is used<a name="line.92"></a> |
| <span class="sourceLineNo">093</span> * to navigate to the previous block without having to go to the block index<a name="line.93"></a> |
| <span class="sourceLineNo">094</span> * <li>4: For minorVersions &gt;=1, the ordinal describing checksum type (1 byte)<a name="line.94"></a> |
| <span class="sourceLineNo">095</span> * <li>5: For minorVersions &gt;=1, the number of data bytes/checksum chunk (4 bytes)<a name="line.95"></a> |
| <span class="sourceLineNo">096</span> * <li>6: onDiskDataSizeWithHeader: For minorVersions &gt;=1, the size of data 'on disk', including<a name="line.96"></a> |
| <span class="sourceLineNo">097</span> * header, excluding checksums (4 bytes)<a name="line.97"></a> |
| <span class="sourceLineNo">098</span> * </ul><a name="line.98"></a> |
| <span class="sourceLineNo">099</span> * </li><a name="line.99"></a> |
| <span class="sourceLineNo">100</span> * <li><b>Raw/Compressed/Encrypted/Encoded data:</b> The compression algorithm is the same for all<a name="line.100"></a> |
| <span class="sourceLineNo">101</span> * the blocks in an {@link HFile}. If compression is NONE, this is just raw, serialized Cells.<a name="line.101"></a> |
| <span class="sourceLineNo">102</span> * <li><b>Tail:</b> For minorVersions &gt;=1, a series of 4 byte checksums, one each for the number<a name="line.102"></a> |
| <span class="sourceLineNo">103</span> * of bytes specified by bytesPerChecksum.<a name="line.103"></a> |
| <span class="sourceLineNo">104</span> * </ul><a name="line.104"></a> |
| <span class="sourceLineNo">105</span> * <h3>Caching</h3> Caches cache whole blocks with trailing checksums if any. We then tag on some<a name="line.105"></a> |
| <span class="sourceLineNo">106</span> * metadata, the content of BLOCK_METADATA_SPACE which will be flag on if we are doing 'hbase'<a name="line.106"></a> |
| <span class="sourceLineNo">107</span> * checksums and then the offset into the file which is needed when we re-make a cache key when we<a name="line.107"></a> |
| <span class="sourceLineNo">108</span> * return the block to the cache as 'done'. See {@link Cacheable#serialize(ByteBuffer, boolean)} and<a name="line.108"></a> |
| <span class="sourceLineNo">109</span> * {@link Cacheable#getDeserializer()}.<a name="line.109"></a> |
| <span class="sourceLineNo">110</span> * <p><a name="line.110"></a> |
| <span class="sourceLineNo">111</span> * TODO: Should we cache the checksums? Down in Writer#getBlockForCaching(CacheConfig) where we make<a name="line.111"></a> |
| <span class="sourceLineNo">112</span> * a block to cache-on-write, there is an attempt at turning off checksums. This is not the only<a name="line.112"></a> |
| <span class="sourceLineNo">113</span> * place we get blocks to cache. We also will cache the raw return from an hdfs read. In this case,<a name="line.113"></a> |
| <span class="sourceLineNo">114</span> * the checksums may be present. If the cache is backed by something that doesn't do ECC, say an<a name="line.114"></a> |
| <span class="sourceLineNo">115</span> * SSD, we might want to preserve checksums. For now this is open question.<a name="line.115"></a> |
| <span class="sourceLineNo">116</span> * <p><a name="line.116"></a> |
| <span class="sourceLineNo">117</span> * TODO: Over in BucketCache, we save a block allocation by doing a custom serialization. Be sure to<a name="line.117"></a> |
| <span class="sourceLineNo">118</span> * change it if serialization changes in here. Could we add a method here that takes an IOEngine and<a name="line.118"></a> |
| <span class="sourceLineNo">119</span> * that then serializes to it rather than expose our internals over in BucketCache? IOEngine is in<a name="line.119"></a> |
| <span class="sourceLineNo">120</span> * the bucket subpackage. Pull it up? Then this class knows about bucketcache. Ugh.<a name="line.120"></a> |
| <span class="sourceLineNo">121</span> */<a name="line.121"></a> |
| <span class="sourceLineNo">122</span>@InterfaceAudience.Private<a name="line.122"></a> |
| <span class="sourceLineNo">123</span>public class HFileBlock implements Cacheable {<a name="line.123"></a> |
| <span class="sourceLineNo">124</span> private static final Logger LOG = LoggerFactory.getLogger(HFileBlock.class);<a name="line.124"></a> |
| <span class="sourceLineNo">125</span> public static final long FIXED_OVERHEAD = ClassSize.estimateBase(HFileBlock.class, false);<a name="line.125"></a> |
| <span class="sourceLineNo">126</span><a name="line.126"></a> |
| <span class="sourceLineNo">127</span> // Block Header fields.<a name="line.127"></a> |
| <span class="sourceLineNo">128</span><a name="line.128"></a> |
| <span class="sourceLineNo">129</span> // TODO: encapsulate Header related logic in this inner class.<a name="line.129"></a> |
| <span class="sourceLineNo">130</span> static class Header {<a name="line.130"></a> |
| <span class="sourceLineNo">131</span> // Format of header is:<a name="line.131"></a> |
| <span class="sourceLineNo">132</span> // 8 bytes - block magic<a name="line.132"></a> |
| <span class="sourceLineNo">133</span> // 4 bytes int - onDiskSizeWithoutHeader<a name="line.133"></a> |
| <span class="sourceLineNo">134</span> // 4 bytes int - uncompressedSizeWithoutHeader<a name="line.134"></a> |
| <span class="sourceLineNo">135</span> // 8 bytes long - prevBlockOffset<a name="line.135"></a> |
| <span class="sourceLineNo">136</span> // The following 3 are only present if header contains checksum information<a name="line.136"></a> |
| <span class="sourceLineNo">137</span> // 1 byte - checksum type<a name="line.137"></a> |
| <span class="sourceLineNo">138</span> // 4 byte int - bytes per checksum<a name="line.138"></a> |
| <span class="sourceLineNo">139</span> // 4 byte int - onDiskDataSizeWithHeader<a name="line.139"></a> |
| <span class="sourceLineNo">140</span> static int BLOCK_MAGIC_INDEX = 0;<a name="line.140"></a> |
| <span class="sourceLineNo">141</span> static int ON_DISK_SIZE_WITHOUT_HEADER_INDEX = 8;<a name="line.141"></a> |
| <span class="sourceLineNo">142</span> static int UNCOMPRESSED_SIZE_WITHOUT_HEADER_INDEX = 12;<a name="line.142"></a> |
| <span class="sourceLineNo">143</span> static int PREV_BLOCK_OFFSET_INDEX = 16;<a name="line.143"></a> |
| <span class="sourceLineNo">144</span> static int CHECKSUM_TYPE_INDEX = 24;<a name="line.144"></a> |
| <span class="sourceLineNo">145</span> static int BYTES_PER_CHECKSUM_INDEX = 25;<a name="line.145"></a> |
| <span class="sourceLineNo">146</span> static int ON_DISK_DATA_SIZE_WITH_HEADER_INDEX = 29;<a name="line.146"></a> |
| <span class="sourceLineNo">147</span> }<a name="line.147"></a> |
| <span class="sourceLineNo">148</span><a name="line.148"></a> |
| <span class="sourceLineNo">149</span> /** Type of block. Header field 0. */<a name="line.149"></a> |
| <span class="sourceLineNo">150</span> private BlockType blockType;<a name="line.150"></a> |
| <span class="sourceLineNo">151</span><a name="line.151"></a> |
| <span class="sourceLineNo">152</span> /**<a name="line.152"></a> |
| <span class="sourceLineNo">153</span> * Size on disk excluding header, including checksum. Header field 1.<a name="line.153"></a> |
| <span class="sourceLineNo">154</span> * @see Writer#putHeader(byte[], int, int, int, int)<a name="line.154"></a> |
| <span class="sourceLineNo">155</span> */<a name="line.155"></a> |
| <span class="sourceLineNo">156</span> private int onDiskSizeWithoutHeader;<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> * Size of pure data. Does not include header or checksums. Header field 2.<a name="line.159"></a> |
| <span class="sourceLineNo">160</span> * @see Writer#putHeader(byte[], int, int, int, int)<a name="line.160"></a> |
| <span class="sourceLineNo">161</span> */<a name="line.161"></a> |
| <span class="sourceLineNo">162</span> private int uncompressedSizeWithoutHeader;<a name="line.162"></a> |
| <span class="sourceLineNo">163</span><a name="line.163"></a> |
| <span class="sourceLineNo">164</span> /**<a name="line.164"></a> |
| <span class="sourceLineNo">165</span> * The offset of the previous block on disk. Header field 3.<a name="line.165"></a> |
| <span class="sourceLineNo">166</span> * @see Writer#putHeader(byte[], int, int, int, int)<a name="line.166"></a> |
| <span class="sourceLineNo">167</span> */<a name="line.167"></a> |
| <span class="sourceLineNo">168</span> private long prevBlockOffset;<a name="line.168"></a> |
| <span class="sourceLineNo">169</span><a name="line.169"></a> |
| <span class="sourceLineNo">170</span> /**<a name="line.170"></a> |
| <span class="sourceLineNo">171</span> * Size on disk of header + data. Excludes checksum. Header field 6, OR calculated from<a name="line.171"></a> |
| <span class="sourceLineNo">172</span> * {@link #onDiskSizeWithoutHeader} when using HDFS checksum.<a name="line.172"></a> |
| <span class="sourceLineNo">173</span> * @see Writer#putHeader(byte[], int, int, int, int)<a name="line.173"></a> |
| <span class="sourceLineNo">174</span> */<a name="line.174"></a> |
| <span class="sourceLineNo">175</span> private int onDiskDataSizeWithHeader;<a name="line.175"></a> |
| <span class="sourceLineNo">176</span> // End of Block Header fields.<a name="line.176"></a> |
| <span class="sourceLineNo">177</span><a name="line.177"></a> |
| <span class="sourceLineNo">178</span> /**<a name="line.178"></a> |
| <span class="sourceLineNo">179</span> * The in-memory representation of the hfile block. Can be on or offheap. Can be backed by a<a name="line.179"></a> |
| <span class="sourceLineNo">180</span> * single ByteBuffer or by many. Make no assumptions.<a name="line.180"></a> |
| <span class="sourceLineNo">181</span> * <p><a name="line.181"></a> |
| <span class="sourceLineNo">182</span> * Be careful reading from this <code>buf</code>. Duplicate and work on the duplicate or if not,<a name="line.182"></a> |
| <span class="sourceLineNo">183</span> * be sure to reset position and limit else trouble down the road.<a name="line.183"></a> |
| <span class="sourceLineNo">184</span> * <p><a name="line.184"></a> |
| <span class="sourceLineNo">185</span> * TODO: Make this read-only once made.<a name="line.185"></a> |
| <span class="sourceLineNo">186</span> * <p><a name="line.186"></a> |
| <span class="sourceLineNo">187</span> * We are using the ByteBuff type. ByteBuffer is not extensible yet we need to be able to have a<a name="line.187"></a> |
| <span class="sourceLineNo">188</span> * ByteBuffer-like API across multiple ByteBuffers reading from a cache such as BucketCache. So,<a name="line.188"></a> |
| <span class="sourceLineNo">189</span> * we have this ByteBuff type. Unfortunately, it is spread all about HFileBlock. Would be good if<a name="line.189"></a> |
| <span class="sourceLineNo">190</span> * could be confined to cache-use only but hard-to-do.<a name="line.190"></a> |
| <span class="sourceLineNo">191</span> */<a name="line.191"></a> |
| <span class="sourceLineNo">192</span> private ByteBuff buf;<a name="line.192"></a> |
| <span class="sourceLineNo">193</span><a name="line.193"></a> |
| <span class="sourceLineNo">194</span> /**<a name="line.194"></a> |
| <span class="sourceLineNo">195</span> * Meta data that holds meta information on the hfileblock.<a name="line.195"></a> |
| <span class="sourceLineNo">196</span> */<a name="line.196"></a> |
| <span class="sourceLineNo">197</span> private HFileContext fileContext;<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> * The offset of this block in the file. Populated by the reader for convenience of access. This<a name="line.200"></a> |
| <span class="sourceLineNo">201</span> * offset is not part of the block header.<a name="line.201"></a> |
| <span class="sourceLineNo">202</span> */<a name="line.202"></a> |
| <span class="sourceLineNo">203</span> private long offset = UNSET;<a name="line.203"></a> |
| <span class="sourceLineNo">204</span><a name="line.204"></a> |
| <span class="sourceLineNo">205</span> /**<a name="line.205"></a> |
| <span class="sourceLineNo">206</span> * The on-disk size of the next block, including the header and checksums if present. UNSET if<a name="line.206"></a> |
| <span class="sourceLineNo">207</span> * unknown. Blocks try to carry the size of the next block to read in this data member. Usually we<a name="line.207"></a> |
| <span class="sourceLineNo">208</span> * get block sizes from the hfile index but sometimes the index is not available: e.g. when we<a name="line.208"></a> |
| <span class="sourceLineNo">209</span> * read the indexes themselves (indexes are stored in blocks, we do not have an index for the<a name="line.209"></a> |
| <span class="sourceLineNo">210</span> * indexes). Saves seeks especially around file open when there is a flurry of reading in hfile<a name="line.210"></a> |
| <span class="sourceLineNo">211</span> * metadata.<a name="line.211"></a> |
| <span class="sourceLineNo">212</span> */<a name="line.212"></a> |
| <span class="sourceLineNo">213</span> private int nextBlockOnDiskSize = UNSET;<a name="line.213"></a> |
| <span class="sourceLineNo">214</span><a name="line.214"></a> |
| <span class="sourceLineNo">215</span> private ByteBuffAllocator allocator;<a name="line.215"></a> |
| <span class="sourceLineNo">216</span><a name="line.216"></a> |
| <span class="sourceLineNo">217</span> /**<a name="line.217"></a> |
| <span class="sourceLineNo">218</span> * On a checksum failure, do these many succeeding read requests using hdfs checksums before<a name="line.218"></a> |
| <span class="sourceLineNo">219</span> * auto-reenabling hbase checksum verification.<a name="line.219"></a> |
| <span class="sourceLineNo">220</span> */<a name="line.220"></a> |
| <span class="sourceLineNo">221</span> static final int CHECKSUM_VERIFICATION_NUM_IO_THRESHOLD = 3;<a name="line.221"></a> |
| <span class="sourceLineNo">222</span><a name="line.222"></a> |
| <span class="sourceLineNo">223</span> private static int UNSET = -1;<a name="line.223"></a> |
| <span class="sourceLineNo">224</span> public static final boolean FILL_HEADER = true;<a name="line.224"></a> |
| <span class="sourceLineNo">225</span> public static final boolean DONT_FILL_HEADER = false;<a name="line.225"></a> |
| <span class="sourceLineNo">226</span><a name="line.226"></a> |
| <span class="sourceLineNo">227</span> // How to get the estimate correctly? if it is a singleBB?<a name="line.227"></a> |
| <span class="sourceLineNo">228</span> public static final int MULTI_BYTE_BUFFER_HEAP_SIZE =<a name="line.228"></a> |
| <span class="sourceLineNo">229</span> (int) ClassSize.estimateBase(MultiByteBuff.class, false);<a name="line.229"></a> |
| <span class="sourceLineNo">230</span><a name="line.230"></a> |
| <span class="sourceLineNo">231</span> /**<a name="line.231"></a> |
| <span class="sourceLineNo">232</span> * Space for metadata on a block that gets stored along with the block when we cache it. There are<a name="line.232"></a> |
| <span class="sourceLineNo">233</span> * a few bytes stuck on the end of the HFileBlock that we pull in from HDFS. 8 bytes are for the<a name="line.233"></a> |
| <span class="sourceLineNo">234</span> * offset of this block (long) in the file. Offset is important because is is used when we remake<a name="line.234"></a> |
| <span class="sourceLineNo">235</span> * the CacheKey when we return block to the cache when done. There is also a flag on whether<a name="line.235"></a> |
| <span class="sourceLineNo">236</span> * checksumming is being done by hbase or not. See class comment for note on uncertain state of<a name="line.236"></a> |
| <span class="sourceLineNo">237</span> * checksumming of blocks that come out of cache (should we or should we not?). Finally there are<a name="line.237"></a> |
| <span class="sourceLineNo">238</span> * 4 bytes to hold the length of the next block which can save a seek on occasion if available.<a name="line.238"></a> |
| <span class="sourceLineNo">239</span> * (This EXTRA info came in with original commit of the bucketcache, HBASE-7404. It was formerly<a name="line.239"></a> |
| <span class="sourceLineNo">240</span> * known as EXTRA_SERIALIZATION_SPACE).<a name="line.240"></a> |
| <span class="sourceLineNo">241</span> */<a name="line.241"></a> |
| <span class="sourceLineNo">242</span> public static final int BLOCK_METADATA_SPACE =<a name="line.242"></a> |
| <span class="sourceLineNo">243</span> Bytes.SIZEOF_BYTE + Bytes.SIZEOF_LONG + Bytes.SIZEOF_INT;<a name="line.243"></a> |
| <span class="sourceLineNo">244</span><a name="line.244"></a> |
| <span class="sourceLineNo">245</span> /**<a name="line.245"></a> |
| <span class="sourceLineNo">246</span> * Each checksum value is an integer that can be stored in 4 bytes.<a name="line.246"></a> |
| <span class="sourceLineNo">247</span> */<a name="line.247"></a> |
| <span class="sourceLineNo">248</span> static final int CHECKSUM_SIZE = Bytes.SIZEOF_INT;<a name="line.248"></a> |
| <span class="sourceLineNo">249</span><a name="line.249"></a> |
| <span class="sourceLineNo">250</span> static final byte[] DUMMY_HEADER_NO_CHECKSUM =<a name="line.250"></a> |
| <span class="sourceLineNo">251</span> new byte[HConstants.HFILEBLOCK_HEADER_SIZE_NO_CHECKSUM];<a name="line.251"></a> |
| <span class="sourceLineNo">252</span><a name="line.252"></a> |
| <span class="sourceLineNo">253</span> /**<a name="line.253"></a> |
| <span class="sourceLineNo">254</span> * Used deserializing blocks from Cache. <code><a name="line.254"></a> |
| <span class="sourceLineNo">255</span> * ++++++++++++++<a name="line.255"></a> |
| <span class="sourceLineNo">256</span> * + HFileBlock +<a name="line.256"></a> |
| <span class="sourceLineNo">257</span> * ++++++++++++++<a name="line.257"></a> |
| <span class="sourceLineNo">258</span> * + Checksums + <= Optional<a name="line.258"></a> |
| <span class="sourceLineNo">259</span> * ++++++++++++++<a name="line.259"></a> |
| <span class="sourceLineNo">260</span> * + Metadata! + <= See note on BLOCK_METADATA_SPACE above.<a name="line.260"></a> |
| <span class="sourceLineNo">261</span> * ++++++++++++++<a name="line.261"></a> |
| <span class="sourceLineNo">262</span> * </code><a name="line.262"></a> |
| <span class="sourceLineNo">263</span> * @see #serialize(ByteBuffer, boolean)<a name="line.263"></a> |
| <span class="sourceLineNo">264</span> */<a name="line.264"></a> |
| <span class="sourceLineNo">265</span> public static final CacheableDeserializer<Cacheable> BLOCK_DESERIALIZER = new BlockDeserializer();<a name="line.265"></a> |
| <span class="sourceLineNo">266</span><a name="line.266"></a> |
| <span class="sourceLineNo">267</span> public static final class BlockDeserializer implements CacheableDeserializer<Cacheable> {<a name="line.267"></a> |
| <span class="sourceLineNo">268</span> private BlockDeserializer() {<a name="line.268"></a> |
| <span class="sourceLineNo">269</span> }<a name="line.269"></a> |
| <span class="sourceLineNo">270</span><a name="line.270"></a> |
| <span class="sourceLineNo">271</span> @Override<a name="line.271"></a> |
| <span class="sourceLineNo">272</span> public HFileBlock deserialize(ByteBuff buf, ByteBuffAllocator alloc) throws IOException {<a name="line.272"></a> |
| <span class="sourceLineNo">273</span> // The buf has the file block followed by block metadata.<a name="line.273"></a> |
| <span class="sourceLineNo">274</span> // Set limit to just before the BLOCK_METADATA_SPACE then rewind.<a name="line.274"></a> |
| <span class="sourceLineNo">275</span> buf.limit(buf.limit() - BLOCK_METADATA_SPACE).rewind();<a name="line.275"></a> |
| <span class="sourceLineNo">276</span> // Get a new buffer to pass the HFileBlock for it to 'own'.<a name="line.276"></a> |
| <span class="sourceLineNo">277</span> ByteBuff newByteBuff = buf.slice();<a name="line.277"></a> |
| <span class="sourceLineNo">278</span> // Read out the BLOCK_METADATA_SPACE content and shove into our HFileBlock.<a name="line.278"></a> |
| <span class="sourceLineNo">279</span> buf.position(buf.limit());<a name="line.279"></a> |
| <span class="sourceLineNo">280</span> buf.limit(buf.limit() + HFileBlock.BLOCK_METADATA_SPACE);<a name="line.280"></a> |
| <span class="sourceLineNo">281</span> boolean usesChecksum = buf.get() == (byte) 1;<a name="line.281"></a> |
| <span class="sourceLineNo">282</span> long offset = buf.getLong();<a name="line.282"></a> |
| <span class="sourceLineNo">283</span> int nextBlockOnDiskSize = buf.getInt();<a name="line.283"></a> |
| <span class="sourceLineNo">284</span> return createFromBuff(newByteBuff, usesChecksum, offset, nextBlockOnDiskSize, null, alloc);<a name="line.284"></a> |
| <span class="sourceLineNo">285</span> }<a name="line.285"></a> |
| <span class="sourceLineNo">286</span><a name="line.286"></a> |
| <span class="sourceLineNo">287</span> @Override<a name="line.287"></a> |
| <span class="sourceLineNo">288</span> public int getDeserializerIdentifier() {<a name="line.288"></a> |
| <span class="sourceLineNo">289</span> return DESERIALIZER_IDENTIFIER;<a name="line.289"></a> |
| <span class="sourceLineNo">290</span> }<a name="line.290"></a> |
| <span class="sourceLineNo">291</span> }<a name="line.291"></a> |
| <span class="sourceLineNo">292</span><a name="line.292"></a> |
| <span class="sourceLineNo">293</span> private static final int DESERIALIZER_IDENTIFIER;<a name="line.293"></a> |
| <span class="sourceLineNo">294</span> static {<a name="line.294"></a> |
| <span class="sourceLineNo">295</span> DESERIALIZER_IDENTIFIER =<a name="line.295"></a> |
| <span class="sourceLineNo">296</span> CacheableDeserializerIdManager.registerDeserializer(BLOCK_DESERIALIZER);<a name="line.296"></a> |
| <span class="sourceLineNo">297</span> }<a name="line.297"></a> |
| <span class="sourceLineNo">298</span><a name="line.298"></a> |
| <span class="sourceLineNo">299</span> /**<a name="line.299"></a> |
| <span class="sourceLineNo">300</span> * Creates a new {@link HFile} block from the given fields. This constructor is used only while<a name="line.300"></a> |
| <span class="sourceLineNo">301</span> * writing blocks and caching, and is sitting in a byte buffer and we want to stuff the block into<a name="line.301"></a> |
| <span class="sourceLineNo">302</span> * cache.<a name="line.302"></a> |
| <span class="sourceLineNo">303</span> * <p><a name="line.303"></a> |
| <span class="sourceLineNo">304</span> * TODO: The caller presumes no checksumming<a name="line.304"></a> |
| <span class="sourceLineNo">305</span> * <p><a name="line.305"></a> |
| <span class="sourceLineNo">306</span> * TODO: HFile block writer can also off-heap ?<a name="line.306"></a> |
| <span class="sourceLineNo">307</span> * </p><a name="line.307"></a> |
| <span class="sourceLineNo">308</span> * required of this block instance since going into cache; checksum already verified on underlying<a name="line.308"></a> |
| <span class="sourceLineNo">309</span> * block data pulled in from filesystem. Is that correct? What if cache is SSD?<a name="line.309"></a> |
| <span class="sourceLineNo">310</span> * @param blockType the type of this block, see {@link BlockType}<a name="line.310"></a> |
| <span class="sourceLineNo">311</span> * @param onDiskSizeWithoutHeader see {@link #onDiskSizeWithoutHeader}<a name="line.311"></a> |
| <span class="sourceLineNo">312</span> * @param uncompressedSizeWithoutHeader see {@link #uncompressedSizeWithoutHeader}<a name="line.312"></a> |
| <span class="sourceLineNo">313</span> * @param prevBlockOffset see {@link #prevBlockOffset}<a name="line.313"></a> |
| <span class="sourceLineNo">314</span> * @param buf block buffer with header<a name="line.314"></a> |
| <span class="sourceLineNo">315</span> * ({@link HConstants#HFILEBLOCK_HEADER_SIZE} bytes)<a name="line.315"></a> |
| <span class="sourceLineNo">316</span> * @param fillHeader when true, write the first 4 header fields into passed<a name="line.316"></a> |
| <span class="sourceLineNo">317</span> * buffer.<a name="line.317"></a> |
| <span class="sourceLineNo">318</span> * @param offset the file offset the block was read from<a name="line.318"></a> |
| <span class="sourceLineNo">319</span> * @param onDiskDataSizeWithHeader see {@link #onDiskDataSizeWithHeader}<a name="line.319"></a> |
| <span class="sourceLineNo">320</span> * @param fileContext HFile meta data<a name="line.320"></a> |
| <span class="sourceLineNo">321</span> */<a name="line.321"></a> |
| <span class="sourceLineNo">322</span> public HFileBlock(BlockType blockType, int onDiskSizeWithoutHeader,<a name="line.322"></a> |
| <span class="sourceLineNo">323</span> int uncompressedSizeWithoutHeader, long prevBlockOffset, ByteBuff buf, boolean fillHeader,<a name="line.323"></a> |
| <span class="sourceLineNo">324</span> long offset, int nextBlockOnDiskSize, int onDiskDataSizeWithHeader, HFileContext fileContext,<a name="line.324"></a> |
| <span class="sourceLineNo">325</span> ByteBuffAllocator allocator) {<a name="line.325"></a> |
| <span class="sourceLineNo">326</span> this.blockType = blockType;<a name="line.326"></a> |
| <span class="sourceLineNo">327</span> this.onDiskSizeWithoutHeader = onDiskSizeWithoutHeader;<a name="line.327"></a> |
| <span class="sourceLineNo">328</span> this.uncompressedSizeWithoutHeader = uncompressedSizeWithoutHeader;<a name="line.328"></a> |
| <span class="sourceLineNo">329</span> this.prevBlockOffset = prevBlockOffset;<a name="line.329"></a> |
| <span class="sourceLineNo">330</span> this.offset = offset;<a name="line.330"></a> |
| <span class="sourceLineNo">331</span> this.onDiskDataSizeWithHeader = onDiskDataSizeWithHeader;<a name="line.331"></a> |
| <span class="sourceLineNo">332</span> this.nextBlockOnDiskSize = nextBlockOnDiskSize;<a name="line.332"></a> |
| <span class="sourceLineNo">333</span> this.fileContext = fileContext;<a name="line.333"></a> |
| <span class="sourceLineNo">334</span> this.allocator = allocator;<a name="line.334"></a> |
| <span class="sourceLineNo">335</span> this.buf = buf;<a name="line.335"></a> |
| <span class="sourceLineNo">336</span> if (fillHeader) {<a name="line.336"></a> |
| <span class="sourceLineNo">337</span> overwriteHeader();<a name="line.337"></a> |
| <span class="sourceLineNo">338</span> }<a name="line.338"></a> |
| <span class="sourceLineNo">339</span> this.buf.rewind();<a name="line.339"></a> |
| <span class="sourceLineNo">340</span> }<a name="line.340"></a> |
| <span class="sourceLineNo">341</span><a name="line.341"></a> |
| <span class="sourceLineNo">342</span> /**<a name="line.342"></a> |
| <span class="sourceLineNo">343</span> * Creates a block from an existing buffer starting with a header. Rewinds and takes ownership of<a name="line.343"></a> |
| <span class="sourceLineNo">344</span> * the buffer. By definition of rewind, ignores the buffer position, but if you slice the buffer<a name="line.344"></a> |
| <span class="sourceLineNo">345</span> * beforehand, it will rewind to that point.<a name="line.345"></a> |
| <span class="sourceLineNo">346</span> * @param buf Has header, content, and trailing checksums if present.<a name="line.346"></a> |
| <span class="sourceLineNo">347</span> */<a name="line.347"></a> |
| <span class="sourceLineNo">348</span> static HFileBlock createFromBuff(ByteBuff buf, boolean usesHBaseChecksum, final long offset,<a name="line.348"></a> |
| <span class="sourceLineNo">349</span> final int nextBlockOnDiskSize, HFileContext fileContext, ByteBuffAllocator allocator)<a name="line.349"></a> |
| <span class="sourceLineNo">350</span> throws IOException {<a name="line.350"></a> |
| <span class="sourceLineNo">351</span> buf.rewind();<a name="line.351"></a> |
| <span class="sourceLineNo">352</span> final BlockType blockType = BlockType.read(buf);<a name="line.352"></a> |
| <span class="sourceLineNo">353</span> final int onDiskSizeWithoutHeader = buf.getInt(Header.ON_DISK_SIZE_WITHOUT_HEADER_INDEX);<a name="line.353"></a> |
| <span class="sourceLineNo">354</span> final int uncompressedSizeWithoutHeader =<a name="line.354"></a> |
| <span class="sourceLineNo">355</span> buf.getInt(Header.UNCOMPRESSED_SIZE_WITHOUT_HEADER_INDEX);<a name="line.355"></a> |
| <span class="sourceLineNo">356</span> final long prevBlockOffset = buf.getLong(Header.PREV_BLOCK_OFFSET_INDEX);<a name="line.356"></a> |
| <span class="sourceLineNo">357</span> // This constructor is called when we deserialize a block from cache and when we read a block in<a name="line.357"></a> |
| <span class="sourceLineNo">358</span> // from the fs. fileCache is null when deserialized from cache so need to make up one.<a name="line.358"></a> |
| <span class="sourceLineNo">359</span> HFileContextBuilder fileContextBuilder =<a name="line.359"></a> |
| <span class="sourceLineNo">360</span> fileContext != null ? new HFileContextBuilder(fileContext) : new HFileContextBuilder();<a name="line.360"></a> |
| <span class="sourceLineNo">361</span> fileContextBuilder.withHBaseCheckSum(usesHBaseChecksum);<a name="line.361"></a> |
| <span class="sourceLineNo">362</span> int onDiskDataSizeWithHeader;<a name="line.362"></a> |
| <span class="sourceLineNo">363</span> if (usesHBaseChecksum) {<a name="line.363"></a> |
| <span class="sourceLineNo">364</span> byte checksumType = buf.get(Header.CHECKSUM_TYPE_INDEX);<a name="line.364"></a> |
| <span class="sourceLineNo">365</span> int bytesPerChecksum = buf.getInt(Header.BYTES_PER_CHECKSUM_INDEX);<a name="line.365"></a> |
| <span class="sourceLineNo">366</span> onDiskDataSizeWithHeader = buf.getInt(Header.ON_DISK_DATA_SIZE_WITH_HEADER_INDEX);<a name="line.366"></a> |
| <span class="sourceLineNo">367</span> // Use the checksum type and bytes per checksum from header, not from fileContext.<a name="line.367"></a> |
| <span class="sourceLineNo">368</span> fileContextBuilder.withChecksumType(ChecksumType.codeToType(checksumType));<a name="line.368"></a> |
| <span class="sourceLineNo">369</span> fileContextBuilder.withBytesPerCheckSum(bytesPerChecksum);<a name="line.369"></a> |
| <span class="sourceLineNo">370</span> } else {<a name="line.370"></a> |
| <span class="sourceLineNo">371</span> fileContextBuilder.withChecksumType(ChecksumType.NULL);<a name="line.371"></a> |
| <span class="sourceLineNo">372</span> fileContextBuilder.withBytesPerCheckSum(0);<a name="line.372"></a> |
| <span class="sourceLineNo">373</span> // Need to fix onDiskDataSizeWithHeader; there are not checksums after-block-data<a name="line.373"></a> |
| <span class="sourceLineNo">374</span> onDiskDataSizeWithHeader = onDiskSizeWithoutHeader + headerSize(usesHBaseChecksum);<a name="line.374"></a> |
| <span class="sourceLineNo">375</span> }<a name="line.375"></a> |
| <span class="sourceLineNo">376</span> fileContext = fileContextBuilder.build();<a name="line.376"></a> |
| <span class="sourceLineNo">377</span> assert usesHBaseChecksum == fileContext.isUseHBaseChecksum();<a name="line.377"></a> |
| <span class="sourceLineNo">378</span> return new HFileBlockBuilder().withBlockType(blockType)<a name="line.378"></a> |
| <span class="sourceLineNo">379</span> .withOnDiskSizeWithoutHeader(onDiskSizeWithoutHeader)<a name="line.379"></a> |
| <span class="sourceLineNo">380</span> .withUncompressedSizeWithoutHeader(uncompressedSizeWithoutHeader)<a name="line.380"></a> |
| <span class="sourceLineNo">381</span> .withPrevBlockOffset(prevBlockOffset).withOffset(offset)<a name="line.381"></a> |
| <span class="sourceLineNo">382</span> .withOnDiskDataSizeWithHeader(onDiskDataSizeWithHeader)<a name="line.382"></a> |
| <span class="sourceLineNo">383</span> .withNextBlockOnDiskSize(nextBlockOnDiskSize).withHFileContext(fileContext)<a name="line.383"></a> |
| <span class="sourceLineNo">384</span> .withByteBuffAllocator(allocator).withByteBuff(buf.rewind()).withShared(!buf.hasArray())<a name="line.384"></a> |
| <span class="sourceLineNo">385</span> .build();<a name="line.385"></a> |
| <span class="sourceLineNo">386</span> }<a name="line.386"></a> |
| <span class="sourceLineNo">387</span><a name="line.387"></a> |
| <span class="sourceLineNo">388</span> /**<a name="line.388"></a> |
| <span class="sourceLineNo">389</span> * Parse total on disk size including header and checksum.<a name="line.389"></a> |
| <span class="sourceLineNo">390</span> * @param headerBuf Header ByteBuffer. Presumed exact size of header.<a name="line.390"></a> |
| <span class="sourceLineNo">391</span> * @param verifyChecksum true if checksum verification is in use.<a name="line.391"></a> |
| <span class="sourceLineNo">392</span> * @return Size of the block with header included.<a name="line.392"></a> |
| <span class="sourceLineNo">393</span> */<a name="line.393"></a> |
| <span class="sourceLineNo">394</span> private static int getOnDiskSizeWithHeader(final ByteBuff headerBuf, boolean verifyChecksum) {<a name="line.394"></a> |
| <span class="sourceLineNo">395</span> return headerBuf.getInt(Header.ON_DISK_SIZE_WITHOUT_HEADER_INDEX) + headerSize(verifyChecksum);<a name="line.395"></a> |
| <span class="sourceLineNo">396</span> }<a name="line.396"></a> |
| <span class="sourceLineNo">397</span><a name="line.397"></a> |
| <span class="sourceLineNo">398</span> /**<a name="line.398"></a> |
| <span class="sourceLineNo">399</span> * @return the on-disk size of the next block (including the header size and any checksums if<a name="line.399"></a> |
| <span class="sourceLineNo">400</span> * present) read by peeking into the next block's header; use as a hint when doing a read<a name="line.400"></a> |
| <span class="sourceLineNo">401</span> * of the next block when scanning or running over a file.<a name="line.401"></a> |
| <span class="sourceLineNo">402</span> */<a name="line.402"></a> |
| <span class="sourceLineNo">403</span> int getNextBlockOnDiskSize() {<a name="line.403"></a> |
| <span class="sourceLineNo">404</span> return nextBlockOnDiskSize;<a name="line.404"></a> |
| <span class="sourceLineNo">405</span> }<a name="line.405"></a> |
| <span class="sourceLineNo">406</span><a name="line.406"></a> |
| <span class="sourceLineNo">407</span> @Override<a name="line.407"></a> |
| <span class="sourceLineNo">408</span> public BlockType getBlockType() {<a name="line.408"></a> |
| <span class="sourceLineNo">409</span> return blockType;<a name="line.409"></a> |
| <span class="sourceLineNo">410</span> }<a name="line.410"></a> |
| <span class="sourceLineNo">411</span><a name="line.411"></a> |
| <span class="sourceLineNo">412</span> @Override<a name="line.412"></a> |
| <span class="sourceLineNo">413</span> public int refCnt() {<a name="line.413"></a> |
| <span class="sourceLineNo">414</span> return buf.refCnt();<a name="line.414"></a> |
| <span class="sourceLineNo">415</span> }<a name="line.415"></a> |
| <span class="sourceLineNo">416</span><a name="line.416"></a> |
| <span class="sourceLineNo">417</span> @Override<a name="line.417"></a> |
| <span class="sourceLineNo">418</span> public HFileBlock retain() {<a name="line.418"></a> |
| <span class="sourceLineNo">419</span> buf.retain();<a name="line.419"></a> |
| <span class="sourceLineNo">420</span> return this;<a name="line.420"></a> |
| <span class="sourceLineNo">421</span> }<a name="line.421"></a> |
| <span class="sourceLineNo">422</span><a name="line.422"></a> |
| <span class="sourceLineNo">423</span> /**<a name="line.423"></a> |
| <span class="sourceLineNo">424</span> * Call {@link ByteBuff#release()} to decrease the reference count, if no other reference, it will<a name="line.424"></a> |
| <span class="sourceLineNo">425</span> * return back the {@link ByteBuffer} to {@link org.apache.hadoop.hbase.io.ByteBuffAllocator}<a name="line.425"></a> |
| <span class="sourceLineNo">426</span> */<a name="line.426"></a> |
| <span class="sourceLineNo">427</span> @Override<a name="line.427"></a> |
| <span class="sourceLineNo">428</span> public boolean release() {<a name="line.428"></a> |
| <span class="sourceLineNo">429</span> return buf.release();<a name="line.429"></a> |
| <span class="sourceLineNo">430</span> }<a name="line.430"></a> |
| <span class="sourceLineNo">431</span><a name="line.431"></a> |
| <span class="sourceLineNo">432</span> /**<a name="line.432"></a> |
| <span class="sourceLineNo">433</span> * Calling this method in strategic locations where HFileBlocks are referenced may help diagnose<a name="line.433"></a> |
| <span class="sourceLineNo">434</span> * potential buffer leaks. We pass the block itself as a default hint, but one can use<a name="line.434"></a> |
| <span class="sourceLineNo">435</span> * {@link #touch(Object)} to pass their own hint as well.<a name="line.435"></a> |
| <span class="sourceLineNo">436</span> */<a name="line.436"></a> |
| <span class="sourceLineNo">437</span> @Override<a name="line.437"></a> |
| <span class="sourceLineNo">438</span> public HFileBlock touch() {<a name="line.438"></a> |
| <span class="sourceLineNo">439</span> return touch(this);<a name="line.439"></a> |
| <span class="sourceLineNo">440</span> }<a name="line.440"></a> |
| <span class="sourceLineNo">441</span><a name="line.441"></a> |
| <span class="sourceLineNo">442</span> @Override<a name="line.442"></a> |
| <span class="sourceLineNo">443</span> public HFileBlock touch(Object hint) {<a name="line.443"></a> |
| <span class="sourceLineNo">444</span> buf.touch(hint);<a name="line.444"></a> |
| <span class="sourceLineNo">445</span> return this;<a name="line.445"></a> |
| <span class="sourceLineNo">446</span> }<a name="line.446"></a> |
| <span class="sourceLineNo">447</span><a name="line.447"></a> |
| <span class="sourceLineNo">448</span> /** Returns get data block encoding id that was used to encode this block */<a name="line.448"></a> |
| <span class="sourceLineNo">449</span> short getDataBlockEncodingId() {<a name="line.449"></a> |
| <span class="sourceLineNo">450</span> if (blockType != BlockType.ENCODED_DATA) {<a name="line.450"></a> |
| <span class="sourceLineNo">451</span> throw new IllegalArgumentException("Querying encoder ID of a block " + "of type other than "<a name="line.451"></a> |
| <span class="sourceLineNo">452</span> + BlockType.ENCODED_DATA + ": " + blockType);<a name="line.452"></a> |
| <span class="sourceLineNo">453</span> }<a name="line.453"></a> |
| <span class="sourceLineNo">454</span> return buf.getShort(headerSize());<a name="line.454"></a> |
| <span class="sourceLineNo">455</span> }<a name="line.455"></a> |
| <span class="sourceLineNo">456</span><a name="line.456"></a> |
| <span class="sourceLineNo">457</span> /** Returns the on-disk size of header + data part + checksum. */<a name="line.457"></a> |
| <span class="sourceLineNo">458</span> public int getOnDiskSizeWithHeader() {<a name="line.458"></a> |
| <span class="sourceLineNo">459</span> return onDiskSizeWithoutHeader + headerSize();<a name="line.459"></a> |
| <span class="sourceLineNo">460</span> }<a name="line.460"></a> |
| <span class="sourceLineNo">461</span><a name="line.461"></a> |
| <span class="sourceLineNo">462</span> /** Returns the on-disk size of the data part + checksum (header excluded). */<a name="line.462"></a> |
| <span class="sourceLineNo">463</span> int getOnDiskSizeWithoutHeader() {<a name="line.463"></a> |
| <span class="sourceLineNo">464</span> return onDiskSizeWithoutHeader;<a name="line.464"></a> |
| <span class="sourceLineNo">465</span> }<a name="line.465"></a> |
| <span class="sourceLineNo">466</span><a name="line.466"></a> |
| <span class="sourceLineNo">467</span> /** Returns the uncompressed size of data part (header and checksum excluded). */<a name="line.467"></a> |
| <span class="sourceLineNo">468</span> public int getUncompressedSizeWithoutHeader() {<a name="line.468"></a> |
| <span class="sourceLineNo">469</span> return uncompressedSizeWithoutHeader;<a name="line.469"></a> |
| <span class="sourceLineNo">470</span> }<a name="line.470"></a> |
| <span class="sourceLineNo">471</span><a name="line.471"></a> |
| <span class="sourceLineNo">472</span> /** Returns the offset of the previous block of the same type in the file, or -1 if unknown */<a name="line.472"></a> |
| <span class="sourceLineNo">473</span> long getPrevBlockOffset() {<a name="line.473"></a> |
| <span class="sourceLineNo">474</span> return prevBlockOffset;<a name="line.474"></a> |
| <span class="sourceLineNo">475</span> }<a name="line.475"></a> |
| <span class="sourceLineNo">476</span><a name="line.476"></a> |
| <span class="sourceLineNo">477</span> /**<a name="line.477"></a> |
| <span class="sourceLineNo">478</span> * Rewinds {@code buf} and writes first 4 header fields. {@code buf} position is modified as<a name="line.478"></a> |
| <span class="sourceLineNo">479</span> * side-effect.<a name="line.479"></a> |
| <span class="sourceLineNo">480</span> */<a name="line.480"></a> |
| <span class="sourceLineNo">481</span> private void overwriteHeader() {<a name="line.481"></a> |
| <span class="sourceLineNo">482</span> buf.rewind();<a name="line.482"></a> |
| <span class="sourceLineNo">483</span> blockType.write(buf);<a name="line.483"></a> |
| <span class="sourceLineNo">484</span> buf.putInt(onDiskSizeWithoutHeader);<a name="line.484"></a> |
| <span class="sourceLineNo">485</span> buf.putInt(uncompressedSizeWithoutHeader);<a name="line.485"></a> |
| <span class="sourceLineNo">486</span> buf.putLong(prevBlockOffset);<a name="line.486"></a> |
| <span class="sourceLineNo">487</span> if (this.fileContext.isUseHBaseChecksum()) {<a name="line.487"></a> |
| <span class="sourceLineNo">488</span> buf.put(fileContext.getChecksumType().getCode());<a name="line.488"></a> |
| <span class="sourceLineNo">489</span> buf.putInt(fileContext.getBytesPerChecksum());<a name="line.489"></a> |
| <span class="sourceLineNo">490</span> buf.putInt(onDiskDataSizeWithHeader);<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><a name="line.493"></a> |
| <span class="sourceLineNo">494</span> /**<a name="line.494"></a> |
| <span class="sourceLineNo">495</span> * Returns a buffer that does not include the header and checksum.<a name="line.495"></a> |
| <span class="sourceLineNo">496</span> * @return the buffer with header skipped and checksum omitted.<a name="line.496"></a> |
| <span class="sourceLineNo">497</span> */<a name="line.497"></a> |
| <span class="sourceLineNo">498</span> public ByteBuff getBufferWithoutHeader() {<a name="line.498"></a> |
| <span class="sourceLineNo">499</span> ByteBuff dup = getBufferReadOnly();<a name="line.499"></a> |
| <span class="sourceLineNo">500</span> return dup.position(headerSize()).slice();<a name="line.500"></a> |
| <span class="sourceLineNo">501</span> }<a name="line.501"></a> |
| <span class="sourceLineNo">502</span><a name="line.502"></a> |
| <span class="sourceLineNo">503</span> /**<a name="line.503"></a> |
| <span class="sourceLineNo">504</span> * Returns a read-only duplicate of the buffer this block stores internally ready to be read.<a name="line.504"></a> |
| <span class="sourceLineNo">505</span> * Clients must not modify the buffer object though they may set position and limit on the<a name="line.505"></a> |
| <span class="sourceLineNo">506</span> * returned buffer since we pass back a duplicate. This method has to be public because it is used<a name="line.506"></a> |
| <span class="sourceLineNo">507</span> * in {@link CompoundBloomFilter} to avoid object creation on every Bloom filter lookup, but has<a name="line.507"></a> |
| <span class="sourceLineNo">508</span> * to be used with caution. Buffer holds header, block content, and any follow-on checksums if<a name="line.508"></a> |
| <span class="sourceLineNo">509</span> * present.<a name="line.509"></a> |
| <span class="sourceLineNo">510</span> * @return the buffer of this block for read-only operations<a name="line.510"></a> |
| <span class="sourceLineNo">511</span> */<a name="line.511"></a> |
| <span class="sourceLineNo">512</span> public ByteBuff getBufferReadOnly() {<a name="line.512"></a> |
| <span class="sourceLineNo">513</span> // TODO: ByteBuf does not support asReadOnlyBuffer(). Fix.<a name="line.513"></a> |
| <span class="sourceLineNo">514</span> ByteBuff dup = this.buf.duplicate();<a name="line.514"></a> |
| <span class="sourceLineNo">515</span> assert dup.position() == 0;<a name="line.515"></a> |
| <span class="sourceLineNo">516</span> return dup;<a name="line.516"></a> |
| <span class="sourceLineNo">517</span> }<a name="line.517"></a> |
| <span class="sourceLineNo">518</span><a name="line.518"></a> |
| <span class="sourceLineNo">519</span> public ByteBuffAllocator getByteBuffAllocator() {<a name="line.519"></a> |
| <span class="sourceLineNo">520</span> return this.allocator;<a name="line.520"></a> |
| <span class="sourceLineNo">521</span> }<a name="line.521"></a> |
| <span class="sourceLineNo">522</span><a name="line.522"></a> |
| <span class="sourceLineNo">523</span> private void sanityCheckAssertion(long valueFromBuf, long valueFromField, String fieldName)<a name="line.523"></a> |
| <span class="sourceLineNo">524</span> throws IOException {<a name="line.524"></a> |
| <span class="sourceLineNo">525</span> if (valueFromBuf != valueFromField) {<a name="line.525"></a> |
| <span class="sourceLineNo">526</span> throw new AssertionError(fieldName + " in the buffer (" + valueFromBuf<a name="line.526"></a> |
| <span class="sourceLineNo">527</span> + ") is different from that in the field (" + valueFromField + ")");<a name="line.527"></a> |
| <span class="sourceLineNo">528</span> }<a name="line.528"></a> |
| <span class="sourceLineNo">529</span> }<a name="line.529"></a> |
| <span class="sourceLineNo">530</span><a name="line.530"></a> |
| <span class="sourceLineNo">531</span> private void sanityCheckAssertion(BlockType valueFromBuf, BlockType valueFromField)<a name="line.531"></a> |
| <span class="sourceLineNo">532</span> throws IOException {<a name="line.532"></a> |
| <span class="sourceLineNo">533</span> if (valueFromBuf != valueFromField) {<a name="line.533"></a> |
| <span class="sourceLineNo">534</span> throw new IOException("Block type stored in the buffer: " + valueFromBuf<a name="line.534"></a> |
| <span class="sourceLineNo">535</span> + ", block type field: " + valueFromField);<a name="line.535"></a> |
| <span class="sourceLineNo">536</span> }<a name="line.536"></a> |
| <span class="sourceLineNo">537</span> }<a name="line.537"></a> |
| <span class="sourceLineNo">538</span><a name="line.538"></a> |
| <span class="sourceLineNo">539</span> /**<a name="line.539"></a> |
| <span class="sourceLineNo">540</span> * Checks if the block is internally consistent, i.e. the first<a name="line.540"></a> |
| <span class="sourceLineNo">541</span> * {@link HConstants#HFILEBLOCK_HEADER_SIZE} bytes of the buffer contain a valid header consistent<a name="line.541"></a> |
| <span class="sourceLineNo">542</span> * with the fields. Assumes a packed block structure. This function is primary for testing and<a name="line.542"></a> |
| <span class="sourceLineNo">543</span> * debugging, and is not thread-safe, because it alters the internal buffer pointer. Used by tests<a name="line.543"></a> |
| <span class="sourceLineNo">544</span> * only.<a name="line.544"></a> |
| <span class="sourceLineNo">545</span> */<a name="line.545"></a> |
| <span class="sourceLineNo">546</span> void sanityCheck() throws IOException {<a name="line.546"></a> |
| <span class="sourceLineNo">547</span> // Duplicate so no side-effects<a name="line.547"></a> |
| <span class="sourceLineNo">548</span> ByteBuff dup = this.buf.duplicate().rewind();<a name="line.548"></a> |
| <span class="sourceLineNo">549</span> sanityCheckAssertion(BlockType.read(dup), blockType);<a name="line.549"></a> |
| <span class="sourceLineNo">550</span><a name="line.550"></a> |
| <span class="sourceLineNo">551</span> sanityCheckAssertion(dup.getInt(), onDiskSizeWithoutHeader, "onDiskSizeWithoutHeader");<a name="line.551"></a> |
| <span class="sourceLineNo">552</span><a name="line.552"></a> |
| <span class="sourceLineNo">553</span> sanityCheckAssertion(dup.getInt(), uncompressedSizeWithoutHeader,<a name="line.553"></a> |
| <span class="sourceLineNo">554</span> "uncompressedSizeWithoutHeader");<a name="line.554"></a> |
| <span class="sourceLineNo">555</span><a name="line.555"></a> |
| <span class="sourceLineNo">556</span> sanityCheckAssertion(dup.getLong(), prevBlockOffset, "prevBlockOffset");<a name="line.556"></a> |
| <span class="sourceLineNo">557</span> if (this.fileContext.isUseHBaseChecksum()) {<a name="line.557"></a> |
| <span class="sourceLineNo">558</span> sanityCheckAssertion(dup.get(), this.fileContext.getChecksumType().getCode(), "checksumType");<a name="line.558"></a> |
| <span class="sourceLineNo">559</span> sanityCheckAssertion(dup.getInt(), this.fileContext.getBytesPerChecksum(),<a name="line.559"></a> |
| <span class="sourceLineNo">560</span> "bytesPerChecksum");<a name="line.560"></a> |
| <span class="sourceLineNo">561</span> sanityCheckAssertion(dup.getInt(), onDiskDataSizeWithHeader, "onDiskDataSizeWithHeader");<a name="line.561"></a> |
| <span class="sourceLineNo">562</span> }<a name="line.562"></a> |
| <span class="sourceLineNo">563</span><a name="line.563"></a> |
| <span class="sourceLineNo">564</span> if (dup.limit() != onDiskDataSizeWithHeader) {<a name="line.564"></a> |
| <span class="sourceLineNo">565</span> throw new AssertionError(<a name="line.565"></a> |
| <span class="sourceLineNo">566</span> "Expected limit " + onDiskDataSizeWithHeader + ", got " + dup.limit());<a name="line.566"></a> |
| <span class="sourceLineNo">567</span> }<a name="line.567"></a> |
| <span class="sourceLineNo">568</span><a name="line.568"></a> |
| <span class="sourceLineNo">569</span> // We might optionally allocate HFILEBLOCK_HEADER_SIZE more bytes to read the next<a name="line.569"></a> |
| <span class="sourceLineNo">570</span> // block's header, so there are two sensible values for buffer capacity.<a name="line.570"></a> |
| <span class="sourceLineNo">571</span> int hdrSize = headerSize();<a name="line.571"></a> |
| <span class="sourceLineNo">572</span> dup.rewind();<a name="line.572"></a> |
| <span class="sourceLineNo">573</span> if (<a name="line.573"></a> |
| <span class="sourceLineNo">574</span> dup.remaining() != onDiskDataSizeWithHeader<a name="line.574"></a> |
| <span class="sourceLineNo">575</span> && dup.remaining() != onDiskDataSizeWithHeader + hdrSize<a name="line.575"></a> |
| <span class="sourceLineNo">576</span> ) {<a name="line.576"></a> |
| <span class="sourceLineNo">577</span> throw new AssertionError("Invalid buffer capacity: " + dup.remaining() + ", expected "<a name="line.577"></a> |
| <span class="sourceLineNo">578</span> + onDiskDataSizeWithHeader + " or " + (onDiskDataSizeWithHeader + hdrSize));<a name="line.578"></a> |
| <span class="sourceLineNo">579</span> }<a name="line.579"></a> |
| <span class="sourceLineNo">580</span> }<a name="line.580"></a> |
| <span class="sourceLineNo">581</span><a name="line.581"></a> |
| <span class="sourceLineNo">582</span> @Override<a name="line.582"></a> |
| <span class="sourceLineNo">583</span> public String toString() {<a name="line.583"></a> |
| <span class="sourceLineNo">584</span> StringBuilder sb = new StringBuilder().append("[").append("blockType=").append(blockType)<a name="line.584"></a> |
| <span class="sourceLineNo">585</span> .append(", fileOffset=").append(offset).append(", headerSize=").append(headerSize())<a name="line.585"></a> |
| <span class="sourceLineNo">586</span> .append(", onDiskSizeWithoutHeader=").append(onDiskSizeWithoutHeader)<a name="line.586"></a> |
| <span class="sourceLineNo">587</span> .append(", uncompressedSizeWithoutHeader=").append(uncompressedSizeWithoutHeader)<a name="line.587"></a> |
| <span class="sourceLineNo">588</span> .append(", prevBlockOffset=").append(prevBlockOffset).append(", isUseHBaseChecksum=")<a name="line.588"></a> |
| <span class="sourceLineNo">589</span> .append(fileContext.isUseHBaseChecksum());<a name="line.589"></a> |
| <span class="sourceLineNo">590</span> if (fileContext.isUseHBaseChecksum()) {<a name="line.590"></a> |
| <span class="sourceLineNo">591</span> sb.append(", checksumType=").append(ChecksumType.codeToType(this.buf.get(24)))<a name="line.591"></a> |
| <span class="sourceLineNo">592</span> .append(", bytesPerChecksum=").append(this.buf.getInt(24 + 1))<a name="line.592"></a> |
| <span class="sourceLineNo">593</span> .append(", onDiskDataSizeWithHeader=").append(onDiskDataSizeWithHeader);<a name="line.593"></a> |
| <span class="sourceLineNo">594</span> } else {<a name="line.594"></a> |
| <span class="sourceLineNo">595</span> sb.append(", onDiskDataSizeWithHeader=").append(onDiskDataSizeWithHeader).append("(")<a name="line.595"></a> |
| <span class="sourceLineNo">596</span> .append(onDiskSizeWithoutHeader).append("+")<a name="line.596"></a> |
| <span class="sourceLineNo">597</span> .append(HConstants.HFILEBLOCK_HEADER_SIZE_NO_CHECKSUM).append(")");<a name="line.597"></a> |
| <span class="sourceLineNo">598</span> }<a name="line.598"></a> |
| <span class="sourceLineNo">599</span> String dataBegin;<a name="line.599"></a> |
| <span class="sourceLineNo">600</span> if (buf.hasArray()) {<a name="line.600"></a> |
| <span class="sourceLineNo">601</span> dataBegin = Bytes.toStringBinary(buf.array(), buf.arrayOffset() + headerSize(),<a name="line.601"></a> |
| <span class="sourceLineNo">602</span> Math.min(32, buf.limit() - buf.arrayOffset() - headerSize()));<a name="line.602"></a> |
| <span class="sourceLineNo">603</span> } else {<a name="line.603"></a> |
| <span class="sourceLineNo">604</span> ByteBuff bufWithoutHeader = getBufferWithoutHeader();<a name="line.604"></a> |
| <span class="sourceLineNo">605</span> byte[] dataBeginBytes =<a name="line.605"></a> |
| <span class="sourceLineNo">606</span> new byte[Math.min(32, bufWithoutHeader.limit() - bufWithoutHeader.position())];<a name="line.606"></a> |
| <span class="sourceLineNo">607</span> bufWithoutHeader.get(dataBeginBytes);<a name="line.607"></a> |
| <span class="sourceLineNo">608</span> dataBegin = Bytes.toStringBinary(dataBeginBytes);<a name="line.608"></a> |
| <span class="sourceLineNo">609</span> }<a name="line.609"></a> |
| <span class="sourceLineNo">610</span> sb.append(", getOnDiskSizeWithHeader=").append(getOnDiskSizeWithHeader())<a name="line.610"></a> |
| <span class="sourceLineNo">611</span> .append(", totalChecksumBytes=").append(totalChecksumBytes()).append(", isUnpacked=")<a name="line.611"></a> |
| <span class="sourceLineNo">612</span> .append(isUnpacked()).append(", buf=[").append(buf).append("]").append(", dataBeginsWith=")<a name="line.612"></a> |
| <span class="sourceLineNo">613</span> .append(dataBegin).append(", fileContext=").append(fileContext)<a name="line.613"></a> |
| <span class="sourceLineNo">614</span> .append(", nextBlockOnDiskSize=").append(nextBlockOnDiskSize).append("]");<a name="line.614"></a> |
| <span class="sourceLineNo">615</span> return sb.toString();<a name="line.615"></a> |
| <span class="sourceLineNo">616</span> }<a name="line.616"></a> |
| <span class="sourceLineNo">617</span><a name="line.617"></a> |
| <span class="sourceLineNo">618</span> /**<a name="line.618"></a> |
| <span class="sourceLineNo">619</span> * Retrieves the decompressed/decrypted view of this block. An encoded block remains in its<a name="line.619"></a> |
| <span class="sourceLineNo">620</span> * encoded structure. Internal structures are shared between instances where applicable.<a name="line.620"></a> |
| <span class="sourceLineNo">621</span> */<a name="line.621"></a> |
| <span class="sourceLineNo">622</span> HFileBlock unpack(HFileContext fileContext, FSReader reader) throws IOException {<a name="line.622"></a> |
| <span class="sourceLineNo">623</span> if (!fileContext.isCompressedOrEncrypted()) {<a name="line.623"></a> |
| <span class="sourceLineNo">624</span> // TODO: cannot use our own fileContext here because HFileBlock(ByteBuffer, boolean),<a name="line.624"></a> |
| <span class="sourceLineNo">625</span> // which is used for block serialization to L2 cache, does not preserve encoding and<a name="line.625"></a> |
| <span class="sourceLineNo">626</span> // encryption details.<a name="line.626"></a> |
| <span class="sourceLineNo">627</span> return this;<a name="line.627"></a> |
| <span class="sourceLineNo">628</span> }<a name="line.628"></a> |
| <span class="sourceLineNo">629</span><a name="line.629"></a> |
| <span class="sourceLineNo">630</span> ByteBuff newBuf = allocateBufferForUnpacking(); // allocates space for the decompressed block<a name="line.630"></a> |
| <span class="sourceLineNo">631</span> HFileBlock unpacked = shallowClone(this, newBuf);<a name="line.631"></a> |
| <span class="sourceLineNo">632</span><a name="line.632"></a> |
| <span class="sourceLineNo">633</span> boolean succ = false;<a name="line.633"></a> |
| <span class="sourceLineNo">634</span> final Context context =<a name="line.634"></a> |
| <span class="sourceLineNo">635</span> Context.current().with(CONTEXT_KEY, new HFileContextAttributesBuilderConsumer(fileContext));<a name="line.635"></a> |
| <span class="sourceLineNo">636</span> try (Scope ignored = context.makeCurrent()) {<a name="line.636"></a> |
| <span class="sourceLineNo">637</span> HFileBlockDecodingContext ctx = blockType == BlockType.ENCODED_DATA<a name="line.637"></a> |
| <span class="sourceLineNo">638</span> ? reader.getBlockDecodingContext()<a name="line.638"></a> |
| <span class="sourceLineNo">639</span> : reader.getDefaultBlockDecodingContext();<a name="line.639"></a> |
| <span class="sourceLineNo">640</span> // Create a duplicated buffer without the header part.<a name="line.640"></a> |
| <span class="sourceLineNo">641</span> int headerSize = this.headerSize();<a name="line.641"></a> |
| <span class="sourceLineNo">642</span> ByteBuff dup = this.buf.duplicate();<a name="line.642"></a> |
| <span class="sourceLineNo">643</span> dup.position(headerSize);<a name="line.643"></a> |
| <span class="sourceLineNo">644</span> dup = dup.slice();<a name="line.644"></a> |
| <span class="sourceLineNo">645</span> // Decode the dup into unpacked#buf<a name="line.645"></a> |
| <span class="sourceLineNo">646</span> ctx.prepareDecoding(unpacked.getOnDiskDataSizeWithHeader() - headerSize,<a name="line.646"></a> |
| <span class="sourceLineNo">647</span> unpacked.getUncompressedSizeWithoutHeader(), unpacked.getBufferWithoutHeader(), dup);<a name="line.647"></a> |
| <span class="sourceLineNo">648</span> succ = true;<a name="line.648"></a> |
| <span class="sourceLineNo">649</span> return unpacked;<a name="line.649"></a> |
| <span class="sourceLineNo">650</span> } finally {<a name="line.650"></a> |
| <span class="sourceLineNo">651</span> if (!succ) {<a name="line.651"></a> |
| <span class="sourceLineNo">652</span> unpacked.release();<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> * Always allocates a new buffer of the correct size. Copies header bytes from the existing<a name="line.658"></a> |
| <span class="sourceLineNo">659</span> * buffer. Does not change header fields. Reserve room to keep checksum bytes too.<a name="line.659"></a> |
| <span class="sourceLineNo">660</span> */<a name="line.660"></a> |
| <span class="sourceLineNo">661</span> private ByteBuff allocateBufferForUnpacking() {<a name="line.661"></a> |
| <span class="sourceLineNo">662</span> int headerSize = headerSize();<a name="line.662"></a> |
| <span class="sourceLineNo">663</span> int capacityNeeded = headerSize + uncompressedSizeWithoutHeader;<a name="line.663"></a> |
| <span class="sourceLineNo">664</span><a name="line.664"></a> |
| <span class="sourceLineNo">665</span> ByteBuff source = buf.duplicate();<a name="line.665"></a> |
| <span class="sourceLineNo">666</span> ByteBuff newBuf = allocator.allocate(capacityNeeded);<a name="line.666"></a> |
| <span class="sourceLineNo">667</span><a name="line.667"></a> |
| <span class="sourceLineNo">668</span> // Copy header bytes into newBuf.<a name="line.668"></a> |
| <span class="sourceLineNo">669</span> source.position(0);<a name="line.669"></a> |
| <span class="sourceLineNo">670</span> newBuf.put(0, source, 0, headerSize);<a name="line.670"></a> |
| <span class="sourceLineNo">671</span><a name="line.671"></a> |
| <span class="sourceLineNo">672</span> // set limit to exclude next block's header<a name="line.672"></a> |
| <span class="sourceLineNo">673</span> newBuf.limit(capacityNeeded);<a name="line.673"></a> |
| <span class="sourceLineNo">674</span> return newBuf;<a name="line.674"></a> |
| <span class="sourceLineNo">675</span> }<a name="line.675"></a> |
| <span class="sourceLineNo">676</span><a name="line.676"></a> |
| <span class="sourceLineNo">677</span> /**<a name="line.677"></a> |
| <span class="sourceLineNo">678</span> * Return true when this block's buffer has been unpacked, false otherwise. Note this is a<a name="line.678"></a> |
| <span class="sourceLineNo">679</span> * calculated heuristic, not tracked attribute of the block.<a name="line.679"></a> |
| <span class="sourceLineNo">680</span> */<a name="line.680"></a> |
| <span class="sourceLineNo">681</span> public boolean isUnpacked() {<a name="line.681"></a> |
| <span class="sourceLineNo">682</span> final int headerSize = headerSize();<a name="line.682"></a> |
| <span class="sourceLineNo">683</span> final int expectedCapacity = headerSize + uncompressedSizeWithoutHeader;<a name="line.683"></a> |
| <span class="sourceLineNo">684</span> final int bufCapacity = buf.remaining();<a name="line.684"></a> |
| <span class="sourceLineNo">685</span> return bufCapacity == expectedCapacity || bufCapacity == expectedCapacity + headerSize;<a name="line.685"></a> |
| <span class="sourceLineNo">686</span> }<a name="line.686"></a> |
| <span class="sourceLineNo">687</span><a name="line.687"></a> |
| <span class="sourceLineNo">688</span> /**<a name="line.688"></a> |
| <span class="sourceLineNo">689</span> * Cannot be {@link #UNSET}. Must be a legitimate value. Used re-making the {@link BlockCacheKey}<a name="line.689"></a> |
| <span class="sourceLineNo">690</span> * when block is returned to the cache.<a name="line.690"></a> |
| <span class="sourceLineNo">691</span> * @return the offset of this block in the file it was read from<a name="line.691"></a> |
| <span class="sourceLineNo">692</span> */<a name="line.692"></a> |
| <span class="sourceLineNo">693</span> long getOffset() {<a name="line.693"></a> |
| <span class="sourceLineNo">694</span> if (offset < 0) {<a name="line.694"></a> |
| <span class="sourceLineNo">695</span> throw new IllegalStateException("HFile block offset not initialized properly");<a name="line.695"></a> |
| <span class="sourceLineNo">696</span> }<a name="line.696"></a> |
| <span class="sourceLineNo">697</span> return offset;<a name="line.697"></a> |
| <span class="sourceLineNo">698</span> }<a name="line.698"></a> |
| <span class="sourceLineNo">699</span><a name="line.699"></a> |
| <span class="sourceLineNo">700</span> /** Returns a byte stream reading the data + checksum of this block */<a name="line.700"></a> |
| <span class="sourceLineNo">701</span> DataInputStream getByteStream() {<a name="line.701"></a> |
| <span class="sourceLineNo">702</span> ByteBuff dup = this.buf.duplicate();<a name="line.702"></a> |
| <span class="sourceLineNo">703</span> dup.position(this.headerSize());<a name="line.703"></a> |
| <span class="sourceLineNo">704</span> return new DataInputStream(new ByteBuffInputStream(dup));<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> @Override<a name="line.707"></a> |
| <span class="sourceLineNo">708</span> public long heapSize() {<a name="line.708"></a> |
| <span class="sourceLineNo">709</span> long size = FIXED_OVERHEAD;<a name="line.709"></a> |
| <span class="sourceLineNo">710</span> size += fileContext.heapSize();<a name="line.710"></a> |
| <span class="sourceLineNo">711</span> if (buf != null) {<a name="line.711"></a> |
| <span class="sourceLineNo">712</span> // Deep overhead of the byte buffer. Needs to be aligned separately.<a name="line.712"></a> |
| <span class="sourceLineNo">713</span> size += ClassSize.align(buf.capacity() + MULTI_BYTE_BUFFER_HEAP_SIZE);<a name="line.713"></a> |
| <span class="sourceLineNo">714</span> }<a name="line.714"></a> |
| <span class="sourceLineNo">715</span> return ClassSize.align(size);<a name="line.715"></a> |
| <span class="sourceLineNo">716</span> }<a name="line.716"></a> |
| <span class="sourceLineNo">717</span><a name="line.717"></a> |
| <span class="sourceLineNo">718</span> /**<a name="line.718"></a> |
| <span class="sourceLineNo">719</span> * Will be override by {@link SharedMemHFileBlock} or {@link ExclusiveMemHFileBlock}. Return true<a name="line.719"></a> |
| <span class="sourceLineNo">720</span> * by default.<a name="line.720"></a> |
| <span class="sourceLineNo">721</span> */<a name="line.721"></a> |
| <span class="sourceLineNo">722</span> public boolean isSharedMem() {<a name="line.722"></a> |
| <span class="sourceLineNo">723</span> return true;<a name="line.723"></a> |
| <span class="sourceLineNo">724</span> }<a name="line.724"></a> |
| <span class="sourceLineNo">725</span><a name="line.725"></a> |
| <span class="sourceLineNo">726</span> /**<a name="line.726"></a> |
| <span class="sourceLineNo">727</span> * Unified version 2 {@link HFile} block writer. The intended usage pattern is as follows:<a name="line.727"></a> |
| <span class="sourceLineNo">728</span> * <ol><a name="line.728"></a> |
| <span class="sourceLineNo">729</span> * <li>Construct an {@link HFileBlock.Writer}, providing a compression algorithm.<a name="line.729"></a> |
| <span class="sourceLineNo">730</span> * <li>Call {@link Writer#startWriting} and get a data stream to write to.<a name="line.730"></a> |
| <span class="sourceLineNo">731</span> * <li>Write your data into the stream.<a name="line.731"></a> |
| <span class="sourceLineNo">732</span> * <li>Call Writer#writeHeaderAndData(FSDataOutputStream) as many times as you need to. store the<a name="line.732"></a> |
| <span class="sourceLineNo">733</span> * serialized block into an external stream.<a name="line.733"></a> |
| <span class="sourceLineNo">734</span> * <li>Repeat to write more blocks.<a name="line.734"></a> |
| <span class="sourceLineNo">735</span> * </ol><a name="line.735"></a> |
| <span class="sourceLineNo">736</span> * <p><a name="line.736"></a> |
| <span class="sourceLineNo">737</span> */<a name="line.737"></a> |
| <span class="sourceLineNo">738</span> static class Writer implements ShipperListener {<a name="line.738"></a> |
| <span class="sourceLineNo">739</span> private enum State {<a name="line.739"></a> |
| <span class="sourceLineNo">740</span> INIT,<a name="line.740"></a> |
| <span class="sourceLineNo">741</span> WRITING,<a name="line.741"></a> |
| <span class="sourceLineNo">742</span> BLOCK_READY<a name="line.742"></a> |
| <span class="sourceLineNo">743</span> }<a name="line.743"></a> |
| <span class="sourceLineNo">744</span><a name="line.744"></a> |
| <span class="sourceLineNo">745</span> private int maxSizeUnCompressed;<a name="line.745"></a> |
| <span class="sourceLineNo">746</span><a name="line.746"></a> |
| <span class="sourceLineNo">747</span> private BlockCompressedSizePredicator compressedSizePredicator;<a name="line.747"></a> |
| <span class="sourceLineNo">748</span><a name="line.748"></a> |
| <span class="sourceLineNo">749</span> /** Writer state. Used to ensure the correct usage protocol. */<a name="line.749"></a> |
| <span class="sourceLineNo">750</span> private State state = State.INIT;<a name="line.750"></a> |
| <span class="sourceLineNo">751</span><a name="line.751"></a> |
| <span class="sourceLineNo">752</span> /** Data block encoder used for data blocks */<a name="line.752"></a> |
| <span class="sourceLineNo">753</span> private final HFileDataBlockEncoder dataBlockEncoder;<a name="line.753"></a> |
| <span class="sourceLineNo">754</span><a name="line.754"></a> |
| <span class="sourceLineNo">755</span> private HFileBlockEncodingContext dataBlockEncodingCtx;<a name="line.755"></a> |
| <span class="sourceLineNo">756</span><a name="line.756"></a> |
| <span class="sourceLineNo">757</span> /** block encoding context for non-data blocks */<a name="line.757"></a> |
| <span class="sourceLineNo">758</span> private HFileBlockDefaultEncodingContext defaultBlockEncodingCtx;<a name="line.758"></a> |
| <span class="sourceLineNo">759</span><a name="line.759"></a> |
| <span class="sourceLineNo">760</span> /**<a name="line.760"></a> |
| <span class="sourceLineNo">761</span> * The stream we use to accumulate data into a block in an uncompressed format. We reset this<a name="line.761"></a> |
| <span class="sourceLineNo">762</span> * stream at the end of each block and reuse it. The header is written as the first<a name="line.762"></a> |
| <span class="sourceLineNo">763</span> * {@link HConstants#HFILEBLOCK_HEADER_SIZE} bytes into this stream.<a name="line.763"></a> |
| <span class="sourceLineNo">764</span> */<a name="line.764"></a> |
| <span class="sourceLineNo">765</span> private ByteArrayOutputStream baosInMemory;<a name="line.765"></a> |
| <span class="sourceLineNo">766</span><a name="line.766"></a> |
| <span class="sourceLineNo">767</span> /**<a name="line.767"></a> |
| <span class="sourceLineNo">768</span> * Current block type. Set in {@link #startWriting(BlockType)}. Could be changed in<a name="line.768"></a> |
| <span class="sourceLineNo">769</span> * {@link #finishBlock()} from {@link BlockType#DATA} to {@link BlockType#ENCODED_DATA}.<a name="line.769"></a> |
| <span class="sourceLineNo">770</span> */<a name="line.770"></a> |
| <span class="sourceLineNo">771</span> private BlockType blockType;<a name="line.771"></a> |
| <span class="sourceLineNo">772</span><a name="line.772"></a> |
| <span class="sourceLineNo">773</span> /**<a name="line.773"></a> |
| <span class="sourceLineNo">774</span> * A stream that we write uncompressed bytes to, which compresses them and writes them to<a name="line.774"></a> |
| <span class="sourceLineNo">775</span> * {@link #baosInMemory}.<a name="line.775"></a> |
| <span class="sourceLineNo">776</span> */<a name="line.776"></a> |
| <span class="sourceLineNo">777</span> private DataOutputStream userDataStream;<a name="line.777"></a> |
| <span class="sourceLineNo">778</span><a name="line.778"></a> |
| <span class="sourceLineNo">779</span> /**<a name="line.779"></a> |
| <span class="sourceLineNo">780</span> * Bytes to be written to the file system, including the header. Compressed if compression is<a name="line.780"></a> |
| <span class="sourceLineNo">781</span> * turned on. It also includes the checksum data that immediately follows the block data.<a name="line.781"></a> |
| <span class="sourceLineNo">782</span> * (header + data + checksums)<a name="line.782"></a> |
| <span class="sourceLineNo">783</span> */<a name="line.783"></a> |
| <span class="sourceLineNo">784</span> private ByteArrayOutputStream onDiskBlockBytesWithHeader;<a name="line.784"></a> |
| <span class="sourceLineNo">785</span><a name="line.785"></a> |
| <span class="sourceLineNo">786</span> /**<a name="line.786"></a> |
| <span class="sourceLineNo">787</span> * The size of the checksum data on disk. It is used only if data is not compressed. If data is<a name="line.787"></a> |
| <span class="sourceLineNo">788</span> * compressed, then the checksums are already part of onDiskBytesWithHeader. If data is<a name="line.788"></a> |
| <span class="sourceLineNo">789</span> * uncompressed, then this variable stores the checksum data for this block.<a name="line.789"></a> |
| <span class="sourceLineNo">790</span> */<a name="line.790"></a> |
| <span class="sourceLineNo">791</span> private byte[] onDiskChecksum = HConstants.EMPTY_BYTE_ARRAY;<a name="line.791"></a> |
| <span class="sourceLineNo">792</span><a name="line.792"></a> |
| <span class="sourceLineNo">793</span> /**<a name="line.793"></a> |
| <span class="sourceLineNo">794</span> * Current block's start offset in the {@link HFile}. Set in<a name="line.794"></a> |
| <span class="sourceLineNo">795</span> * {@link #writeHeaderAndData(FSDataOutputStream)}.<a name="line.795"></a> |
| <span class="sourceLineNo">796</span> */<a name="line.796"></a> |
| <span class="sourceLineNo">797</span> private long startOffset;<a name="line.797"></a> |
| <span class="sourceLineNo">798</span><a name="line.798"></a> |
| <span class="sourceLineNo">799</span> /**<a name="line.799"></a> |
| <span class="sourceLineNo">800</span> * Offset of previous block by block type. Updated when the next block is started.<a name="line.800"></a> |
| <span class="sourceLineNo">801</span> */<a name="line.801"></a> |
| <span class="sourceLineNo">802</span> private long[] prevOffsetByType;<a name="line.802"></a> |
| <span class="sourceLineNo">803</span><a name="line.803"></a> |
| <span class="sourceLineNo">804</span> /** The offset of the previous block of the same type */<a name="line.804"></a> |
| <span class="sourceLineNo">805</span> private long prevOffset;<a name="line.805"></a> |
| <span class="sourceLineNo">806</span> /** Meta data that holds information about the hfileblock **/<a name="line.806"></a> |
| <span class="sourceLineNo">807</span> private HFileContext fileContext;<a name="line.807"></a> |
| <span class="sourceLineNo">808</span><a name="line.808"></a> |
| <span class="sourceLineNo">809</span> private final ByteBuffAllocator allocator;<a name="line.809"></a> |
| <span class="sourceLineNo">810</span><a name="line.810"></a> |
| <span class="sourceLineNo">811</span> @Override<a name="line.811"></a> |
| <span class="sourceLineNo">812</span> public void beforeShipped() {<a name="line.812"></a> |
| <span class="sourceLineNo">813</span> if (getEncodingState() != null) {<a name="line.813"></a> |
| <span class="sourceLineNo">814</span> getEncodingState().beforeShipped();<a name="line.814"></a> |
| <span class="sourceLineNo">815</span> }<a name="line.815"></a> |
| <span class="sourceLineNo">816</span> }<a name="line.816"></a> |
| <span class="sourceLineNo">817</span><a name="line.817"></a> |
| <span class="sourceLineNo">818</span> EncodingState getEncodingState() {<a name="line.818"></a> |
| <span class="sourceLineNo">819</span> return dataBlockEncodingCtx.getEncodingState();<a name="line.819"></a> |
| <span class="sourceLineNo">820</span> }<a name="line.820"></a> |
| <span class="sourceLineNo">821</span><a name="line.821"></a> |
| <span class="sourceLineNo">822</span> /**<a name="line.822"></a> |
| <span class="sourceLineNo">823</span> * @param dataBlockEncoder data block encoding algorithm to use<a name="line.823"></a> |
| <span class="sourceLineNo">824</span> */<a name="line.824"></a> |
| <span class="sourceLineNo">825</span> public Writer(Configuration conf, HFileDataBlockEncoder dataBlockEncoder,<a name="line.825"></a> |
| <span class="sourceLineNo">826</span> HFileContext fileContext) {<a name="line.826"></a> |
| <span class="sourceLineNo">827</span> this(conf, dataBlockEncoder, fileContext, ByteBuffAllocator.HEAP, fileContext.getBlocksize());<a name="line.827"></a> |
| <span class="sourceLineNo">828</span> }<a name="line.828"></a> |
| <span class="sourceLineNo">829</span><a name="line.829"></a> |
| <span class="sourceLineNo">830</span> public Writer(Configuration conf, HFileDataBlockEncoder dataBlockEncoder,<a name="line.830"></a> |
| <span class="sourceLineNo">831</span> HFileContext fileContext, ByteBuffAllocator allocator, int maxSizeUnCompressed) {<a name="line.831"></a> |
| <span class="sourceLineNo">832</span> if (fileContext.getBytesPerChecksum() < HConstants.HFILEBLOCK_HEADER_SIZE) {<a name="line.832"></a> |
| <span class="sourceLineNo">833</span> throw new RuntimeException("Unsupported value of bytesPerChecksum. " + " Minimum is "<a name="line.833"></a> |
| <span class="sourceLineNo">834</span> + HConstants.HFILEBLOCK_HEADER_SIZE + " but the configured value is "<a name="line.834"></a> |
| <span class="sourceLineNo">835</span> + fileContext.getBytesPerChecksum());<a name="line.835"></a> |
| <span class="sourceLineNo">836</span> }<a name="line.836"></a> |
| <span class="sourceLineNo">837</span> this.allocator = allocator;<a name="line.837"></a> |
| <span class="sourceLineNo">838</span> this.dataBlockEncoder =<a name="line.838"></a> |
| <span class="sourceLineNo">839</span> dataBlockEncoder != null ? dataBlockEncoder : NoOpDataBlockEncoder.INSTANCE;<a name="line.839"></a> |
| <span class="sourceLineNo">840</span> this.dataBlockEncodingCtx = this.dataBlockEncoder.newDataBlockEncodingContext(conf,<a name="line.840"></a> |
| <span class="sourceLineNo">841</span> HConstants.HFILEBLOCK_DUMMY_HEADER, fileContext);<a name="line.841"></a> |
| <span class="sourceLineNo">842</span> // TODO: This should be lazily instantiated<a name="line.842"></a> |
| <span class="sourceLineNo">843</span> this.defaultBlockEncodingCtx = new HFileBlockDefaultEncodingContext(conf, null,<a name="line.843"></a> |
| <span class="sourceLineNo">844</span> HConstants.HFILEBLOCK_DUMMY_HEADER, fileContext);<a name="line.844"></a> |
| <span class="sourceLineNo">845</span> // TODO: Set BAOS initial size. Use fileContext.getBlocksize() and add for header/checksum<a name="line.845"></a> |
| <span class="sourceLineNo">846</span> baosInMemory = new ByteArrayOutputStream();<a name="line.846"></a> |
| <span class="sourceLineNo">847</span> prevOffsetByType = new long[BlockType.values().length];<a name="line.847"></a> |
| <span class="sourceLineNo">848</span> for (int i = 0; i < prevOffsetByType.length; ++i) {<a name="line.848"></a> |
| <span class="sourceLineNo">849</span> prevOffsetByType[i] = UNSET;<a name="line.849"></a> |
| <span class="sourceLineNo">850</span> }<a name="line.850"></a> |
| <span class="sourceLineNo">851</span> // TODO: Why fileContext saved away when we have dataBlockEncoder and/or<a name="line.851"></a> |
| <span class="sourceLineNo">852</span> // defaultDataBlockEncoder?<a name="line.852"></a> |
| <span class="sourceLineNo">853</span> this.fileContext = fileContext;<a name="line.853"></a> |
| <span class="sourceLineNo">854</span> this.compressedSizePredicator = (BlockCompressedSizePredicator) ReflectionUtils.newInstance(<a name="line.854"></a> |
| <span class="sourceLineNo">855</span> conf.getClass(BLOCK_COMPRESSED_SIZE_PREDICATOR, UncompressedBlockSizePredicator.class),<a name="line.855"></a> |
| <span class="sourceLineNo">856</span> new Configuration(conf));<a name="line.856"></a> |
| <span class="sourceLineNo">857</span> this.maxSizeUnCompressed = maxSizeUnCompressed;<a name="line.857"></a> |
| <span class="sourceLineNo">858</span> }<a name="line.858"></a> |
| <span class="sourceLineNo">859</span><a name="line.859"></a> |
| <span class="sourceLineNo">860</span> /**<a name="line.860"></a> |
| <span class="sourceLineNo">861</span> * Starts writing into the block. The previous block's data is discarded.<a name="line.861"></a> |
| <span class="sourceLineNo">862</span> * @return the stream the user can write their data into<a name="line.862"></a> |
| <span class="sourceLineNo">863</span> */<a name="line.863"></a> |
| <span class="sourceLineNo">864</span> DataOutputStream startWriting(BlockType newBlockType) throws IOException {<a name="line.864"></a> |
| <span class="sourceLineNo">865</span> if (state == State.BLOCK_READY && startOffset != -1) {<a name="line.865"></a> |
| <span class="sourceLineNo">866</span> // We had a previous block that was written to a stream at a specific<a name="line.866"></a> |
| <span class="sourceLineNo">867</span> // offset. Save that offset as the last offset of a block of that type.<a name="line.867"></a> |
| <span class="sourceLineNo">868</span> prevOffsetByType[blockType.getId()] = startOffset;<a name="line.868"></a> |
| <span class="sourceLineNo">869</span> }<a name="line.869"></a> |
| <span class="sourceLineNo">870</span><a name="line.870"></a> |
| <span class="sourceLineNo">871</span> startOffset = -1;<a name="line.871"></a> |
| <span class="sourceLineNo">872</span> blockType = newBlockType;<a name="line.872"></a> |
| <span class="sourceLineNo">873</span><a name="line.873"></a> |
| <span class="sourceLineNo">874</span> baosInMemory.reset();<a name="line.874"></a> |
| <span class="sourceLineNo">875</span> baosInMemory.write(HConstants.HFILEBLOCK_DUMMY_HEADER);<a name="line.875"></a> |
| <span class="sourceLineNo">876</span><a name="line.876"></a> |
| <span class="sourceLineNo">877</span> state = State.WRITING;<a name="line.877"></a> |
| <span class="sourceLineNo">878</span><a name="line.878"></a> |
| <span class="sourceLineNo">879</span> // We will compress it later in finishBlock()<a name="line.879"></a> |
| <span class="sourceLineNo">880</span> userDataStream = new ByteBufferWriterDataOutputStream(baosInMemory);<a name="line.880"></a> |
| <span class="sourceLineNo">881</span> if (newBlockType == BlockType.DATA) {<a name="line.881"></a> |
| <span class="sourceLineNo">882</span> this.dataBlockEncoder.startBlockEncoding(dataBlockEncodingCtx, userDataStream);<a name="line.882"></a> |
| <span class="sourceLineNo">883</span> }<a name="line.883"></a> |
| <span class="sourceLineNo">884</span> return userDataStream;<a name="line.884"></a> |
| <span class="sourceLineNo">885</span> }<a name="line.885"></a> |
| <span class="sourceLineNo">886</span><a name="line.886"></a> |
| <span class="sourceLineNo">887</span> /**<a name="line.887"></a> |
| <span class="sourceLineNo">888</span> * Writes the Cell to this block<a name="line.888"></a> |
| <span class="sourceLineNo">889</span> */<a name="line.889"></a> |
| <span class="sourceLineNo">890</span> void write(Cell cell) throws IOException {<a name="line.890"></a> |
| <span class="sourceLineNo">891</span> expectState(State.WRITING);<a name="line.891"></a> |
| <span class="sourceLineNo">892</span> this.dataBlockEncoder.encode(cell, dataBlockEncodingCtx, this.userDataStream);<a name="line.892"></a> |
| <span class="sourceLineNo">893</span> }<a name="line.893"></a> |
| <span class="sourceLineNo">894</span><a name="line.894"></a> |
| <span class="sourceLineNo">895</span> /**<a name="line.895"></a> |
| <span class="sourceLineNo">896</span> * Transitions the block writer from the "writing" state to the "block ready" state. Does<a name="line.896"></a> |
| <span class="sourceLineNo">897</span> * nothing if a block is already finished.<a name="line.897"></a> |
| <span class="sourceLineNo">898</span> */<a name="line.898"></a> |
| <span class="sourceLineNo">899</span> void ensureBlockReady() throws IOException {<a name="line.899"></a> |
| <span class="sourceLineNo">900</span> Preconditions.checkState(state != State.INIT, "Unexpected state: " + state);<a name="line.900"></a> |
| <span class="sourceLineNo">901</span><a name="line.901"></a> |
| <span class="sourceLineNo">902</span> if (state == State.BLOCK_READY) {<a name="line.902"></a> |
| <span class="sourceLineNo">903</span> return;<a name="line.903"></a> |
| <span class="sourceLineNo">904</span> }<a name="line.904"></a> |
| <span class="sourceLineNo">905</span><a name="line.905"></a> |
| <span class="sourceLineNo">906</span> // This will set state to BLOCK_READY.<a name="line.906"></a> |
| <span class="sourceLineNo">907</span> finishBlock();<a name="line.907"></a> |
| <span class="sourceLineNo">908</span> }<a name="line.908"></a> |
| <span class="sourceLineNo">909</span><a name="line.909"></a> |
| <span class="sourceLineNo">910</span> public boolean checkBoundariesWithPredicate() {<a name="line.910"></a> |
| <span class="sourceLineNo">911</span> int rawBlockSize = encodedBlockSizeWritten();<a name="line.911"></a> |
| <span class="sourceLineNo">912</span> if (rawBlockSize >= maxSizeUnCompressed) {<a name="line.912"></a> |
| <span class="sourceLineNo">913</span> return true;<a name="line.913"></a> |
| <span class="sourceLineNo">914</span> } else {<a name="line.914"></a> |
| <span class="sourceLineNo">915</span> return compressedSizePredicator.shouldFinishBlock(rawBlockSize);<a name="line.915"></a> |
| <span class="sourceLineNo">916</span> }<a name="line.916"></a> |
| <span class="sourceLineNo">917</span> }<a name="line.917"></a> |
| <span class="sourceLineNo">918</span><a name="line.918"></a> |
| <span class="sourceLineNo">919</span> /**<a name="line.919"></a> |
| <span class="sourceLineNo">920</span> * Finish up writing of the block. Flushes the compressing stream (if using compression), fills<a name="line.920"></a> |
| <span class="sourceLineNo">921</span> * out the header, does any compression/encryption of bytes to flush out to disk, and manages<a name="line.921"></a> |
| <span class="sourceLineNo">922</span> * the cache on write content, if applicable. Sets block write state to "block ready".<a name="line.922"></a> |
| <span class="sourceLineNo">923</span> */<a name="line.923"></a> |
| <span class="sourceLineNo">924</span> private void finishBlock() throws IOException {<a name="line.924"></a> |
| <span class="sourceLineNo">925</span> if (blockType == BlockType.DATA) {<a name="line.925"></a> |
| <span class="sourceLineNo">926</span> this.dataBlockEncoder.endBlockEncoding(dataBlockEncodingCtx, userDataStream,<a name="line.926"></a> |
| <span class="sourceLineNo">927</span> baosInMemory.getBuffer(), blockType);<a name="line.927"></a> |
| <span class="sourceLineNo">928</span> blockType = dataBlockEncodingCtx.getBlockType();<a name="line.928"></a> |
| <span class="sourceLineNo">929</span> }<a name="line.929"></a> |
| <span class="sourceLineNo">930</span> userDataStream.flush();<a name="line.930"></a> |
| <span class="sourceLineNo">931</span> prevOffset = prevOffsetByType[blockType.getId()];<a name="line.931"></a> |
| <span class="sourceLineNo">932</span><a name="line.932"></a> |
| <span class="sourceLineNo">933</span> // We need to cache the unencoded/uncompressed size before changing the block state<a name="line.933"></a> |
| <span class="sourceLineNo">934</span> int rawBlockSize = 0;<a name="line.934"></a> |
| <span class="sourceLineNo">935</span> if (this.getEncodingState() != null) {<a name="line.935"></a> |
| <span class="sourceLineNo">936</span> rawBlockSize = encodedBlockSizeWritten();<a name="line.936"></a> |
| <span class="sourceLineNo">937</span> }<a name="line.937"></a> |
| <span class="sourceLineNo">938</span> // We need to set state before we can package the block up for cache-on-write. In a way, the<a name="line.938"></a> |
| <span class="sourceLineNo">939</span> // block is ready, but not yet encoded or compressed.<a name="line.939"></a> |
| <span class="sourceLineNo">940</span> state = State.BLOCK_READY;<a name="line.940"></a> |
| <span class="sourceLineNo">941</span> Bytes compressAndEncryptDat;<a name="line.941"></a> |
| <span class="sourceLineNo">942</span> if (blockType == BlockType.DATA || blockType == BlockType.ENCODED_DATA) {<a name="line.942"></a> |
| <span class="sourceLineNo">943</span> compressAndEncryptDat =<a name="line.943"></a> |
| <span class="sourceLineNo">944</span> dataBlockEncodingCtx.compressAndEncrypt(baosInMemory.getBuffer(), 0, baosInMemory.size());<a name="line.944"></a> |
| <span class="sourceLineNo">945</span> } else {<a name="line.945"></a> |
| <span class="sourceLineNo">946</span> compressAndEncryptDat = defaultBlockEncodingCtx.compressAndEncrypt(baosInMemory.getBuffer(),<a name="line.946"></a> |
| <span class="sourceLineNo">947</span> 0, baosInMemory.size());<a name="line.947"></a> |
| <span class="sourceLineNo">948</span> }<a name="line.948"></a> |
| <span class="sourceLineNo">949</span> if (compressAndEncryptDat == null) {<a name="line.949"></a> |
| <span class="sourceLineNo">950</span> compressAndEncryptDat = new Bytes(baosInMemory.getBuffer(), 0, baosInMemory.size());<a name="line.950"></a> |
| <span class="sourceLineNo">951</span> }<a name="line.951"></a> |
| <span class="sourceLineNo">952</span> if (onDiskBlockBytesWithHeader == null) {<a name="line.952"></a> |
| <span class="sourceLineNo">953</span> onDiskBlockBytesWithHeader = new ByteArrayOutputStream(compressAndEncryptDat.getLength());<a name="line.953"></a> |
| <span class="sourceLineNo">954</span> }<a name="line.954"></a> |
| <span class="sourceLineNo">955</span> onDiskBlockBytesWithHeader.reset();<a name="line.955"></a> |
| <span class="sourceLineNo">956</span> onDiskBlockBytesWithHeader.write(compressAndEncryptDat.get(),<a name="line.956"></a> |
| <span class="sourceLineNo">957</span> compressAndEncryptDat.getOffset(), compressAndEncryptDat.getLength());<a name="line.957"></a> |
| <span class="sourceLineNo">958</span> // Update raw and compressed sizes in the predicate<a name="line.958"></a> |
| <span class="sourceLineNo">959</span> compressedSizePredicator.updateLatestBlockSizes(fileContext, rawBlockSize,<a name="line.959"></a> |
| <span class="sourceLineNo">960</span> onDiskBlockBytesWithHeader.size());<a name="line.960"></a> |
| <span class="sourceLineNo">961</span><a name="line.961"></a> |
| <span class="sourceLineNo">962</span> // Calculate how many bytes we need for checksum on the tail of the block.<a name="line.962"></a> |
| <span class="sourceLineNo">963</span> int numBytes = (int) ChecksumUtil.numBytes(onDiskBlockBytesWithHeader.size(),<a name="line.963"></a> |
| <span class="sourceLineNo">964</span> fileContext.getBytesPerChecksum());<a name="line.964"></a> |
| <span class="sourceLineNo">965</span><a name="line.965"></a> |
| <span class="sourceLineNo">966</span> // Put the header for the on disk bytes; header currently is unfilled-out<a name="line.966"></a> |
| <span class="sourceLineNo">967</span> putHeader(onDiskBlockBytesWithHeader, onDiskBlockBytesWithHeader.size() + numBytes,<a name="line.967"></a> |
| <span class="sourceLineNo">968</span> baosInMemory.size(), onDiskBlockBytesWithHeader.size());<a name="line.968"></a> |
| <span class="sourceLineNo">969</span><a name="line.969"></a> |
| <span class="sourceLineNo">970</span> if (onDiskChecksum.length != numBytes) {<a name="line.970"></a> |
| <span class="sourceLineNo">971</span> onDiskChecksum = new byte[numBytes];<a name="line.971"></a> |
| <span class="sourceLineNo">972</span> }<a name="line.972"></a> |
| <span class="sourceLineNo">973</span> ChecksumUtil.generateChecksums(onDiskBlockBytesWithHeader.getBuffer(), 0,<a name="line.973"></a> |
| <span class="sourceLineNo">974</span> onDiskBlockBytesWithHeader.size(), onDiskChecksum, 0, fileContext.getChecksumType(),<a name="line.974"></a> |
| <span class="sourceLineNo">975</span> fileContext.getBytesPerChecksum());<a name="line.975"></a> |
| <span class="sourceLineNo">976</span> }<a name="line.976"></a> |
| <span class="sourceLineNo">977</span><a name="line.977"></a> |
| <span class="sourceLineNo">978</span> /**<a name="line.978"></a> |
| <span class="sourceLineNo">979</span> * Put the header into the given byte array at the given offset.<a name="line.979"></a> |
| <span class="sourceLineNo">980</span> * @param onDiskSize size of the block on disk header + data + checksum<a name="line.980"></a> |
| <span class="sourceLineNo">981</span> * @param uncompressedSize size of the block after decompression (but before optional data block<a name="line.981"></a> |
| <span class="sourceLineNo">982</span> * decoding) including header<a name="line.982"></a> |
| <span class="sourceLineNo">983</span> * @param onDiskDataSize size of the block on disk with header and data but not including the<a name="line.983"></a> |
| <span class="sourceLineNo">984</span> * checksums<a name="line.984"></a> |
| <span class="sourceLineNo">985</span> */<a name="line.985"></a> |
| <span class="sourceLineNo">986</span> private void putHeader(byte[] dest, int offset, int onDiskSize, int uncompressedSize,<a name="line.986"></a> |
| <span class="sourceLineNo">987</span> int onDiskDataSize) {<a name="line.987"></a> |
| <span class="sourceLineNo">988</span> offset = blockType.put(dest, offset);<a name="line.988"></a> |
| <span class="sourceLineNo">989</span> offset = Bytes.putInt(dest, offset, onDiskSize - HConstants.HFILEBLOCK_HEADER_SIZE);<a name="line.989"></a> |
| <span class="sourceLineNo">990</span> offset = Bytes.putInt(dest, offset, uncompressedSize - HConstants.HFILEBLOCK_HEADER_SIZE);<a name="line.990"></a> |
| <span class="sourceLineNo">991</span> offset = Bytes.putLong(dest, offset, prevOffset);<a name="line.991"></a> |
| <span class="sourceLineNo">992</span> offset = Bytes.putByte(dest, offset, fileContext.getChecksumType().getCode());<a name="line.992"></a> |
| <span class="sourceLineNo">993</span> offset = Bytes.putInt(dest, offset, fileContext.getBytesPerChecksum());<a name="line.993"></a> |
| <span class="sourceLineNo">994</span> Bytes.putInt(dest, offset, onDiskDataSize);<a name="line.994"></a> |
| <span class="sourceLineNo">995</span> }<a name="line.995"></a> |
| <span class="sourceLineNo">996</span><a name="line.996"></a> |
| <span class="sourceLineNo">997</span> private void putHeader(ByteBuff buff, int onDiskSize, int uncompressedSize,<a name="line.997"></a> |
| <span class="sourceLineNo">998</span> int onDiskDataSize) {<a name="line.998"></a> |
| <span class="sourceLineNo">999</span> buff.rewind();<a name="line.999"></a> |
| <span class="sourceLineNo">1000</span> blockType.write(buff);<a name="line.1000"></a> |
| <span class="sourceLineNo">1001</span> buff.putInt(onDiskSize - HConstants.HFILEBLOCK_HEADER_SIZE);<a name="line.1001"></a> |
| <span class="sourceLineNo">1002</span> buff.putInt(uncompressedSize - HConstants.HFILEBLOCK_HEADER_SIZE);<a name="line.1002"></a> |
| <span class="sourceLineNo">1003</span> buff.putLong(prevOffset);<a name="line.1003"></a> |
| <span class="sourceLineNo">1004</span> buff.put(fileContext.getChecksumType().getCode());<a name="line.1004"></a> |
| <span class="sourceLineNo">1005</span> buff.putInt(fileContext.getBytesPerChecksum());<a name="line.1005"></a> |
| <span class="sourceLineNo">1006</span> buff.putInt(onDiskDataSize);<a name="line.1006"></a> |
| <span class="sourceLineNo">1007</span> }<a name="line.1007"></a> |
| <span class="sourceLineNo">1008</span><a name="line.1008"></a> |
| <span class="sourceLineNo">1009</span> private void putHeader(ByteArrayOutputStream dest, int onDiskSize, int uncompressedSize,<a name="line.1009"></a> |
| <span class="sourceLineNo">1010</span> int onDiskDataSize) {<a name="line.1010"></a> |
| <span class="sourceLineNo">1011</span> putHeader(dest.getBuffer(), 0, onDiskSize, uncompressedSize, onDiskDataSize);<a name="line.1011"></a> |
| <span class="sourceLineNo">1012</span> }<a name="line.1012"></a> |
| <span class="sourceLineNo">1013</span><a name="line.1013"></a> |
| <span class="sourceLineNo">1014</span> /**<a name="line.1014"></a> |
| <span class="sourceLineNo">1015</span> * Similar to {@link #writeHeaderAndData(FSDataOutputStream)}, but records the offset of this<a name="line.1015"></a> |
| <span class="sourceLineNo">1016</span> * block so that it can be referenced in the next block of the same type.<a name="line.1016"></a> |
| <span class="sourceLineNo">1017</span> */<a name="line.1017"></a> |
| <span class="sourceLineNo">1018</span> void writeHeaderAndData(FSDataOutputStream out) throws IOException {<a name="line.1018"></a> |
| <span class="sourceLineNo">1019</span> long offset = out.getPos();<a name="line.1019"></a> |
| <span class="sourceLineNo">1020</span> if (startOffset != UNSET && offset != startOffset) {<a name="line.1020"></a> |
| <span class="sourceLineNo">1021</span> throw new IOException("A " + blockType + " block written to a "<a name="line.1021"></a> |
| <span class="sourceLineNo">1022</span> + "stream twice, first at offset " + startOffset + ", then at " + offset);<a name="line.1022"></a> |
| <span class="sourceLineNo">1023</span> }<a name="line.1023"></a> |
| <span class="sourceLineNo">1024</span> startOffset = offset;<a name="line.1024"></a> |
| <span class="sourceLineNo">1025</span> finishBlockAndWriteHeaderAndData(out);<a name="line.1025"></a> |
| <span class="sourceLineNo">1026</span> }<a name="line.1026"></a> |
| <span class="sourceLineNo">1027</span><a name="line.1027"></a> |
| <span class="sourceLineNo">1028</span> /**<a name="line.1028"></a> |
| <span class="sourceLineNo">1029</span> * Writes the header and the compressed data of this block (or uncompressed data when not using<a name="line.1029"></a> |
| <span class="sourceLineNo">1030</span> * compression) into the given stream. Can be called in the "writing" state or in the "block<a name="line.1030"></a> |
| <span class="sourceLineNo">1031</span> * ready" state. If called in the "writing" state, transitions the writer to the "block ready"<a name="line.1031"></a> |
| <span class="sourceLineNo">1032</span> * state.<a name="line.1032"></a> |
| <span class="sourceLineNo">1033</span> * @param out the output stream to write the<a name="line.1033"></a> |
| <span class="sourceLineNo">1034</span> */<a name="line.1034"></a> |
| <span class="sourceLineNo">1035</span> protected void finishBlockAndWriteHeaderAndData(DataOutputStream out) throws IOException {<a name="line.1035"></a> |
| <span class="sourceLineNo">1036</span> ensureBlockReady();<a name="line.1036"></a> |
| <span class="sourceLineNo">1037</span> long startTime = EnvironmentEdgeManager.currentTime();<a name="line.1037"></a> |
| <span class="sourceLineNo">1038</span> out.write(onDiskBlockBytesWithHeader.getBuffer(), 0, onDiskBlockBytesWithHeader.size());<a name="line.1038"></a> |
| <span class="sourceLineNo">1039</span> out.write(onDiskChecksum);<a name="line.1039"></a> |
| <span class="sourceLineNo">1040</span> HFile.updateWriteLatency(EnvironmentEdgeManager.currentTime() - startTime);<a name="line.1040"></a> |
| <span class="sourceLineNo">1041</span> }<a name="line.1041"></a> |
| <span class="sourceLineNo">1042</span><a name="line.1042"></a> |
| <span class="sourceLineNo">1043</span> /**<a name="line.1043"></a> |
| <span class="sourceLineNo">1044</span> * Returns the header or the compressed data (or uncompressed data when not using compression)<a name="line.1044"></a> |
| <span class="sourceLineNo">1045</span> * as a byte array. Can be called in the "writing" state or in the "block ready" state. If<a name="line.1045"></a> |
| <span class="sourceLineNo">1046</span> * called in the "writing" state, transitions the writer to the "block ready" state. This<a name="line.1046"></a> |
| <span class="sourceLineNo">1047</span> * returns the header + data + checksums stored on disk.<a name="line.1047"></a> |
| <span class="sourceLineNo">1048</span> * @return header and data as they would be stored on disk in a byte array<a name="line.1048"></a> |
| <span class="sourceLineNo">1049</span> */<a name="line.1049"></a> |
| <span class="sourceLineNo">1050</span> byte[] getHeaderAndDataForTest() throws IOException {<a name="line.1050"></a> |
| <span class="sourceLineNo">1051</span> ensureBlockReady();<a name="line.1051"></a> |
| <span class="sourceLineNo">1052</span> // This is not very optimal, because we are doing an extra copy.<a name="line.1052"></a> |
| <span class="sourceLineNo">1053</span> // But this method is used only by unit tests.<a name="line.1053"></a> |
| <span class="sourceLineNo">1054</span> byte[] output = new byte[onDiskBlockBytesWithHeader.size() + onDiskChecksum.length];<a name="line.1054"></a> |
| <span class="sourceLineNo">1055</span> System.arraycopy(onDiskBlockBytesWithHeader.getBuffer(), 0, output, 0,<a name="line.1055"></a> |
| <span class="sourceLineNo">1056</span> onDiskBlockBytesWithHeader.size());<a name="line.1056"></a> |
| <span class="sourceLineNo">1057</span> System.arraycopy(onDiskChecksum, 0, output, onDiskBlockBytesWithHeader.size(),<a name="line.1057"></a> |
| <span class="sourceLineNo">1058</span> onDiskChecksum.length);<a name="line.1058"></a> |
| <span class="sourceLineNo">1059</span> return output;<a name="line.1059"></a> |
| <span class="sourceLineNo">1060</span> }<a name="line.1060"></a> |
| <span class="sourceLineNo">1061</span><a name="line.1061"></a> |
| <span class="sourceLineNo">1062</span> /**<a name="line.1062"></a> |
| <span class="sourceLineNo">1063</span> * Releases resources used by this writer.<a name="line.1063"></a> |
| <span class="sourceLineNo">1064</span> */<a name="line.1064"></a> |
| <span class="sourceLineNo">1065</span> void release() {<a name="line.1065"></a> |
| <span class="sourceLineNo">1066</span> if (dataBlockEncodingCtx != null) {<a name="line.1066"></a> |
| <span class="sourceLineNo">1067</span> dataBlockEncodingCtx.close();<a name="line.1067"></a> |
| <span class="sourceLineNo">1068</span> dataBlockEncodingCtx = null;<a name="line.1068"></a> |
| <span class="sourceLineNo">1069</span> }<a name="line.1069"></a> |
| <span class="sourceLineNo">1070</span> if (defaultBlockEncodingCtx != null) {<a name="line.1070"></a> |
| <span class="sourceLineNo">1071</span> defaultBlockEncodingCtx.close();<a name="line.1071"></a> |
| <span class="sourceLineNo">1072</span> defaultBlockEncodingCtx = null;<a name="line.1072"></a> |
| <span class="sourceLineNo">1073</span> }<a name="line.1073"></a> |
| <span class="sourceLineNo">1074</span> }<a name="line.1074"></a> |
| <span class="sourceLineNo">1075</span><a name="line.1075"></a> |
| <span class="sourceLineNo">1076</span> /**<a name="line.1076"></a> |
| <span class="sourceLineNo">1077</span> * Returns the on-disk size of the data portion of the block. This is the compressed size if<a name="line.1077"></a> |
| <span class="sourceLineNo">1078</span> * compression is enabled. Can only be called in the "block ready" state. Header is not<a name="line.1078"></a> |
| <span class="sourceLineNo">1079</span> * compressed, and its size is not included in the return value.<a name="line.1079"></a> |
| <span class="sourceLineNo">1080</span> * @return the on-disk size of the block, not including the header.<a name="line.1080"></a> |
| <span class="sourceLineNo">1081</span> */<a name="line.1081"></a> |
| <span class="sourceLineNo">1082</span> int getOnDiskSizeWithoutHeader() {<a name="line.1082"></a> |
| <span class="sourceLineNo">1083</span> expectState(State.BLOCK_READY);<a name="line.1083"></a> |
| <span class="sourceLineNo">1084</span> return onDiskBlockBytesWithHeader.size() + onDiskChecksum.length<a name="line.1084"></a> |
| <span class="sourceLineNo">1085</span> - HConstants.HFILEBLOCK_HEADER_SIZE;<a name="line.1085"></a> |
| <span class="sourceLineNo">1086</span> }<a name="line.1086"></a> |
| <span class="sourceLineNo">1087</span><a name="line.1087"></a> |
| <span class="sourceLineNo">1088</span> /**<a name="line.1088"></a> |
| <span class="sourceLineNo">1089</span> * Returns the on-disk size of the block. Can only be called in the "block ready" state.<a name="line.1089"></a> |
| <span class="sourceLineNo">1090</span> * @return the on-disk size of the block ready to be written, including the header size, the<a name="line.1090"></a> |
| <span class="sourceLineNo">1091</span> * data and the checksum data.<a name="line.1091"></a> |
| <span class="sourceLineNo">1092</span> */<a name="line.1092"></a> |
| <span class="sourceLineNo">1093</span> int getOnDiskSizeWithHeader() {<a name="line.1093"></a> |
| <span class="sourceLineNo">1094</span> expectState(State.BLOCK_READY);<a name="line.1094"></a> |
| <span class="sourceLineNo">1095</span> return onDiskBlockBytesWithHeader.size() + onDiskChecksum.length;<a name="line.1095"></a> |
| <span class="sourceLineNo">1096</span> }<a name="line.1096"></a> |
| <span class="sourceLineNo">1097</span><a name="line.1097"></a> |
| <span class="sourceLineNo">1098</span> /**<a name="line.1098"></a> |
| <span class="sourceLineNo">1099</span> * The uncompressed size of the block data. Does not include header size.<a name="line.1099"></a> |
| <span class="sourceLineNo">1100</span> */<a name="line.1100"></a> |
| <span class="sourceLineNo">1101</span> int getUncompressedSizeWithoutHeader() {<a name="line.1101"></a> |
| <span class="sourceLineNo">1102</span> expectState(State.BLOCK_READY);<a name="line.1102"></a> |
| <span class="sourceLineNo">1103</span> return baosInMemory.size() - HConstants.HFILEBLOCK_HEADER_SIZE;<a name="line.1103"></a> |
| <span class="sourceLineNo">1104</span> }<a name="line.1104"></a> |
| <span class="sourceLineNo">1105</span><a name="line.1105"></a> |
| <span class="sourceLineNo">1106</span> /**<a name="line.1106"></a> |
| <span class="sourceLineNo">1107</span> * The uncompressed size of the block data, including header size.<a name="line.1107"></a> |
| <span class="sourceLineNo">1108</span> */<a name="line.1108"></a> |
| <span class="sourceLineNo">1109</span> public int getUncompressedSizeWithHeader() {<a name="line.1109"></a> |
| <span class="sourceLineNo">1110</span> expectState(State.BLOCK_READY);<a name="line.1110"></a> |
| <span class="sourceLineNo">1111</span> return baosInMemory.size();<a name="line.1111"></a> |
| <span class="sourceLineNo">1112</span> }<a name="line.1112"></a> |
| <span class="sourceLineNo">1113</span><a name="line.1113"></a> |
| <span class="sourceLineNo">1114</span> /** Returns true if a block is being written */<a name="line.1114"></a> |
| <span class="sourceLineNo">1115</span> boolean isWriting() {<a name="line.1115"></a> |
| <span class="sourceLineNo">1116</span> return state == State.WRITING;<a name="line.1116"></a> |
| <span class="sourceLineNo">1117</span> }<a name="line.1117"></a> |
| <span class="sourceLineNo">1118</span><a name="line.1118"></a> |
| <span class="sourceLineNo">1119</span> /**<a name="line.1119"></a> |
| <span class="sourceLineNo">1120</span> * Returns the number of bytes written into the current block so far, or zero if not writing the<a name="line.1120"></a> |
| <span class="sourceLineNo">1121</span> * block at the moment. Note that this will return zero in the "block ready" state as well.<a name="line.1121"></a> |
| <span class="sourceLineNo">1122</span> * @return the number of bytes written<a name="line.1122"></a> |
| <span class="sourceLineNo">1123</span> */<a name="line.1123"></a> |
| <span class="sourceLineNo">1124</span> public int encodedBlockSizeWritten() {<a name="line.1124"></a> |
| <span class="sourceLineNo">1125</span> return state != State.WRITING ? 0 : this.getEncodingState().getEncodedDataSizeWritten();<a name="line.1125"></a> |
| <span class="sourceLineNo">1126</span> }<a name="line.1126"></a> |
| <span class="sourceLineNo">1127</span><a name="line.1127"></a> |
| <span class="sourceLineNo">1128</span> /**<a name="line.1128"></a> |
| <span class="sourceLineNo">1129</span> * Returns the number of bytes written into the current block so far, or zero if not writing the<a name="line.1129"></a> |
| <span class="sourceLineNo">1130</span> * block at the moment. Note that this will return zero in the "block ready" state as well.<a name="line.1130"></a> |
| <span class="sourceLineNo">1131</span> * @return the number of bytes written<a name="line.1131"></a> |
| <span class="sourceLineNo">1132</span> */<a name="line.1132"></a> |
| <span class="sourceLineNo">1133</span> public int blockSizeWritten() {<a name="line.1133"></a> |
| <span class="sourceLineNo">1134</span> return state != State.WRITING ? 0 : this.getEncodingState().getUnencodedDataSizeWritten();<a name="line.1134"></a> |
| <span class="sourceLineNo">1135</span> }<a name="line.1135"></a> |
| <span class="sourceLineNo">1136</span><a name="line.1136"></a> |
| <span class="sourceLineNo">1137</span> /**<a name="line.1137"></a> |
| <span class="sourceLineNo">1138</span> * Clones the header followed by the uncompressed data, even if using compression. This is<a name="line.1138"></a> |
| <span class="sourceLineNo">1139</span> * needed for storing uncompressed blocks in the block cache. Can be called in the "writing"<a name="line.1139"></a> |
| <span class="sourceLineNo">1140</span> * state or the "block ready" state. Returns only the header and data, does not include checksum<a name="line.1140"></a> |
| <span class="sourceLineNo">1141</span> * data.<a name="line.1141"></a> |
| <span class="sourceLineNo">1142</span> * @return Returns an uncompressed block ByteBuff for caching on write<a name="line.1142"></a> |
| <span class="sourceLineNo">1143</span> */<a name="line.1143"></a> |
| <span class="sourceLineNo">1144</span> ByteBuff cloneUncompressedBufferWithHeader() {<a name="line.1144"></a> |
| <span class="sourceLineNo">1145</span> expectState(State.BLOCK_READY);<a name="line.1145"></a> |
| <span class="sourceLineNo">1146</span> ByteBuff bytebuff = allocator.allocate(baosInMemory.size());<a name="line.1146"></a> |
| <span class="sourceLineNo">1147</span> baosInMemory.toByteBuff(bytebuff);<a name="line.1147"></a> |
| <span class="sourceLineNo">1148</span> int numBytes = (int) ChecksumUtil.numBytes(onDiskBlockBytesWithHeader.size(),<a name="line.1148"></a> |
| <span class="sourceLineNo">1149</span> fileContext.getBytesPerChecksum());<a name="line.1149"></a> |
| <span class="sourceLineNo">1150</span> putHeader(bytebuff, onDiskBlockBytesWithHeader.size() + numBytes, baosInMemory.size(),<a name="line.1150"></a> |
| <span class="sourceLineNo">1151</span> onDiskBlockBytesWithHeader.size());<a name="line.1151"></a> |
| <span class="sourceLineNo">1152</span> bytebuff.rewind();<a name="line.1152"></a> |
| <span class="sourceLineNo">1153</span> return bytebuff;<a name="line.1153"></a> |
| <span class="sourceLineNo">1154</span> }<a name="line.1154"></a> |
| <span class="sourceLineNo">1155</span><a name="line.1155"></a> |
| <span class="sourceLineNo">1156</span> /**<a name="line.1156"></a> |
| <span class="sourceLineNo">1157</span> * Clones the header followed by the on-disk (compressed/encoded/encrypted) data. This is needed<a name="line.1157"></a> |
| <span class="sourceLineNo">1158</span> * for storing packed blocks in the block cache. Returns only the header and data, Does not<a name="line.1158"></a> |
| <span class="sourceLineNo">1159</span> * include checksum data.<a name="line.1159"></a> |
| <span class="sourceLineNo">1160</span> * @return Returns a copy of block bytes for caching on write<a name="line.1160"></a> |
| <span class="sourceLineNo">1161</span> */<a name="line.1161"></a> |
| <span class="sourceLineNo">1162</span> private ByteBuff cloneOnDiskBufferWithHeader() {<a name="line.1162"></a> |
| <span class="sourceLineNo">1163</span> expectState(State.BLOCK_READY);<a name="line.1163"></a> |
| <span class="sourceLineNo">1164</span> ByteBuff bytebuff = allocator.allocate(onDiskBlockBytesWithHeader.size());<a name="line.1164"></a> |
| <span class="sourceLineNo">1165</span> onDiskBlockBytesWithHeader.toByteBuff(bytebuff);<a name="line.1165"></a> |
| <span class="sourceLineNo">1166</span> bytebuff.rewind();<a name="line.1166"></a> |
| <span class="sourceLineNo">1167</span> return bytebuff;<a name="line.1167"></a> |
| <span class="sourceLineNo">1168</span> }<a name="line.1168"></a> |
| <span class="sourceLineNo">1169</span><a name="line.1169"></a> |
| <span class="sourceLineNo">1170</span> private void expectState(State expectedState) {<a name="line.1170"></a> |
| <span class="sourceLineNo">1171</span> if (state != expectedState) {<a name="line.1171"></a> |
| <span class="sourceLineNo">1172</span> throw new IllegalStateException(<a name="line.1172"></a> |
| <span class="sourceLineNo">1173</span> "Expected state: " + expectedState + ", actual state: " + state);<a name="line.1173"></a> |
| <span class="sourceLineNo">1174</span> }<a name="line.1174"></a> |
| <span class="sourceLineNo">1175</span> }<a name="line.1175"></a> |
| <span class="sourceLineNo">1176</span><a name="line.1176"></a> |
| <span class="sourceLineNo">1177</span> /**<a name="line.1177"></a> |
| <span class="sourceLineNo">1178</span> * Takes the given {@link BlockWritable} instance, creates a new block of its appropriate type,<a name="line.1178"></a> |
| <span class="sourceLineNo">1179</span> * writes the writable into this block, and flushes the block into the output stream. The writer<a name="line.1179"></a> |
| <span class="sourceLineNo">1180</span> * is instructed not to buffer uncompressed bytes for cache-on-write.<a name="line.1180"></a> |
| <span class="sourceLineNo">1181</span> * @param bw the block-writable object to write as a block<a name="line.1181"></a> |
| <span class="sourceLineNo">1182</span> * @param out the file system output stream<a name="line.1182"></a> |
| <span class="sourceLineNo">1183</span> */<a name="line.1183"></a> |
| <span class="sourceLineNo">1184</span> void writeBlock(BlockWritable bw, FSDataOutputStream out) throws IOException {<a name="line.1184"></a> |
| <span class="sourceLineNo">1185</span> bw.writeToBlock(startWriting(bw.getBlockType()));<a name="line.1185"></a> |
| <span class="sourceLineNo">1186</span> writeHeaderAndData(out);<a name="line.1186"></a> |
| <span class="sourceLineNo">1187</span> }<a name="line.1187"></a> |
| <span class="sourceLineNo">1188</span><a name="line.1188"></a> |
| <span class="sourceLineNo">1189</span> /**<a name="line.1189"></a> |
| <span class="sourceLineNo">1190</span> * Creates a new HFileBlock. Checksums have already been validated, so the byte buffer passed<a name="line.1190"></a> |
| <span class="sourceLineNo">1191</span> * into the constructor of this newly created block does not have checksum data even though the<a name="line.1191"></a> |
| <span class="sourceLineNo">1192</span> * header minor version is MINOR_VERSION_WITH_CHECKSUM. This is indicated by setting a 0 value<a name="line.1192"></a> |
| <span class="sourceLineNo">1193</span> * in bytesPerChecksum. This method copies the on-disk or uncompressed data to build the<a name="line.1193"></a> |
| <span class="sourceLineNo">1194</span> * HFileBlock which is used only while writing blocks and caching.<a name="line.1194"></a> |
| <span class="sourceLineNo">1195</span> * <p><a name="line.1195"></a> |
| <span class="sourceLineNo">1196</span> * TODO: Should there be an option where a cache can ask that hbase preserve block checksums for<a name="line.1196"></a> |
| <span class="sourceLineNo">1197</span> * checking after a block comes out of the cache? Otehrwise, cache is responsible for blocks<a name="line.1197"></a> |
| <span class="sourceLineNo">1198</span> * being wholesome (ECC memory or if file-backed, it does checksumming).<a name="line.1198"></a> |
| <span class="sourceLineNo">1199</span> */<a name="line.1199"></a> |
| <span class="sourceLineNo">1200</span> HFileBlock getBlockForCaching(CacheConfig cacheConf) {<a name="line.1200"></a> |
| <span class="sourceLineNo">1201</span> HFileContext newContext = new HFileContextBuilder().withBlockSize(fileContext.getBlocksize())<a name="line.1201"></a> |
| <span class="sourceLineNo">1202</span> .withBytesPerCheckSum(0).withChecksumType(ChecksumType.NULL) // no checksums in cached data<a name="line.1202"></a> |
| <span class="sourceLineNo">1203</span> .withCompression(fileContext.getCompression())<a name="line.1203"></a> |
| <span class="sourceLineNo">1204</span> .withDataBlockEncoding(fileContext.getDataBlockEncoding())<a name="line.1204"></a> |
| <span class="sourceLineNo">1205</span> .withHBaseCheckSum(fileContext.isUseHBaseChecksum())<a name="line.1205"></a> |
| <span class="sourceLineNo">1206</span> .withCompressTags(fileContext.isCompressTags())<a name="line.1206"></a> |
| <span class="sourceLineNo">1207</span> .withIncludesMvcc(fileContext.isIncludesMvcc())<a name="line.1207"></a> |
| <span class="sourceLineNo">1208</span> .withIncludesTags(fileContext.isIncludesTags())<a name="line.1208"></a> |
| <span class="sourceLineNo">1209</span> .withColumnFamily(fileContext.getColumnFamily()).withTableName(fileContext.getTableName())<a name="line.1209"></a> |
| <span class="sourceLineNo">1210</span> .build();<a name="line.1210"></a> |
| <span class="sourceLineNo">1211</span> // Build the HFileBlock.<a name="line.1211"></a> |
| <span class="sourceLineNo">1212</span> HFileBlockBuilder builder = new HFileBlockBuilder();<a name="line.1212"></a> |
| <span class="sourceLineNo">1213</span> ByteBuff buff;<a name="line.1213"></a> |
| <span class="sourceLineNo">1214</span> if (cacheConf.shouldCacheCompressed(blockType.getCategory())) {<a name="line.1214"></a> |
| <span class="sourceLineNo">1215</span> buff = cloneOnDiskBufferWithHeader();<a name="line.1215"></a> |
| <span class="sourceLineNo">1216</span> } else {<a name="line.1216"></a> |
| <span class="sourceLineNo">1217</span> buff = cloneUncompressedBufferWithHeader();<a name="line.1217"></a> |
| <span class="sourceLineNo">1218</span> }<a name="line.1218"></a> |
| <span class="sourceLineNo">1219</span> return builder.withBlockType(blockType)<a name="line.1219"></a> |
| <span class="sourceLineNo">1220</span> .withOnDiskSizeWithoutHeader(getOnDiskSizeWithoutHeader())<a name="line.1220"></a> |
| <span class="sourceLineNo">1221</span> .withUncompressedSizeWithoutHeader(getUncompressedSizeWithoutHeader())<a name="line.1221"></a> |
| <span class="sourceLineNo">1222</span> .withPrevBlockOffset(prevOffset).withByteBuff(buff).withFillHeader(FILL_HEADER)<a name="line.1222"></a> |
| <span class="sourceLineNo">1223</span> .withOffset(startOffset).withNextBlockOnDiskSize(UNSET)<a name="line.1223"></a> |
| <span class="sourceLineNo">1224</span> .withOnDiskDataSizeWithHeader(onDiskBlockBytesWithHeader.size() + onDiskChecksum.length)<a name="line.1224"></a> |
| <span class="sourceLineNo">1225</span> .withHFileContext(newContext).withByteBuffAllocator(cacheConf.getByteBuffAllocator())<a name="line.1225"></a> |
| <span class="sourceLineNo">1226</span> .withShared(!buff.hasArray()).build();<a name="line.1226"></a> |
| <span class="sourceLineNo">1227</span> }<a name="line.1227"></a> |
| <span class="sourceLineNo">1228</span> }<a name="line.1228"></a> |
| <span class="sourceLineNo">1229</span><a name="line.1229"></a> |
| <span class="sourceLineNo">1230</span> /** Something that can be written into a block. */<a name="line.1230"></a> |
| <span class="sourceLineNo">1231</span> interface BlockWritable {<a name="line.1231"></a> |
| <span class="sourceLineNo">1232</span> /** The type of block this data should use. */<a name="line.1232"></a> |
| <span class="sourceLineNo">1233</span> BlockType getBlockType();<a name="line.1233"></a> |
| <span class="sourceLineNo">1234</span><a name="line.1234"></a> |
| <span class="sourceLineNo">1235</span> /**<a name="line.1235"></a> |
| <span class="sourceLineNo">1236</span> * Writes the block to the provided stream. Must not write any magic records.<a name="line.1236"></a> |
| <span class="sourceLineNo">1237</span> * @param out a stream to write uncompressed data into<a name="line.1237"></a> |
| <span class="sourceLineNo">1238</span> */<a name="line.1238"></a> |
| <span class="sourceLineNo">1239</span> void writeToBlock(DataOutput out) throws IOException;<a name="line.1239"></a> |
| <span class="sourceLineNo">1240</span> }<a name="line.1240"></a> |
| <span class="sourceLineNo">1241</span><a name="line.1241"></a> |
| <span class="sourceLineNo">1242</span> /**<a name="line.1242"></a> |
| <span class="sourceLineNo">1243</span> * Iterator for reading {@link HFileBlock}s in load-on-open-section, such as root data index<a name="line.1243"></a> |
| <span class="sourceLineNo">1244</span> * block, meta index block, file info block etc.<a name="line.1244"></a> |
| <span class="sourceLineNo">1245</span> */<a name="line.1245"></a> |
| <span class="sourceLineNo">1246</span> interface BlockIterator {<a name="line.1246"></a> |
| <span class="sourceLineNo">1247</span> /**<a name="line.1247"></a> |
| <span class="sourceLineNo">1248</span> * Get the next block, or null if there are no more blocks to iterate.<a name="line.1248"></a> |
| <span class="sourceLineNo">1249</span> */<a name="line.1249"></a> |
| <span class="sourceLineNo">1250</span> HFileBlock nextBlock() throws IOException;<a name="line.1250"></a> |
| <span class="sourceLineNo">1251</span><a name="line.1251"></a> |
| <span class="sourceLineNo">1252</span> /**<a name="line.1252"></a> |
| <span class="sourceLineNo">1253</span> * Similar to {@link #nextBlock()} but checks block type, throws an exception if incorrect, and<a name="line.1253"></a> |
| <span class="sourceLineNo">1254</span> * returns the HFile block<a name="line.1254"></a> |
| <span class="sourceLineNo">1255</span> */<a name="line.1255"></a> |
| <span class="sourceLineNo">1256</span> HFileBlock nextBlockWithBlockType(BlockType blockType) throws IOException;<a name="line.1256"></a> |
| <span class="sourceLineNo">1257</span><a name="line.1257"></a> |
| <span class="sourceLineNo">1258</span> /**<a name="line.1258"></a> |
| <span class="sourceLineNo">1259</span> * Now we use the {@link ByteBuffAllocator} to manage the nio ByteBuffers for HFileBlocks, so we<a name="line.1259"></a> |
| <span class="sourceLineNo">1260</span> * must deallocate all of the ByteBuffers in the end life. the BlockIterator's life cycle is<a name="line.1260"></a> |
| <span class="sourceLineNo">1261</span> * starting from opening an HFileReader and stopped when the HFileReader#close, so we will keep<a name="line.1261"></a> |
| <span class="sourceLineNo">1262</span> * track all the read blocks until we call {@link BlockIterator#freeBlocks()} when closing the<a name="line.1262"></a> |
| <span class="sourceLineNo">1263</span> * HFileReader. Sum bytes of those blocks in load-on-open section should be quite small, so<a name="line.1263"></a> |
| <span class="sourceLineNo">1264</span> * tracking them should be OK.<a name="line.1264"></a> |
| <span class="sourceLineNo">1265</span> */<a name="line.1265"></a> |
| <span class="sourceLineNo">1266</span> void freeBlocks();<a name="line.1266"></a> |
| <span class="sourceLineNo">1267</span> }<a name="line.1267"></a> |
| <span class="sourceLineNo">1268</span><a name="line.1268"></a> |
| <span class="sourceLineNo">1269</span> /** An HFile block reader with iteration ability. */<a name="line.1269"></a> |
| <span class="sourceLineNo">1270</span> interface FSReader {<a name="line.1270"></a> |
| <span class="sourceLineNo">1271</span> /**<a name="line.1271"></a> |
| <span class="sourceLineNo">1272</span> * Reads the block at the given offset in the file with the given on-disk size and uncompressed<a name="line.1272"></a> |
| <span class="sourceLineNo">1273</span> * size.<a name="line.1273"></a> |
| <span class="sourceLineNo">1274</span> * @param offset of the file to read<a name="line.1274"></a> |
| <span class="sourceLineNo">1275</span> * @param onDiskSize the on-disk size of the entire block, including all applicable headers,<a name="line.1275"></a> |
| <span class="sourceLineNo">1276</span> * or -1 if unknown<a name="line.1276"></a> |
| <span class="sourceLineNo">1277</span> * @param pread true to use pread, otherwise use the stream read.<a name="line.1277"></a> |
| <span class="sourceLineNo">1278</span> * @param updateMetrics update the metrics or not.<a name="line.1278"></a> |
| <span class="sourceLineNo">1279</span> * @param intoHeap allocate the block's ByteBuff by {@link ByteBuffAllocator} or JVM heap.<a name="line.1279"></a> |
| <span class="sourceLineNo">1280</span> * For LRUBlockCache, we must ensure that the block to cache is an heap<a name="line.1280"></a> |
| <span class="sourceLineNo">1281</span> * one, because the memory occupation is based on heap now, also for<a name="line.1281"></a> |
| <span class="sourceLineNo">1282</span> * {@link CombinedBlockCache}, we use the heap LRUBlockCache as L1 cache to<a name="line.1282"></a> |
| <span class="sourceLineNo">1283</span> * cache small blocks such as IndexBlock or MetaBlock for faster access. So<a name="line.1283"></a> |
| <span class="sourceLineNo">1284</span> * introduce an flag here to decide whether allocate from JVM heap or not<a name="line.1284"></a> |
| <span class="sourceLineNo">1285</span> * so that we can avoid an extra off-heap to heap memory copy when using<a name="line.1285"></a> |
| <span class="sourceLineNo">1286</span> * LRUBlockCache. For most cases, we known what's the expected block type<a name="line.1286"></a> |
| <span class="sourceLineNo">1287</span> * we'll read, while for some special case (Example:<a name="line.1287"></a> |
| <span class="sourceLineNo">1288</span> * HFileReaderImpl#readNextDataBlock()), we cannot pre-decide what's the<a name="line.1288"></a> |
| <span class="sourceLineNo">1289</span> * expected block type, then we can only allocate block's ByteBuff from<a name="line.1289"></a> |
| <span class="sourceLineNo">1290</span> * {@link ByteBuffAllocator} firstly, and then when caching it in<a name="line.1290"></a> |
| <span class="sourceLineNo">1291</span> * {@link LruBlockCache} we'll check whether the ByteBuff is from heap or<a name="line.1291"></a> |
| <span class="sourceLineNo">1292</span> * not, if not then we'll clone it to an heap one and cache it.<a name="line.1292"></a> |
| <span class="sourceLineNo">1293</span> * @return the newly read block<a name="line.1293"></a> |
| <span class="sourceLineNo">1294</span> */<a name="line.1294"></a> |
| <span class="sourceLineNo">1295</span> HFileBlock readBlockData(long offset, long onDiskSize, boolean pread, boolean updateMetrics,<a name="line.1295"></a> |
| <span class="sourceLineNo">1296</span> boolean intoHeap) throws IOException;<a name="line.1296"></a> |
| <span class="sourceLineNo">1297</span><a name="line.1297"></a> |
| <span class="sourceLineNo">1298</span> /**<a name="line.1298"></a> |
| <span class="sourceLineNo">1299</span> * Creates a block iterator over the given portion of the {@link HFile}. The iterator returns<a name="line.1299"></a> |
| <span class="sourceLineNo">1300</span> * blocks starting with offset such that offset &lt;= startOffset &lt; endOffset. Returned<a name="line.1300"></a> |
| <span class="sourceLineNo">1301</span> * blocks are always unpacked. Used when no hfile index available; e.g. reading in the hfile<a name="line.1301"></a> |
| <span class="sourceLineNo">1302</span> * index blocks themselves on file open.<a name="line.1302"></a> |
| <span class="sourceLineNo">1303</span> * @param startOffset the offset of the block to start iteration with<a name="line.1303"></a> |
| <span class="sourceLineNo">1304</span> * @param endOffset the offset to end iteration at (exclusive)<a name="line.1304"></a> |
| <span class="sourceLineNo">1305</span> * @return an iterator of blocks between the two given offsets<a name="line.1305"></a> |
| <span class="sourceLineNo">1306</span> */<a name="line.1306"></a> |
| <span class="sourceLineNo">1307</span> BlockIterator blockRange(long startOffset, long endOffset);<a name="line.1307"></a> |
| <span class="sourceLineNo">1308</span><a name="line.1308"></a> |
| <span class="sourceLineNo">1309</span> /** Closes the backing streams */<a name="line.1309"></a> |
| <span class="sourceLineNo">1310</span> void closeStreams() throws IOException;<a name="line.1310"></a> |
| <span class="sourceLineNo">1311</span><a name="line.1311"></a> |
| <span class="sourceLineNo">1312</span> /** Get a decoder for {@link BlockType#ENCODED_DATA} blocks from this file. */<a name="line.1312"></a> |
| <span class="sourceLineNo">1313</span> HFileBlockDecodingContext getBlockDecodingContext();<a name="line.1313"></a> |
| <span class="sourceLineNo">1314</span><a name="line.1314"></a> |
| <span class="sourceLineNo">1315</span> /** Get the default decoder for blocks from this file. */<a name="line.1315"></a> |
| <span class="sourceLineNo">1316</span> HFileBlockDecodingContext getDefaultBlockDecodingContext();<a name="line.1316"></a> |
| <span class="sourceLineNo">1317</span><a name="line.1317"></a> |
| <span class="sourceLineNo">1318</span> void setIncludesMemStoreTS(boolean includesMemstoreTS);<a name="line.1318"></a> |
| <span class="sourceLineNo">1319</span><a name="line.1319"></a> |
| <span class="sourceLineNo">1320</span> void setDataBlockEncoder(HFileDataBlockEncoder encoder, Configuration conf);<a name="line.1320"></a> |
| <span class="sourceLineNo">1321</span><a name="line.1321"></a> |
| <span class="sourceLineNo">1322</span> /**<a name="line.1322"></a> |
| <span class="sourceLineNo">1323</span> * To close the stream's socket. Note: This can be concurrently called from multiple threads and<a name="line.1323"></a> |
| <span class="sourceLineNo">1324</span> * implementation should take care of thread safety.<a name="line.1324"></a> |
| <span class="sourceLineNo">1325</span> */<a name="line.1325"></a> |
| <span class="sourceLineNo">1326</span> void unbufferStream();<a name="line.1326"></a> |
| <span class="sourceLineNo">1327</span> }<a name="line.1327"></a> |
| <span class="sourceLineNo">1328</span><a name="line.1328"></a> |
| <span class="sourceLineNo">1329</span> /**<a name="line.1329"></a> |
| <span class="sourceLineNo">1330</span> * Data-structure to use caching the header of the NEXT block. Only works if next read that comes<a name="line.1330"></a> |
| <span class="sourceLineNo">1331</span> * in here is next in sequence in this block. When we read, we read current block and the next<a name="line.1331"></a> |
| <span class="sourceLineNo">1332</span> * blocks' header. We do this so we have the length of the next block to read if the hfile index<a name="line.1332"></a> |
| <span class="sourceLineNo">1333</span> * is not available (rare, at hfile open only).<a name="line.1333"></a> |
| <span class="sourceLineNo">1334</span> */<a name="line.1334"></a> |
| <span class="sourceLineNo">1335</span> private static class PrefetchedHeader {<a name="line.1335"></a> |
| <span class="sourceLineNo">1336</span> long offset = -1;<a name="line.1336"></a> |
| <span class="sourceLineNo">1337</span> byte[] header = new byte[HConstants.HFILEBLOCK_HEADER_SIZE];<a name="line.1337"></a> |
| <span class="sourceLineNo">1338</span> final ByteBuff buf = new SingleByteBuff(ByteBuffer.wrap(header, 0, header.length));<a name="line.1338"></a> |
| <span class="sourceLineNo">1339</span><a name="line.1339"></a> |
| <span class="sourceLineNo">1340</span> @Override<a name="line.1340"></a> |
| <span class="sourceLineNo">1341</span> public String toString() {<a name="line.1341"></a> |
| <span class="sourceLineNo">1342</span> return "offset=" + this.offset + ", header=" + Bytes.toStringBinary(header);<a name="line.1342"></a> |
| <span class="sourceLineNo">1343</span> }<a name="line.1343"></a> |
| <span class="sourceLineNo">1344</span> }<a name="line.1344"></a> |
| <span class="sourceLineNo">1345</span><a name="line.1345"></a> |
| <span class="sourceLineNo">1346</span> /**<a name="line.1346"></a> |
| <span class="sourceLineNo">1347</span> * Reads version 2 HFile blocks from the filesystem.<a name="line.1347"></a> |
| <span class="sourceLineNo">1348</span> */<a name="line.1348"></a> |
| <span class="sourceLineNo">1349</span> static class FSReaderImpl implements FSReader {<a name="line.1349"></a> |
| <span class="sourceLineNo">1350</span> /**<a name="line.1350"></a> |
| <span class="sourceLineNo">1351</span> * The file system stream of the underlying {@link HFile} that does or doesn't do checksum<a name="line.1351"></a> |
| <span class="sourceLineNo">1352</span> * validations in the filesystem<a name="line.1352"></a> |
| <span class="sourceLineNo">1353</span> */<a name="line.1353"></a> |
| <span class="sourceLineNo">1354</span> private FSDataInputStreamWrapper streamWrapper;<a name="line.1354"></a> |
| <span class="sourceLineNo">1355</span><a name="line.1355"></a> |
| <span class="sourceLineNo">1356</span> private HFileBlockDecodingContext encodedBlockDecodingCtx;<a name="line.1356"></a> |
| <span class="sourceLineNo">1357</span><a name="line.1357"></a> |
| <span class="sourceLineNo">1358</span> /** Default context used when BlockType != {@link BlockType#ENCODED_DATA}. */<a name="line.1358"></a> |
| <span class="sourceLineNo">1359</span> private final HFileBlockDefaultDecodingContext defaultDecodingCtx;<a name="line.1359"></a> |
| <span class="sourceLineNo">1360</span><a name="line.1360"></a> |
| <span class="sourceLineNo">1361</span> /**<a name="line.1361"></a> |
| <span class="sourceLineNo">1362</span> * Cache of the NEXT header after this. Check it is indeed next blocks header before using it.<a name="line.1362"></a> |
| <span class="sourceLineNo">1363</span> * TODO: Review. This overread into next block to fetch next blocks header seems unnecessary<a name="line.1363"></a> |
| <span class="sourceLineNo">1364</span> * given we usually get the block size from the hfile index. Review!<a name="line.1364"></a> |
| <span class="sourceLineNo">1365</span> */<a name="line.1365"></a> |
| <span class="sourceLineNo">1366</span> private AtomicReference<PrefetchedHeader> prefetchedHeader =<a name="line.1366"></a> |
| <span class="sourceLineNo">1367</span> new AtomicReference<>(new PrefetchedHeader());<a name="line.1367"></a> |
| <span class="sourceLineNo">1368</span><a name="line.1368"></a> |
| <span class="sourceLineNo">1369</span> /** The size of the file we are reading from, or -1 if unknown. */<a name="line.1369"></a> |
| <span class="sourceLineNo">1370</span> private long fileSize;<a name="line.1370"></a> |
| <span class="sourceLineNo">1371</span><a name="line.1371"></a> |
| <span class="sourceLineNo">1372</span> /** The size of the header */<a name="line.1372"></a> |
| <span class="sourceLineNo">1373</span> protected final int hdrSize;<a name="line.1373"></a> |
| <span class="sourceLineNo">1374</span><a name="line.1374"></a> |
| <span class="sourceLineNo">1375</span> /** The filesystem used to access data */<a name="line.1375"></a> |
| <span class="sourceLineNo">1376</span> private HFileSystem hfs;<a name="line.1376"></a> |
| <span class="sourceLineNo">1377</span><a name="line.1377"></a> |
| <span class="sourceLineNo">1378</span> private HFileContext fileContext;<a name="line.1378"></a> |
| <span class="sourceLineNo">1379</span> // Cache the fileName<a name="line.1379"></a> |
| <span class="sourceLineNo">1380</span> private String pathName;<a name="line.1380"></a> |
| <span class="sourceLineNo">1381</span><a name="line.1381"></a> |
| <span class="sourceLineNo">1382</span> private final ByteBuffAllocator allocator;<a name="line.1382"></a> |
| <span class="sourceLineNo">1383</span><a name="line.1383"></a> |
| <span class="sourceLineNo">1384</span> private final Lock streamLock = new ReentrantLock();<a name="line.1384"></a> |
| <span class="sourceLineNo">1385</span><a name="line.1385"></a> |
| <span class="sourceLineNo">1386</span> private final boolean isPreadAllBytes;<a name="line.1386"></a> |
| <span class="sourceLineNo">1387</span><a name="line.1387"></a> |
| <span class="sourceLineNo">1388</span> FSReaderImpl(ReaderContext readerContext, HFileContext fileContext, ByteBuffAllocator allocator,<a name="line.1388"></a> |
| <span class="sourceLineNo">1389</span> Configuration conf) throws IOException {<a name="line.1389"></a> |
| <span class="sourceLineNo">1390</span> this.fileSize = readerContext.getFileSize();<a name="line.1390"></a> |
| <span class="sourceLineNo">1391</span> this.hfs = readerContext.getFileSystem();<a name="line.1391"></a> |
| <span class="sourceLineNo">1392</span> if (readerContext.getFilePath() != null) {<a name="line.1392"></a> |
| <span class="sourceLineNo">1393</span> this.pathName = readerContext.getFilePath().toString();<a name="line.1393"></a> |
| <span class="sourceLineNo">1394</span> }<a name="line.1394"></a> |
| <span class="sourceLineNo">1395</span> this.fileContext = fileContext;<a name="line.1395"></a> |
| <span class="sourceLineNo">1396</span> this.hdrSize = headerSize(fileContext.isUseHBaseChecksum());<a name="line.1396"></a> |
| <span class="sourceLineNo">1397</span> this.allocator = allocator;<a name="line.1397"></a> |
| <span class="sourceLineNo">1398</span><a name="line.1398"></a> |
| <span class="sourceLineNo">1399</span> this.streamWrapper = readerContext.getInputStreamWrapper();<a name="line.1399"></a> |
| <span class="sourceLineNo">1400</span> // Older versions of HBase didn't support checksum.<a name="line.1400"></a> |
| <span class="sourceLineNo">1401</span> this.streamWrapper.prepareForBlockReader(!fileContext.isUseHBaseChecksum());<a name="line.1401"></a> |
| <span class="sourceLineNo">1402</span> defaultDecodingCtx = new HFileBlockDefaultDecodingContext(conf, fileContext);<a name="line.1402"></a> |
| <span class="sourceLineNo">1403</span> encodedBlockDecodingCtx = defaultDecodingCtx;<a name="line.1403"></a> |
| <span class="sourceLineNo">1404</span> isPreadAllBytes = readerContext.isPreadAllBytes();<a name="line.1404"></a> |
| <span class="sourceLineNo">1405</span> }<a name="line.1405"></a> |
| <span class="sourceLineNo">1406</span><a name="line.1406"></a> |
| <span class="sourceLineNo">1407</span> @Override<a name="line.1407"></a> |
| <span class="sourceLineNo">1408</span> public BlockIterator blockRange(final long startOffset, final long endOffset) {<a name="line.1408"></a> |
| <span class="sourceLineNo">1409</span> final FSReader owner = this; // handle for inner class<a name="line.1409"></a> |
| <span class="sourceLineNo">1410</span> return new BlockIterator() {<a name="line.1410"></a> |
| <span class="sourceLineNo">1411</span> private volatile boolean freed = false;<a name="line.1411"></a> |
| <span class="sourceLineNo">1412</span> // Tracking all read blocks until we call freeBlocks.<a name="line.1412"></a> |
| <span class="sourceLineNo">1413</span> private List<HFileBlock> blockTracker = new ArrayList<>();<a name="line.1413"></a> |
| <span class="sourceLineNo">1414</span> private long offset = startOffset;<a name="line.1414"></a> |
| <span class="sourceLineNo">1415</span> // Cache length of next block. Current block has the length of next block in it.<a name="line.1415"></a> |
| <span class="sourceLineNo">1416</span> private long length = -1;<a name="line.1416"></a> |
| <span class="sourceLineNo">1417</span><a name="line.1417"></a> |
| <span class="sourceLineNo">1418</span> @Override<a name="line.1418"></a> |
| <span class="sourceLineNo">1419</span> public HFileBlock nextBlock() throws IOException {<a name="line.1419"></a> |
| <span class="sourceLineNo">1420</span> if (offset >= endOffset) {<a name="line.1420"></a> |
| <span class="sourceLineNo">1421</span> return null;<a name="line.1421"></a> |
| <span class="sourceLineNo">1422</span> }<a name="line.1422"></a> |
| <span class="sourceLineNo">1423</span> HFileBlock b = readBlockData(offset, length, false, false, true);<a name="line.1423"></a> |
| <span class="sourceLineNo">1424</span> offset += b.getOnDiskSizeWithHeader();<a name="line.1424"></a> |
| <span class="sourceLineNo">1425</span> length = b.getNextBlockOnDiskSize();<a name="line.1425"></a> |
| <span class="sourceLineNo">1426</span> HFileBlock uncompressed = b.unpack(fileContext, owner);<a name="line.1426"></a> |
| <span class="sourceLineNo">1427</span> if (uncompressed != b) {<a name="line.1427"></a> |
| <span class="sourceLineNo">1428</span> b.release(); // Need to release the compressed Block now.<a name="line.1428"></a> |
| <span class="sourceLineNo">1429</span> }<a name="line.1429"></a> |
| <span class="sourceLineNo">1430</span> blockTracker.add(uncompressed);<a name="line.1430"></a> |
| <span class="sourceLineNo">1431</span> return uncompressed;<a name="line.1431"></a> |
| <span class="sourceLineNo">1432</span> }<a name="line.1432"></a> |
| <span class="sourceLineNo">1433</span><a name="line.1433"></a> |
| <span class="sourceLineNo">1434</span> @Override<a name="line.1434"></a> |
| <span class="sourceLineNo">1435</span> public HFileBlock nextBlockWithBlockType(BlockType blockType) throws IOException {<a name="line.1435"></a> |
| <span class="sourceLineNo">1436</span> HFileBlock blk = nextBlock();<a name="line.1436"></a> |
| <span class="sourceLineNo">1437</span> if (blk.getBlockType() != blockType) {<a name="line.1437"></a> |
| <span class="sourceLineNo">1438</span> throw new IOException(<a name="line.1438"></a> |
| <span class="sourceLineNo">1439</span> "Expected block of type " + blockType + " but found " + blk.getBlockType());<a name="line.1439"></a> |
| <span class="sourceLineNo">1440</span> }<a name="line.1440"></a> |
| <span class="sourceLineNo">1441</span> return blk;<a name="line.1441"></a> |
| <span class="sourceLineNo">1442</span> }<a name="line.1442"></a> |
| <span class="sourceLineNo">1443</span><a name="line.1443"></a> |
| <span class="sourceLineNo">1444</span> @Override<a name="line.1444"></a> |
| <span class="sourceLineNo">1445</span> public void freeBlocks() {<a name="line.1445"></a> |
| <span class="sourceLineNo">1446</span> if (freed) {<a name="line.1446"></a> |
| <span class="sourceLineNo">1447</span> return;<a name="line.1447"></a> |
| <span class="sourceLineNo">1448</span> }<a name="line.1448"></a> |
| <span class="sourceLineNo">1449</span> blockTracker.forEach(HFileBlock::release);<a name="line.1449"></a> |
| <span class="sourceLineNo">1450</span> blockTracker = null;<a name="line.1450"></a> |
| <span class="sourceLineNo">1451</span> freed = true;<a name="line.1451"></a> |
| <span class="sourceLineNo">1452</span> }<a name="line.1452"></a> |
| <span class="sourceLineNo">1453</span> };<a name="line.1453"></a> |
| <span class="sourceLineNo">1454</span> }<a name="line.1454"></a> |
| <span class="sourceLineNo">1455</span><a name="line.1455"></a> |
| <span class="sourceLineNo">1456</span> /**<a name="line.1456"></a> |
| <span class="sourceLineNo">1457</span> * Does a positional read or a seek and read into the given byte buffer. We need take care that<a name="line.1457"></a> |
| <span class="sourceLineNo">1458</span> * we will call the {@link ByteBuff#release()} for every exit to deallocate the ByteBuffers,<a name="line.1458"></a> |
| <span class="sourceLineNo">1459</span> * otherwise the memory leak may happen.<a name="line.1459"></a> |
| <span class="sourceLineNo">1460</span> * @param dest destination buffer<a name="line.1460"></a> |
| <span class="sourceLineNo">1461</span> * @param size size of read<a name="line.1461"></a> |
| <span class="sourceLineNo">1462</span> * @param peekIntoNextBlock whether to read the next block's on-disk size<a name="line.1462"></a> |
| <span class="sourceLineNo">1463</span> * @param fileOffset position in the stream to read at<a name="line.1463"></a> |
| <span class="sourceLineNo">1464</span> * @param pread whether we should do a positional read<a name="line.1464"></a> |
| <span class="sourceLineNo">1465</span> * @param istream The input source of data<a name="line.1465"></a> |
| <span class="sourceLineNo">1466</span> * @return true to indicate the destination buffer include the next block header, otherwise only<a name="line.1466"></a> |
| <span class="sourceLineNo">1467</span> * include the current block data without the next block header.<a name="line.1467"></a> |
| <span class="sourceLineNo">1468</span> * @throws IOException if any IO error happen.<a name="line.1468"></a> |
| <span class="sourceLineNo">1469</span> */<a name="line.1469"></a> |
| <span class="sourceLineNo">1470</span> protected boolean readAtOffset(FSDataInputStream istream, ByteBuff dest, int size,<a name="line.1470"></a> |
| <span class="sourceLineNo">1471</span> boolean peekIntoNextBlock, long fileOffset, boolean pread) throws IOException {<a name="line.1471"></a> |
| <span class="sourceLineNo">1472</span> if (!pread) {<a name="line.1472"></a> |
| <span class="sourceLineNo">1473</span> // Seek + read. Better for scanning.<a name="line.1473"></a> |
| <span class="sourceLineNo">1474</span> istream.seek(fileOffset);<a name="line.1474"></a> |
| <span class="sourceLineNo">1475</span> long realOffset = istream.getPos();<a name="line.1475"></a> |
| <span class="sourceLineNo">1476</span> if (realOffset != fileOffset) {<a name="line.1476"></a> |
| <span class="sourceLineNo">1477</span> throw new IOException("Tried to seek to " + fileOffset + " to read " + size<a name="line.1477"></a> |
| <span class="sourceLineNo">1478</span> + " bytes, but pos=" + realOffset + " after seek");<a name="line.1478"></a> |
| <span class="sourceLineNo">1479</span> }<a name="line.1479"></a> |
| <span class="sourceLineNo">1480</span> if (!peekIntoNextBlock) {<a name="line.1480"></a> |
| <span class="sourceLineNo">1481</span> BlockIOUtils.readFully(dest, istream, size);<a name="line.1481"></a> |
| <span class="sourceLineNo">1482</span> return false;<a name="line.1482"></a> |
| <span class="sourceLineNo">1483</span> }<a name="line.1483"></a> |
| <span class="sourceLineNo">1484</span><a name="line.1484"></a> |
| <span class="sourceLineNo">1485</span> // Try to read the next block header<a name="line.1485"></a> |
| <span class="sourceLineNo">1486</span> if (!BlockIOUtils.readWithExtra(dest, istream, size, hdrSize)) {<a name="line.1486"></a> |
| <span class="sourceLineNo">1487</span> // did not read the next block header.<a name="line.1487"></a> |
| <span class="sourceLineNo">1488</span> return false;<a name="line.1488"></a> |
| <span class="sourceLineNo">1489</span> }<a name="line.1489"></a> |
| <span class="sourceLineNo">1490</span> } else {<a name="line.1490"></a> |
| <span class="sourceLineNo">1491</span> // Positional read. Better for random reads; or when the streamLock is already locked.<a name="line.1491"></a> |
| <span class="sourceLineNo">1492</span> int extraSize = peekIntoNextBlock ? hdrSize : 0;<a name="line.1492"></a> |
| <span class="sourceLineNo">1493</span> if (<a name="line.1493"></a> |
| <span class="sourceLineNo">1494</span> !BlockIOUtils.preadWithExtra(dest, istream, fileOffset, size, extraSize, isPreadAllBytes)<a name="line.1494"></a> |
| <span class="sourceLineNo">1495</span> ) {<a name="line.1495"></a> |
| <span class="sourceLineNo">1496</span> // did not read the next block header.<a name="line.1496"></a> |
| <span class="sourceLineNo">1497</span> return false;<a name="line.1497"></a> |
| <span class="sourceLineNo">1498</span> }<a name="line.1498"></a> |
| <span class="sourceLineNo">1499</span> }<a name="line.1499"></a> |
| <span class="sourceLineNo">1500</span> assert peekIntoNextBlock;<a name="line.1500"></a> |
| <span class="sourceLineNo">1501</span> return true;<a name="line.1501"></a> |
| <span class="sourceLineNo">1502</span> }<a name="line.1502"></a> |
| <span class="sourceLineNo">1503</span><a name="line.1503"></a> |
| <span class="sourceLineNo">1504</span> /**<a name="line.1504"></a> |
| <span class="sourceLineNo">1505</span> * Reads a version 2 block (version 1 blocks not supported and not expected). Tries to do as<a name="line.1505"></a> |
| <span class="sourceLineNo">1506</span> * little memory allocation as possible, using the provided on-disk size.<a name="line.1506"></a> |
| <span class="sourceLineNo">1507</span> * @param offset the offset in the stream to read at<a name="line.1507"></a> |
| <span class="sourceLineNo">1508</span> * @param onDiskSizeWithHeaderL the on-disk size of the block, including the header, or -1 if<a name="line.1508"></a> |
| <span class="sourceLineNo">1509</span> * unknown; i.e. when iterating over blocks reading in the file<a name="line.1509"></a> |
| <span class="sourceLineNo">1510</span> * metadata info.<a name="line.1510"></a> |
| <span class="sourceLineNo">1511</span> * @param pread whether to use a positional read<a name="line.1511"></a> |
| <span class="sourceLineNo">1512</span> * @param updateMetrics whether to update the metrics<a name="line.1512"></a> |
| <span class="sourceLineNo">1513</span> * @param intoHeap allocate ByteBuff of block from heap or off-heap.<a name="line.1513"></a> |
| <span class="sourceLineNo">1514</span> * @see FSReader#readBlockData(long, long, boolean, boolean, boolean) for more details about the<a name="line.1514"></a> |
| <span class="sourceLineNo">1515</span> * useHeap.<a name="line.1515"></a> |
| <span class="sourceLineNo">1516</span> */<a name="line.1516"></a> |
| <span class="sourceLineNo">1517</span> @Override<a name="line.1517"></a> |
| <span class="sourceLineNo">1518</span> public HFileBlock readBlockData(long offset, long onDiskSizeWithHeaderL, boolean pread,<a name="line.1518"></a> |
| <span class="sourceLineNo">1519</span> boolean updateMetrics, boolean intoHeap) throws IOException {<a name="line.1519"></a> |
| <span class="sourceLineNo">1520</span> // Get a copy of the current state of whether to validate<a name="line.1520"></a> |
| <span class="sourceLineNo">1521</span> // hbase checksums or not for this read call. This is not<a name="line.1521"></a> |
| <span class="sourceLineNo">1522</span> // thread-safe but the one constraint is that if we decide<a name="line.1522"></a> |
| <span class="sourceLineNo">1523</span> // to skip hbase checksum verification then we are<a name="line.1523"></a> |
| <span class="sourceLineNo">1524</span> // guaranteed to use hdfs checksum verification.<a name="line.1524"></a> |
| <span class="sourceLineNo">1525</span> boolean doVerificationThruHBaseChecksum = streamWrapper.shouldUseHBaseChecksum();<a name="line.1525"></a> |
| <span class="sourceLineNo">1526</span> FSDataInputStream is = streamWrapper.getStream(doVerificationThruHBaseChecksum);<a name="line.1526"></a> |
| <span class="sourceLineNo">1527</span> final Context context = Context.current().with(CONTEXT_KEY,<a name="line.1527"></a> |
| <span class="sourceLineNo">1528</span> new HFileContextAttributesBuilderConsumer(fileContext)<a name="line.1528"></a> |
| <span class="sourceLineNo">1529</span> .setSkipChecksum(doVerificationThruHBaseChecksum)<a name="line.1529"></a> |
| <span class="sourceLineNo">1530</span> .setReadType(pread ? ReadType.POSITIONAL_READ : ReadType.SEEK_PLUS_READ));<a name="line.1530"></a> |
| <span class="sourceLineNo">1531</span> try (Scope ignored = context.makeCurrent()) {<a name="line.1531"></a> |
| <span class="sourceLineNo">1532</span> HFileBlock blk = readBlockDataInternal(is, offset, onDiskSizeWithHeaderL, pread,<a name="line.1532"></a> |
| <span class="sourceLineNo">1533</span> doVerificationThruHBaseChecksum, updateMetrics, intoHeap);<a name="line.1533"></a> |
| <span class="sourceLineNo">1534</span> if (blk == null) {<a name="line.1534"></a> |
| <span class="sourceLineNo">1535</span> HFile.LOG.warn("HBase checksum verification failed for file {} at offset {} filesize {}."<a name="line.1535"></a> |
| <span class="sourceLineNo">1536</span> + " Retrying read with HDFS checksums turned on...", pathName, offset, fileSize);<a name="line.1536"></a> |
| <span class="sourceLineNo">1537</span><a name="line.1537"></a> |
| <span class="sourceLineNo">1538</span> if (!doVerificationThruHBaseChecksum) {<a name="line.1538"></a> |
| <span class="sourceLineNo">1539</span> String msg = "HBase checksum verification failed for file " + pathName + " at offset "<a name="line.1539"></a> |
| <span class="sourceLineNo">1540</span> + offset + " filesize " + fileSize + " but this cannot happen because doVerify is "<a name="line.1540"></a> |
| <span class="sourceLineNo">1541</span> + doVerificationThruHBaseChecksum;<a name="line.1541"></a> |
| <span class="sourceLineNo">1542</span> HFile.LOG.warn(msg);<a name="line.1542"></a> |
| <span class="sourceLineNo">1543</span> throw new IOException(msg); // cannot happen case here<a name="line.1543"></a> |
| <span class="sourceLineNo">1544</span> }<a name="line.1544"></a> |
| <span class="sourceLineNo">1545</span> HFile.CHECKSUM_FAILURES.increment(); // update metrics<a name="line.1545"></a> |
| <span class="sourceLineNo">1546</span><a name="line.1546"></a> |
| <span class="sourceLineNo">1547</span> // If we have a checksum failure, we fall back into a mode where<a name="line.1547"></a> |
| <span class="sourceLineNo">1548</span> // the next few reads use HDFS level checksums. We aim to make the<a name="line.1548"></a> |
| <span class="sourceLineNo">1549</span> // next CHECKSUM_VERIFICATION_NUM_IO_THRESHOLD reads avoid<a name="line.1549"></a> |
| <span class="sourceLineNo">1550</span> // hbase checksum verification, but since this value is set without<a name="line.1550"></a> |
| <span class="sourceLineNo">1551</span> // holding any locks, it can so happen that we might actually do<a name="line.1551"></a> |
| <span class="sourceLineNo">1552</span> // a few more than precisely this number.<a name="line.1552"></a> |
| <span class="sourceLineNo">1553</span> is = this.streamWrapper.fallbackToFsChecksum(CHECKSUM_VERIFICATION_NUM_IO_THRESHOLD);<a name="line.1553"></a> |
| <span class="sourceLineNo">1554</span> doVerificationThruHBaseChecksum = false;<a name="line.1554"></a> |
| <span class="sourceLineNo">1555</span> blk = readBlockDataInternal(is, offset, onDiskSizeWithHeaderL, pread,<a name="line.1555"></a> |
| <span class="sourceLineNo">1556</span> doVerificationThruHBaseChecksum, updateMetrics, intoHeap);<a name="line.1556"></a> |
| <span class="sourceLineNo">1557</span> if (blk != null) {<a name="line.1557"></a> |
| <span class="sourceLineNo">1558</span> HFile.LOG.warn(<a name="line.1558"></a> |
| <span class="sourceLineNo">1559</span> "HDFS checksum verification succeeded for file {} at offset {} filesize" + " {}",<a name="line.1559"></a> |
| <span class="sourceLineNo">1560</span> pathName, offset, fileSize);<a name="line.1560"></a> |
| <span class="sourceLineNo">1561</span> }<a name="line.1561"></a> |
| <span class="sourceLineNo">1562</span> }<a name="line.1562"></a> |
| <span class="sourceLineNo">1563</span> if (blk == null && !doVerificationThruHBaseChecksum) {<a name="line.1563"></a> |
| <span class="sourceLineNo">1564</span> String msg =<a name="line.1564"></a> |
| <span class="sourceLineNo">1565</span> "readBlockData failed, possibly due to " + "checksum verification failed for file "<a name="line.1565"></a> |
| <span class="sourceLineNo">1566</span> + pathName + " at offset " + offset + " filesize " + fileSize;<a name="line.1566"></a> |
| <span class="sourceLineNo">1567</span> HFile.LOG.warn(msg);<a name="line.1567"></a> |
| <span class="sourceLineNo">1568</span> throw new IOException(msg);<a name="line.1568"></a> |
| <span class="sourceLineNo">1569</span> }<a name="line.1569"></a> |
| <span class="sourceLineNo">1570</span><a name="line.1570"></a> |
| <span class="sourceLineNo">1571</span> // If there is a checksum mismatch earlier, then retry with<a name="line.1571"></a> |
| <span class="sourceLineNo">1572</span> // HBase checksums switched off and use HDFS checksum verification.<a name="line.1572"></a> |
| <span class="sourceLineNo">1573</span> // This triggers HDFS to detect and fix corrupt replicas. The<a name="line.1573"></a> |
| <span class="sourceLineNo">1574</span> // next checksumOffCount read requests will use HDFS checksums.<a name="line.1574"></a> |
| <span class="sourceLineNo">1575</span> // The decrementing of this.checksumOffCount is not thread-safe,<a name="line.1575"></a> |
| <span class="sourceLineNo">1576</span> // but it is harmless because eventually checksumOffCount will be<a name="line.1576"></a> |
| <span class="sourceLineNo">1577</span> // a negative number.<a name="line.1577"></a> |
| <span class="sourceLineNo">1578</span> streamWrapper.checksumOk();<a name="line.1578"></a> |
| <span class="sourceLineNo">1579</span> return blk;<a name="line.1579"></a> |
| <span class="sourceLineNo">1580</span> }<a name="line.1580"></a> |
| <span class="sourceLineNo">1581</span> }<a name="line.1581"></a> |
| <span class="sourceLineNo">1582</span><a name="line.1582"></a> |
| <span class="sourceLineNo">1583</span> /**<a name="line.1583"></a> |
| <span class="sourceLineNo">1584</span> * Returns Check <code>onDiskSizeWithHeaderL</code> size is healthy and then return it as an int<a name="line.1584"></a> |
| <span class="sourceLineNo">1585</span> */<a name="line.1585"></a> |
| <span class="sourceLineNo">1586</span> private static int checkAndGetSizeAsInt(final long onDiskSizeWithHeaderL, final int hdrSize)<a name="line.1586"></a> |
| <span class="sourceLineNo">1587</span> throws IOException {<a name="line.1587"></a> |
| <span class="sourceLineNo">1588</span> if (<a name="line.1588"></a> |
| <span class="sourceLineNo">1589</span> (onDiskSizeWithHeaderL < hdrSize && onDiskSizeWithHeaderL != -1)<a name="line.1589"></a> |
| <span class="sourceLineNo">1590</span> || onDiskSizeWithHeaderL >= Integer.MAX_VALUE<a name="line.1590"></a> |
| <span class="sourceLineNo">1591</span> ) {<a name="line.1591"></a> |
| <span class="sourceLineNo">1592</span> throw new IOException(<a name="line.1592"></a> |
| <span class="sourceLineNo">1593</span> "Invalid onDisksize=" + onDiskSizeWithHeaderL + ": expected to be at least " + hdrSize<a name="line.1593"></a> |
| <span class="sourceLineNo">1594</span> + " and at most " + Integer.MAX_VALUE + ", or -1");<a name="line.1594"></a> |
| <span class="sourceLineNo">1595</span> }<a name="line.1595"></a> |
| <span class="sourceLineNo">1596</span> return (int) onDiskSizeWithHeaderL;<a name="line.1596"></a> |
| <span class="sourceLineNo">1597</span> }<a name="line.1597"></a> |
| <span class="sourceLineNo">1598</span><a name="line.1598"></a> |
| <span class="sourceLineNo">1599</span> /**<a name="line.1599"></a> |
| <span class="sourceLineNo">1600</span> * Verify the passed in onDiskSizeWithHeader aligns with what is in the header else something is<a name="line.1600"></a> |
| <span class="sourceLineNo">1601</span> * not right.<a name="line.1601"></a> |
| <span class="sourceLineNo">1602</span> */<a name="line.1602"></a> |
| <span class="sourceLineNo">1603</span> private void verifyOnDiskSizeMatchesHeader(final int passedIn, final ByteBuff headerBuf,<a name="line.1603"></a> |
| <span class="sourceLineNo">1604</span> final long offset, boolean verifyChecksum) throws IOException {<a name="line.1604"></a> |
| <span class="sourceLineNo">1605</span> // Assert size provided aligns with what is in the header<a name="line.1605"></a> |
| <span class="sourceLineNo">1606</span> int fromHeader = getOnDiskSizeWithHeader(headerBuf, verifyChecksum);<a name="line.1606"></a> |
| <span class="sourceLineNo">1607</span> if (passedIn != fromHeader) {<a name="line.1607"></a> |
| <span class="sourceLineNo">1608</span> throw new IOException("Passed in onDiskSizeWithHeader=" + passedIn + " != " + fromHeader<a name="line.1608"></a> |
| <span class="sourceLineNo">1609</span> + ", offset=" + offset + ", fileContext=" + this.fileContext);<a name="line.1609"></a> |
| <span class="sourceLineNo">1610</span> }<a name="line.1610"></a> |
| <span class="sourceLineNo">1611</span> }<a name="line.1611"></a> |
| <span class="sourceLineNo">1612</span><a name="line.1612"></a> |
| <span class="sourceLineNo">1613</span> /**<a name="line.1613"></a> |
| <span class="sourceLineNo">1614</span> * Check atomic reference cache for this block's header. Cache only good if next read coming<a name="line.1614"></a> |
| <span class="sourceLineNo">1615</span> * through is next in sequence in the block. We read next block's header on the tail of reading<a name="line.1615"></a> |
| <span class="sourceLineNo">1616</span> * the previous block to save a seek. Otherwise, we have to do a seek to read the header before<a name="line.1616"></a> |
| <span class="sourceLineNo">1617</span> * we can pull in the block OR we have to backup the stream because we over-read (the next<a name="line.1617"></a> |
| <span class="sourceLineNo">1618</span> * block's header).<a name="line.1618"></a> |
| <span class="sourceLineNo">1619</span> * @see PrefetchedHeader<a name="line.1619"></a> |
| <span class="sourceLineNo">1620</span> * @return The cached block header or null if not found.<a name="line.1620"></a> |
| <span class="sourceLineNo">1621</span> * @see #cacheNextBlockHeader(long, ByteBuff, int, int)<a name="line.1621"></a> |
| <span class="sourceLineNo">1622</span> */<a name="line.1622"></a> |
| <span class="sourceLineNo">1623</span> private ByteBuff getCachedHeader(final long offset) {<a name="line.1623"></a> |
| <span class="sourceLineNo">1624</span> PrefetchedHeader ph = this.prefetchedHeader.get();<a name="line.1624"></a> |
| <span class="sourceLineNo">1625</span> return ph != null && ph.offset == offset ? ph.buf : null;<a name="line.1625"></a> |
| <span class="sourceLineNo">1626</span> }<a name="line.1626"></a> |
| <span class="sourceLineNo">1627</span><a name="line.1627"></a> |
| <span class="sourceLineNo">1628</span> /**<a name="line.1628"></a> |
| <span class="sourceLineNo">1629</span> * Save away the next blocks header in atomic reference.<a name="line.1629"></a> |
| <span class="sourceLineNo">1630</span> * @see #getCachedHeader(long)<a name="line.1630"></a> |
| <span class="sourceLineNo">1631</span> * @see PrefetchedHeader<a name="line.1631"></a> |
| <span class="sourceLineNo">1632</span> */<a name="line.1632"></a> |
| <span class="sourceLineNo">1633</span> private void cacheNextBlockHeader(final long offset, ByteBuff onDiskBlock,<a name="line.1633"></a> |
| <span class="sourceLineNo">1634</span> int onDiskSizeWithHeader, int headerLength) {<a name="line.1634"></a> |
| <span class="sourceLineNo">1635</span> PrefetchedHeader ph = new PrefetchedHeader();<a name="line.1635"></a> |
| <span class="sourceLineNo">1636</span> ph.offset = offset;<a name="line.1636"></a> |
| <span class="sourceLineNo">1637</span> onDiskBlock.get(onDiskSizeWithHeader, ph.header, 0, headerLength);<a name="line.1637"></a> |
| <span class="sourceLineNo">1638</span> this.prefetchedHeader.set(ph);<a name="line.1638"></a> |
| <span class="sourceLineNo">1639</span> }<a name="line.1639"></a> |
| <span class="sourceLineNo">1640</span><a name="line.1640"></a> |
| <span class="sourceLineNo">1641</span> private int getNextBlockOnDiskSize(boolean readNextHeader, ByteBuff onDiskBlock,<a name="line.1641"></a> |
| <span class="sourceLineNo">1642</span> int onDiskSizeWithHeader) {<a name="line.1642"></a> |
| <span class="sourceLineNo">1643</span> int nextBlockOnDiskSize = -1;<a name="line.1643"></a> |
| <span class="sourceLineNo">1644</span> if (readNextHeader) {<a name="line.1644"></a> |
| <span class="sourceLineNo">1645</span> nextBlockOnDiskSize =<a name="line.1645"></a> |
| <span class="sourceLineNo">1646</span> onDiskBlock.getIntAfterPosition(onDiskSizeWithHeader + BlockType.MAGIC_LENGTH) + hdrSize;<a name="line.1646"></a> |
| <span class="sourceLineNo">1647</span> }<a name="line.1647"></a> |
| <span class="sourceLineNo">1648</span> return nextBlockOnDiskSize;<a name="line.1648"></a> |
| <span class="sourceLineNo">1649</span> }<a name="line.1649"></a> |
| <span class="sourceLineNo">1650</span><a name="line.1650"></a> |
| <span class="sourceLineNo">1651</span> private ByteBuff allocate(int size, boolean intoHeap) {<a name="line.1651"></a> |
| <span class="sourceLineNo">1652</span> return intoHeap ? HEAP.allocate(size) : allocator.allocate(size);<a name="line.1652"></a> |
| <span class="sourceLineNo">1653</span> }<a name="line.1653"></a> |
| <span class="sourceLineNo">1654</span><a name="line.1654"></a> |
| <span class="sourceLineNo">1655</span> /**<a name="line.1655"></a> |
| <span class="sourceLineNo">1656</span> * Reads a version 2 block.<a name="line.1656"></a> |
| <span class="sourceLineNo">1657</span> * @param offset the offset in the stream to read at.<a name="line.1657"></a> |
| <span class="sourceLineNo">1658</span> * @param onDiskSizeWithHeaderL the on-disk size of the block, including the header and<a name="line.1658"></a> |
| <span class="sourceLineNo">1659</span> * checksums if present or -1 if unknown (as a long). Can be -1 if<a name="line.1659"></a> |
| <span class="sourceLineNo">1660</span> * we are doing raw iteration of blocks as when loading up file<a name="line.1660"></a> |
| <span class="sourceLineNo">1661</span> * metadata; i.e. the first read of a new file. Usually non-null<a name="line.1661"></a> |
| <span class="sourceLineNo">1662</span> * gotten from the file index.<a name="line.1662"></a> |
| <span class="sourceLineNo">1663</span> * @param pread whether to use a positional read<a name="line.1663"></a> |
| <span class="sourceLineNo">1664</span> * @param verifyChecksum Whether to use HBase checksums. If HBase checksum is switched<a name="line.1664"></a> |
| <span class="sourceLineNo">1665</span> * off, then use HDFS checksum. Can also flip on/off reading same<a name="line.1665"></a> |
| <span class="sourceLineNo">1666</span> * file if we hit a troublesome patch in an hfile.<a name="line.1666"></a> |
| <span class="sourceLineNo">1667</span> * @param updateMetrics whether need to update the metrics.<a name="line.1667"></a> |
| <span class="sourceLineNo">1668</span> * @param intoHeap allocate the ByteBuff of block from heap or off-heap.<a name="line.1668"></a> |
| <span class="sourceLineNo">1669</span> * @return the HFileBlock or null if there is a HBase checksum mismatch<a name="line.1669"></a> |
| <span class="sourceLineNo">1670</span> */<a name="line.1670"></a> |
| <span class="sourceLineNo">1671</span> protected HFileBlock readBlockDataInternal(FSDataInputStream is, long offset,<a name="line.1671"></a> |
| <span class="sourceLineNo">1672</span> long onDiskSizeWithHeaderL, boolean pread, boolean verifyChecksum, boolean updateMetrics,<a name="line.1672"></a> |
| <span class="sourceLineNo">1673</span> boolean intoHeap) throws IOException {<a name="line.1673"></a> |
| <span class="sourceLineNo">1674</span> if (offset < 0) {<a name="line.1674"></a> |
| <span class="sourceLineNo">1675</span> throw new IOException("Invalid offset=" + offset + " trying to read " + "block (onDiskSize="<a name="line.1675"></a> |
| <span class="sourceLineNo">1676</span> + onDiskSizeWithHeaderL + ")");<a name="line.1676"></a> |
| <span class="sourceLineNo">1677</span> }<a name="line.1677"></a> |
| <span class="sourceLineNo">1678</span><a name="line.1678"></a> |
| <span class="sourceLineNo">1679</span> final Span span = Span.current();<a name="line.1679"></a> |
| <span class="sourceLineNo">1680</span> final AttributesBuilder attributesBuilder = Attributes.builder();<a name="line.1680"></a> |
| <span class="sourceLineNo">1681</span> Optional.of(Context.current()).map(val -> val.get(CONTEXT_KEY))<a name="line.1681"></a> |
| <span class="sourceLineNo">1682</span> .ifPresent(c -> c.accept(attributesBuilder));<a name="line.1682"></a> |
| <span class="sourceLineNo">1683</span> int onDiskSizeWithHeader = checkAndGetSizeAsInt(onDiskSizeWithHeaderL, hdrSize);<a name="line.1683"></a> |
| <span class="sourceLineNo">1684</span> // Try and get cached header. Will serve us in rare case where onDiskSizeWithHeaderL is -1<a name="line.1684"></a> |
| <span class="sourceLineNo">1685</span> // and will save us having to seek the stream backwards to reread the header we<a name="line.1685"></a> |
| <span class="sourceLineNo">1686</span> // read the last time through here.<a name="line.1686"></a> |
| <span class="sourceLineNo">1687</span> ByteBuff headerBuf = getCachedHeader(offset);<a name="line.1687"></a> |
| <span class="sourceLineNo">1688</span> LOG.trace(<a name="line.1688"></a> |
| <span class="sourceLineNo">1689</span> "Reading {} at offset={}, pread={}, verifyChecksum={}, cachedHeader={}, "<a name="line.1689"></a> |
| <span class="sourceLineNo">1690</span> + "onDiskSizeWithHeader={}",<a name="line.1690"></a> |
| <span class="sourceLineNo">1691</span> this.fileContext.getHFileName(), offset, pread, verifyChecksum, headerBuf,<a name="line.1691"></a> |
| <span class="sourceLineNo">1692</span> onDiskSizeWithHeader);<a name="line.1692"></a> |
| <span class="sourceLineNo">1693</span> // This is NOT same as verifyChecksum. This latter is whether to do hbase<a name="line.1693"></a> |
| <span class="sourceLineNo">1694</span> // checksums. Can change with circumstances. The below flag is whether the<a name="line.1694"></a> |
| <span class="sourceLineNo">1695</span> // file has support for checksums (version 2+).<a name="line.1695"></a> |
| <span class="sourceLineNo">1696</span> boolean checksumSupport = this.fileContext.isUseHBaseChecksum();<a name="line.1696"></a> |
| <span class="sourceLineNo">1697</span> long startTime = EnvironmentEdgeManager.currentTime();<a name="line.1697"></a> |
| <span class="sourceLineNo">1698</span> if (onDiskSizeWithHeader <= 0) {<a name="line.1698"></a> |
| <span class="sourceLineNo">1699</span> // We were not passed the block size. Need to get it from the header. If header was<a name="line.1699"></a> |
| <span class="sourceLineNo">1700</span> // not cached (see getCachedHeader above), need to seek to pull it in. This is costly<a name="line.1700"></a> |
| <span class="sourceLineNo">1701</span> // and should happen very rarely. Currently happens on open of a hfile reader where we<a name="line.1701"></a> |
| <span class="sourceLineNo">1702</span> // read the trailer blocks to pull in the indices. Otherwise, we are reading block sizes<a name="line.1702"></a> |
| <span class="sourceLineNo">1703</span> // out of the hfile index. To check, enable TRACE in this file and you'll get an exception<a name="line.1703"></a> |
| <span class="sourceLineNo">1704</span> // in a LOG every time we seek. See HBASE-17072 for more detail.<a name="line.1704"></a> |
| <span class="sourceLineNo">1705</span> if (headerBuf == null) {<a name="line.1705"></a> |
| <span class="sourceLineNo">1706</span> if (LOG.isTraceEnabled()) {<a name="line.1706"></a> |
| <span class="sourceLineNo">1707</span> LOG.trace("Extra seek to get block size!", new RuntimeException());<a name="line.1707"></a> |
| <span class="sourceLineNo">1708</span> }<a name="line.1708"></a> |
| <span class="sourceLineNo">1709</span> span.addEvent("Extra seek to get block size!", attributesBuilder.build());<a name="line.1709"></a> |
| <span class="sourceLineNo">1710</span> headerBuf = HEAP.allocate(hdrSize);<a name="line.1710"></a> |
| <span class="sourceLineNo">1711</span> readAtOffset(is, headerBuf, hdrSize, false, offset, pread);<a name="line.1711"></a> |
| <span class="sourceLineNo">1712</span> headerBuf.rewind();<a name="line.1712"></a> |
| <span class="sourceLineNo">1713</span> }<a name="line.1713"></a> |
| <span class="sourceLineNo">1714</span> onDiskSizeWithHeader = getOnDiskSizeWithHeader(headerBuf, checksumSupport);<a name="line.1714"></a> |
| <span class="sourceLineNo">1715</span> }<a name="line.1715"></a> |
| <span class="sourceLineNo">1716</span> int preReadHeaderSize = headerBuf == null ? 0 : hdrSize;<a name="line.1716"></a> |
| <span class="sourceLineNo">1717</span> // Allocate enough space to fit the next block's header too; saves a seek next time through.<a name="line.1717"></a> |
| <span class="sourceLineNo">1718</span> // onDiskBlock is whole block + header + checksums then extra hdrSize to read next header;<a name="line.1718"></a> |
| <span class="sourceLineNo">1719</span> // onDiskSizeWithHeader is header, body, and any checksums if present. preReadHeaderSize<a name="line.1719"></a> |
| <span class="sourceLineNo">1720</span> // says where to start reading. If we have the header cached, then we don't need to read<a name="line.1720"></a> |
| <span class="sourceLineNo">1721</span> // it again and we can likely read from last place we left off w/o need to backup and reread<a name="line.1721"></a> |
| <span class="sourceLineNo">1722</span> // the header we read last time through here.<a name="line.1722"></a> |
| <span class="sourceLineNo">1723</span> ByteBuff onDiskBlock = this.allocate(onDiskSizeWithHeader + hdrSize, intoHeap);<a name="line.1723"></a> |
| <span class="sourceLineNo">1724</span> boolean initHFileBlockSuccess = false;<a name="line.1724"></a> |
| <span class="sourceLineNo">1725</span> try {<a name="line.1725"></a> |
| <span class="sourceLineNo">1726</span> if (headerBuf != null) {<a name="line.1726"></a> |
| <span class="sourceLineNo">1727</span> onDiskBlock.put(0, headerBuf, 0, hdrSize).position(hdrSize);<a name="line.1727"></a> |
| <span class="sourceLineNo">1728</span> }<a name="line.1728"></a> |
| <span class="sourceLineNo">1729</span> boolean readNextHeader = readAtOffset(is, onDiskBlock,<a name="line.1729"></a> |
| <span class="sourceLineNo">1730</span> onDiskSizeWithHeader - preReadHeaderSize, true, offset + preReadHeaderSize, pread);<a name="line.1730"></a> |
| <span class="sourceLineNo">1731</span> onDiskBlock.rewind(); // in case of moving position when copying a cached header<a name="line.1731"></a> |
| <span class="sourceLineNo">1732</span> int nextBlockOnDiskSize =<a name="line.1732"></a> |
| <span class="sourceLineNo">1733</span> getNextBlockOnDiskSize(readNextHeader, onDiskBlock, onDiskSizeWithHeader);<a name="line.1733"></a> |
| <span class="sourceLineNo">1734</span> if (headerBuf == null) {<a name="line.1734"></a> |
| <span class="sourceLineNo">1735</span> headerBuf = onDiskBlock.duplicate().position(0).limit(hdrSize);<a name="line.1735"></a> |
| <span class="sourceLineNo">1736</span> }<a name="line.1736"></a> |
| <span class="sourceLineNo">1737</span> // Do a few checks before we go instantiate HFileBlock.<a name="line.1737"></a> |
| <span class="sourceLineNo">1738</span> assert onDiskSizeWithHeader > this.hdrSize;<a name="line.1738"></a> |
| <span class="sourceLineNo">1739</span> verifyOnDiskSizeMatchesHeader(onDiskSizeWithHeader, headerBuf, offset, checksumSupport);<a name="line.1739"></a> |
| <span class="sourceLineNo">1740</span> ByteBuff curBlock = onDiskBlock.duplicate().position(0).limit(onDiskSizeWithHeader);<a name="line.1740"></a> |
| <span class="sourceLineNo">1741</span> // Verify checksum of the data before using it for building HFileBlock.<a name="line.1741"></a> |
| <span class="sourceLineNo">1742</span> if (verifyChecksum && !validateChecksum(offset, curBlock, hdrSize)) {<a name="line.1742"></a> |
| <span class="sourceLineNo">1743</span> return null;<a name="line.1743"></a> |
| <span class="sourceLineNo">1744</span> }<a name="line.1744"></a> |
| <span class="sourceLineNo">1745</span> // remove checksum from buffer now that it's verified<a name="line.1745"></a> |
| <span class="sourceLineNo">1746</span> int sizeWithoutChecksum = curBlock.getInt(Header.ON_DISK_DATA_SIZE_WITH_HEADER_INDEX);<a name="line.1746"></a> |
| <span class="sourceLineNo">1747</span> curBlock.limit(sizeWithoutChecksum);<a name="line.1747"></a> |
| <span class="sourceLineNo">1748</span> long duration = EnvironmentEdgeManager.currentTime() - startTime;<a name="line.1748"></a> |
| <span class="sourceLineNo">1749</span> if (updateMetrics) {<a name="line.1749"></a> |
| <span class="sourceLineNo">1750</span> HFile.updateReadLatency(duration, pread);<a name="line.1750"></a> |
| <span class="sourceLineNo">1751</span> }<a name="line.1751"></a> |
| <span class="sourceLineNo">1752</span> // The onDiskBlock will become the headerAndDataBuffer for this block.<a name="line.1752"></a> |
| <span class="sourceLineNo">1753</span> // If nextBlockOnDiskSizeWithHeader is not zero, the onDiskBlock already<a name="line.1753"></a> |
| <span class="sourceLineNo">1754</span> // contains the header of next block, so no need to set next block's header in it.<a name="line.1754"></a> |
| <span class="sourceLineNo">1755</span> HFileBlock hFileBlock = createFromBuff(curBlock, checksumSupport, offset,<a name="line.1755"></a> |
| <span class="sourceLineNo">1756</span> nextBlockOnDiskSize, fileContext, intoHeap ? HEAP : allocator);<a name="line.1756"></a> |
| <span class="sourceLineNo">1757</span> // Run check on uncompressed sizings.<a name="line.1757"></a> |
| <span class="sourceLineNo">1758</span> if (!fileContext.isCompressedOrEncrypted()) {<a name="line.1758"></a> |
| <span class="sourceLineNo">1759</span> hFileBlock.sanityCheckUncompressed();<a name="line.1759"></a> |
| <span class="sourceLineNo">1760</span> }<a name="line.1760"></a> |
| <span class="sourceLineNo">1761</span> LOG.trace("Read {} in {} ms", hFileBlock, duration);<a name="line.1761"></a> |
| <span class="sourceLineNo">1762</span> span.addEvent("Read block", attributesBuilder.build());<a name="line.1762"></a> |
| <span class="sourceLineNo">1763</span> // Cache next block header if we read it for the next time through here.<a name="line.1763"></a> |
| <span class="sourceLineNo">1764</span> if (nextBlockOnDiskSize != -1) {<a name="line.1764"></a> |
| <span class="sourceLineNo">1765</span> cacheNextBlockHeader(offset + hFileBlock.getOnDiskSizeWithHeader(), onDiskBlock,<a name="line.1765"></a> |
| <span class="sourceLineNo">1766</span> onDiskSizeWithHeader, hdrSize);<a name="line.1766"></a> |
| <span class="sourceLineNo">1767</span> }<a name="line.1767"></a> |
| <span class="sourceLineNo">1768</span> initHFileBlockSuccess = true;<a name="line.1768"></a> |
| <span class="sourceLineNo">1769</span> return hFileBlock;<a name="line.1769"></a> |
| <span class="sourceLineNo">1770</span> } finally {<a name="line.1770"></a> |
| <span class="sourceLineNo">1771</span> if (!initHFileBlockSuccess) {<a name="line.1771"></a> |
| <span class="sourceLineNo">1772</span> onDiskBlock.release();<a name="line.1772"></a> |
| <span class="sourceLineNo">1773</span> }<a name="line.1773"></a> |
| <span class="sourceLineNo">1774</span> }<a name="line.1774"></a> |
| <span class="sourceLineNo">1775</span> }<a name="line.1775"></a> |
| <span class="sourceLineNo">1776</span><a name="line.1776"></a> |
| <span class="sourceLineNo">1777</span> @Override<a name="line.1777"></a> |
| <span class="sourceLineNo">1778</span> public void setIncludesMemStoreTS(boolean includesMemstoreTS) {<a name="line.1778"></a> |
| <span class="sourceLineNo">1779</span> this.fileContext =<a name="line.1779"></a> |
| <span class="sourceLineNo">1780</span> new HFileContextBuilder(this.fileContext).withIncludesMvcc(includesMemstoreTS).build();<a name="line.1780"></a> |
| <span class="sourceLineNo">1781</span> }<a name="line.1781"></a> |
| <span class="sourceLineNo">1782</span><a name="line.1782"></a> |
| <span class="sourceLineNo">1783</span> @Override<a name="line.1783"></a> |
| <span class="sourceLineNo">1784</span> public void setDataBlockEncoder(HFileDataBlockEncoder encoder, Configuration conf) {<a name="line.1784"></a> |
| <span class="sourceLineNo">1785</span> encodedBlockDecodingCtx = encoder.newDataBlockDecodingContext(conf, fileContext);<a name="line.1785"></a> |
| <span class="sourceLineNo">1786</span> }<a name="line.1786"></a> |
| <span class="sourceLineNo">1787</span><a name="line.1787"></a> |
| <span class="sourceLineNo">1788</span> @Override<a name="line.1788"></a> |
| <span class="sourceLineNo">1789</span> public HFileBlockDecodingContext getBlockDecodingContext() {<a name="line.1789"></a> |
| <span class="sourceLineNo">1790</span> return this.encodedBlockDecodingCtx;<a name="line.1790"></a> |
| <span class="sourceLineNo">1791</span> }<a name="line.1791"></a> |
| <span class="sourceLineNo">1792</span><a name="line.1792"></a> |
| <span class="sourceLineNo">1793</span> @Override<a name="line.1793"></a> |
| <span class="sourceLineNo">1794</span> public HFileBlockDecodingContext getDefaultBlockDecodingContext() {<a name="line.1794"></a> |
| <span class="sourceLineNo">1795</span> return this.defaultDecodingCtx;<a name="line.1795"></a> |
| <span class="sourceLineNo">1796</span> }<a name="line.1796"></a> |
| <span class="sourceLineNo">1797</span><a name="line.1797"></a> |
| <span class="sourceLineNo">1798</span> /**<a name="line.1798"></a> |
| <span class="sourceLineNo">1799</span> * Generates the checksum for the header as well as the data and then validates it. If the block<a name="line.1799"></a> |
| <span class="sourceLineNo">1800</span> * doesn't uses checksum, returns false.<a name="line.1800"></a> |
| <span class="sourceLineNo">1801</span> * @return True if checksum matches, else false.<a name="line.1801"></a> |
| <span class="sourceLineNo">1802</span> */<a name="line.1802"></a> |
| <span class="sourceLineNo">1803</span> private boolean validateChecksum(long offset, ByteBuff data, int hdrSize) {<a name="line.1803"></a> |
| <span class="sourceLineNo">1804</span> // If this is an older version of the block that does not have checksums, then return false<a name="line.1804"></a> |
| <span class="sourceLineNo">1805</span> // indicating that checksum verification did not succeed. Actually, this method should never<a name="line.1805"></a> |
| <span class="sourceLineNo">1806</span> // be called when the minorVersion is 0, thus this is a defensive check for a cannot-happen<a name="line.1806"></a> |
| <span class="sourceLineNo">1807</span> // case. Since this is a cannot-happen case, it is better to return false to indicate a<a name="line.1807"></a> |
| <span class="sourceLineNo">1808</span> // checksum validation failure.<a name="line.1808"></a> |
| <span class="sourceLineNo">1809</span> if (!fileContext.isUseHBaseChecksum()) {<a name="line.1809"></a> |
| <span class="sourceLineNo">1810</span> return false;<a name="line.1810"></a> |
| <span class="sourceLineNo">1811</span> }<a name="line.1811"></a> |
| <span class="sourceLineNo">1812</span> return ChecksumUtil.validateChecksum(data, pathName, offset, hdrSize);<a name="line.1812"></a> |
| <span class="sourceLineNo">1813</span> }<a name="line.1813"></a> |
| <span class="sourceLineNo">1814</span><a name="line.1814"></a> |
| <span class="sourceLineNo">1815</span> @Override<a name="line.1815"></a> |
| <span class="sourceLineNo">1816</span> public void closeStreams() throws IOException {<a name="line.1816"></a> |
| <span class="sourceLineNo">1817</span> streamWrapper.close();<a name="line.1817"></a> |
| <span class="sourceLineNo">1818</span> }<a name="line.1818"></a> |
| <span class="sourceLineNo">1819</span><a name="line.1819"></a> |
| <span class="sourceLineNo">1820</span> @Override<a name="line.1820"></a> |
| <span class="sourceLineNo">1821</span> public void unbufferStream() {<a name="line.1821"></a> |
| <span class="sourceLineNo">1822</span> // To handle concurrent reads, ensure that no other client is accessing the streams while we<a name="line.1822"></a> |
| <span class="sourceLineNo">1823</span> // unbuffer it.<a name="line.1823"></a> |
| <span class="sourceLineNo">1824</span> if (streamLock.tryLock()) {<a name="line.1824"></a> |
| <span class="sourceLineNo">1825</span> try {<a name="line.1825"></a> |
| <span class="sourceLineNo">1826</span> this.streamWrapper.unbuffer();<a name="line.1826"></a> |
| <span class="sourceLineNo">1827</span> } finally {<a name="line.1827"></a> |
| <span class="sourceLineNo">1828</span> streamLock.unlock();<a name="line.1828"></a> |
| <span class="sourceLineNo">1829</span> }<a name="line.1829"></a> |
| <span class="sourceLineNo">1830</span> }<a name="line.1830"></a> |
| <span class="sourceLineNo">1831</span> }<a name="line.1831"></a> |
| <span class="sourceLineNo">1832</span><a name="line.1832"></a> |
| <span class="sourceLineNo">1833</span> @Override<a name="line.1833"></a> |
| <span class="sourceLineNo">1834</span> public String toString() {<a name="line.1834"></a> |
| <span class="sourceLineNo">1835</span> return "hfs=" + hfs + ", path=" + pathName + ", fileContext=" + fileContext;<a name="line.1835"></a> |
| <span class="sourceLineNo">1836</span> }<a name="line.1836"></a> |
| <span class="sourceLineNo">1837</span> }<a name="line.1837"></a> |
| <span class="sourceLineNo">1838</span><a name="line.1838"></a> |
| <span class="sourceLineNo">1839</span> /** An additional sanity-check in case no compression or encryption is being used. */<a name="line.1839"></a> |
| <span class="sourceLineNo">1840</span> void sanityCheckUncompressed() throws IOException {<a name="line.1840"></a> |
| <span class="sourceLineNo">1841</span> if (onDiskSizeWithoutHeader != uncompressedSizeWithoutHeader + totalChecksumBytes()) {<a name="line.1841"></a> |
| <span class="sourceLineNo">1842</span> throw new IOException("Using no compression but " + "onDiskSizeWithoutHeader="<a name="line.1842"></a> |
| <span class="sourceLineNo">1843</span> + onDiskSizeWithoutHeader + ", " + "uncompressedSizeWithoutHeader="<a name="line.1843"></a> |
| <span class="sourceLineNo">1844</span> + uncompressedSizeWithoutHeader + ", numChecksumbytes=" + totalChecksumBytes());<a name="line.1844"></a> |
| <span class="sourceLineNo">1845</span> }<a name="line.1845"></a> |
| <span class="sourceLineNo">1846</span> }<a name="line.1846"></a> |
| <span class="sourceLineNo">1847</span><a name="line.1847"></a> |
| <span class="sourceLineNo">1848</span> // Cacheable implementation<a name="line.1848"></a> |
| <span class="sourceLineNo">1849</span> @Override<a name="line.1849"></a> |
| <span class="sourceLineNo">1850</span> public int getSerializedLength() {<a name="line.1850"></a> |
| <span class="sourceLineNo">1851</span> if (buf != null) {<a name="line.1851"></a> |
| <span class="sourceLineNo">1852</span> // Include extra bytes for block metadata.<a name="line.1852"></a> |
| <span class="sourceLineNo">1853</span> return this.buf.limit() + BLOCK_METADATA_SPACE;<a name="line.1853"></a> |
| <span class="sourceLineNo">1854</span> }<a name="line.1854"></a> |
| <span class="sourceLineNo">1855</span> return 0;<a name="line.1855"></a> |
| <span class="sourceLineNo">1856</span> }<a name="line.1856"></a> |
| <span class="sourceLineNo">1857</span><a name="line.1857"></a> |
| <span class="sourceLineNo">1858</span> // Cacheable implementation<a name="line.1858"></a> |
| <span class="sourceLineNo">1859</span> @Override<a name="line.1859"></a> |
| <span class="sourceLineNo">1860</span> public void serialize(ByteBuffer destination, boolean includeNextBlockMetadata) {<a name="line.1860"></a> |
| <span class="sourceLineNo">1861</span> this.buf.get(destination, 0, getSerializedLength() - BLOCK_METADATA_SPACE);<a name="line.1861"></a> |
| <span class="sourceLineNo">1862</span> destination = addMetaData(destination, includeNextBlockMetadata);<a name="line.1862"></a> |
| <span class="sourceLineNo">1863</span><a name="line.1863"></a> |
| <span class="sourceLineNo">1864</span> // Make it ready for reading. flip sets position to zero and limit to current position which<a name="line.1864"></a> |
| <span class="sourceLineNo">1865</span> // is what we want if we do not want to serialize the block plus checksums if present plus<a name="line.1865"></a> |
| <span class="sourceLineNo">1866</span> // metadata.<a name="line.1866"></a> |
| <span class="sourceLineNo">1867</span> destination.flip();<a name="line.1867"></a> |
| <span class="sourceLineNo">1868</span> }<a name="line.1868"></a> |
| <span class="sourceLineNo">1869</span><a name="line.1869"></a> |
| <span class="sourceLineNo">1870</span> /**<a name="line.1870"></a> |
| <span class="sourceLineNo">1871</span> * For use by bucketcache. This exposes internals.<a name="line.1871"></a> |
| <span class="sourceLineNo">1872</span> */<a name="line.1872"></a> |
| <span class="sourceLineNo">1873</span> public ByteBuffer getMetaData(ByteBuffer bb) {<a name="line.1873"></a> |
| <span class="sourceLineNo">1874</span> bb = addMetaData(bb, true);<a name="line.1874"></a> |
| <span class="sourceLineNo">1875</span> bb.flip();<a name="line.1875"></a> |
| <span class="sourceLineNo">1876</span> return bb;<a name="line.1876"></a> |
| <span class="sourceLineNo">1877</span> }<a name="line.1877"></a> |
| <span class="sourceLineNo">1878</span><a name="line.1878"></a> |
| <span class="sourceLineNo">1879</span> /**<a name="line.1879"></a> |
| <span class="sourceLineNo">1880</span> * Adds metadata at current position (position is moved forward). Does not flip or reset.<a name="line.1880"></a> |
| <span class="sourceLineNo">1881</span> * @return The passed <code>destination</code> with metadata added.<a name="line.1881"></a> |
| <span class="sourceLineNo">1882</span> */<a name="line.1882"></a> |
| <span class="sourceLineNo">1883</span> private ByteBuffer addMetaData(final ByteBuffer destination, boolean includeNextBlockMetadata) {<a name="line.1883"></a> |
| <span class="sourceLineNo">1884</span> destination.put(this.fileContext.isUseHBaseChecksum() ? (byte) 1 : (byte) 0);<a name="line.1884"></a> |
| <span class="sourceLineNo">1885</span> destination.putLong(this.offset);<a name="line.1885"></a> |
| <span class="sourceLineNo">1886</span> if (includeNextBlockMetadata) {<a name="line.1886"></a> |
| <span class="sourceLineNo">1887</span> destination.putInt(this.nextBlockOnDiskSize);<a name="line.1887"></a> |
| <span class="sourceLineNo">1888</span> }<a name="line.1888"></a> |
| <span class="sourceLineNo">1889</span> return destination;<a name="line.1889"></a> |
| <span class="sourceLineNo">1890</span> }<a name="line.1890"></a> |
| <span class="sourceLineNo">1891</span><a name="line.1891"></a> |
| <span class="sourceLineNo">1892</span> // Cacheable implementation<a name="line.1892"></a> |
| <span class="sourceLineNo">1893</span> @Override<a name="line.1893"></a> |
| <span class="sourceLineNo">1894</span> public CacheableDeserializer<Cacheable> getDeserializer() {<a name="line.1894"></a> |
| <span class="sourceLineNo">1895</span> return HFileBlock.BLOCK_DESERIALIZER;<a name="line.1895"></a> |
| <span class="sourceLineNo">1896</span> }<a name="line.1896"></a> |
| <span class="sourceLineNo">1897</span><a name="line.1897"></a> |
| <span class="sourceLineNo">1898</span> @Override<a name="line.1898"></a> |
| <span class="sourceLineNo">1899</span> public int hashCode() {<a name="line.1899"></a> |
| <span class="sourceLineNo">1900</span> int result = 1;<a name="line.1900"></a> |
| <span class="sourceLineNo">1901</span> result = result * 31 + blockType.hashCode();<a name="line.1901"></a> |
| <span class="sourceLineNo">1902</span> result = result * 31 + nextBlockOnDiskSize;<a name="line.1902"></a> |
| <span class="sourceLineNo">1903</span> result = result * 31 + (int) (offset ^ (offset >>> 32));<a name="line.1903"></a> |
| <span class="sourceLineNo">1904</span> result = result * 31 + onDiskSizeWithoutHeader;<a name="line.1904"></a> |
| <span class="sourceLineNo">1905</span> result = result * 31 + (int) (prevBlockOffset ^ (prevBlockOffset >>> 32));<a name="line.1905"></a> |
| <span class="sourceLineNo">1906</span> result = result * 31 + uncompressedSizeWithoutHeader;<a name="line.1906"></a> |
| <span class="sourceLineNo">1907</span> result = result * 31 + buf.hashCode();<a name="line.1907"></a> |
| <span class="sourceLineNo">1908</span> return result;<a name="line.1908"></a> |
| <span class="sourceLineNo">1909</span> }<a name="line.1909"></a> |
| <span class="sourceLineNo">1910</span><a name="line.1910"></a> |
| <span class="sourceLineNo">1911</span> @Override<a name="line.1911"></a> |
| <span class="sourceLineNo">1912</span> public boolean equals(Object comparison) {<a name="line.1912"></a> |
| <span class="sourceLineNo">1913</span> if (this == comparison) {<a name="line.1913"></a> |
| <span class="sourceLineNo">1914</span> return true;<a name="line.1914"></a> |
| <span class="sourceLineNo">1915</span> }<a name="line.1915"></a> |
| <span class="sourceLineNo">1916</span> if (comparison == null) {<a name="line.1916"></a> |
| <span class="sourceLineNo">1917</span> return false;<a name="line.1917"></a> |
| <span class="sourceLineNo">1918</span> }<a name="line.1918"></a> |
| <span class="sourceLineNo">1919</span> if (!(comparison instanceof HFileBlock)) {<a name="line.1919"></a> |
| <span class="sourceLineNo">1920</span> return false;<a name="line.1920"></a> |
| <span class="sourceLineNo">1921</span> }<a name="line.1921"></a> |
| <span class="sourceLineNo">1922</span><a name="line.1922"></a> |
| <span class="sourceLineNo">1923</span> HFileBlock castedComparison = (HFileBlock) comparison;<a name="line.1923"></a> |
| <span class="sourceLineNo">1924</span><a name="line.1924"></a> |
| <span class="sourceLineNo">1925</span> if (castedComparison.blockType != this.blockType) {<a name="line.1925"></a> |
| <span class="sourceLineNo">1926</span> return false;<a name="line.1926"></a> |
| <span class="sourceLineNo">1927</span> }<a name="line.1927"></a> |
| <span class="sourceLineNo">1928</span> if (castedComparison.nextBlockOnDiskSize != this.nextBlockOnDiskSize) {<a name="line.1928"></a> |
| <span class="sourceLineNo">1929</span> return false;<a name="line.1929"></a> |
| <span class="sourceLineNo">1930</span> }<a name="line.1930"></a> |
| <span class="sourceLineNo">1931</span> // Offset is important. Needed when we have to remake cachekey when block is returned to cache.<a name="line.1931"></a> |
| <span class="sourceLineNo">1932</span> if (castedComparison.offset != this.offset) {<a name="line.1932"></a> |
| <span class="sourceLineNo">1933</span> return false;<a name="line.1933"></a> |
| <span class="sourceLineNo">1934</span> }<a name="line.1934"></a> |
| <span class="sourceLineNo">1935</span> if (castedComparison.onDiskSizeWithoutHeader != this.onDiskSizeWithoutHeader) {<a name="line.1935"></a> |
| <span class="sourceLineNo">1936</span> return false;<a name="line.1936"></a> |
| <span class="sourceLineNo">1937</span> }<a name="line.1937"></a> |
| <span class="sourceLineNo">1938</span> if (castedComparison.prevBlockOffset != this.prevBlockOffset) {<a name="line.1938"></a> |
| <span class="sourceLineNo">1939</span> return false;<a name="line.1939"></a> |
| <span class="sourceLineNo">1940</span> }<a name="line.1940"></a> |
| <span class="sourceLineNo">1941</span> if (castedComparison.uncompressedSizeWithoutHeader != this.uncompressedSizeWithoutHeader) {<a name="line.1941"></a> |
| <span class="sourceLineNo">1942</span> return false;<a name="line.1942"></a> |
| <span class="sourceLineNo">1943</span> }<a name="line.1943"></a> |
| <span class="sourceLineNo">1944</span> if (<a name="line.1944"></a> |
| <span class="sourceLineNo">1945</span> ByteBuff.compareTo(this.buf, 0, this.buf.limit(), castedComparison.buf, 0,<a name="line.1945"></a> |
| <span class="sourceLineNo">1946</span> castedComparison.buf.limit()) != 0<a name="line.1946"></a> |
| <span class="sourceLineNo">1947</span> ) {<a name="line.1947"></a> |
| <span class="sourceLineNo">1948</span> return false;<a name="line.1948"></a> |
| <span class="sourceLineNo">1949</span> }<a name="line.1949"></a> |
| <span class="sourceLineNo">1950</span> return true;<a name="line.1950"></a> |
| <span class="sourceLineNo">1951</span> }<a name="line.1951"></a> |
| <span class="sourceLineNo">1952</span><a name="line.1952"></a> |
| <span class="sourceLineNo">1953</span> DataBlockEncoding getDataBlockEncoding() {<a name="line.1953"></a> |
| <span class="sourceLineNo">1954</span> if (blockType == BlockType.ENCODED_DATA) {<a name="line.1954"></a> |
| <span class="sourceLineNo">1955</span> return DataBlockEncoding.getEncodingById(getDataBlockEncodingId());<a name="line.1955"></a> |
| <span class="sourceLineNo">1956</span> }<a name="line.1956"></a> |
| <span class="sourceLineNo">1957</span> return DataBlockEncoding.NONE;<a name="line.1957"></a> |
| <span class="sourceLineNo">1958</span> }<a name="line.1958"></a> |
| <span class="sourceLineNo">1959</span><a name="line.1959"></a> |
| <span class="sourceLineNo">1960</span> byte getChecksumType() {<a name="line.1960"></a> |
| <span class="sourceLineNo">1961</span> return this.fileContext.getChecksumType().getCode();<a name="line.1961"></a> |
| <span class="sourceLineNo">1962</span> }<a name="line.1962"></a> |
| <span class="sourceLineNo">1963</span><a name="line.1963"></a> |
| <span class="sourceLineNo">1964</span> int getBytesPerChecksum() {<a name="line.1964"></a> |
| <span class="sourceLineNo">1965</span> return this.fileContext.getBytesPerChecksum();<a name="line.1965"></a> |
| <span class="sourceLineNo">1966</span> }<a name="line.1966"></a> |
| <span class="sourceLineNo">1967</span><a name="line.1967"></a> |
| <span class="sourceLineNo">1968</span> /** Returns the size of data on disk + header. Excludes checksum. */<a name="line.1968"></a> |
| <span class="sourceLineNo">1969</span> int getOnDiskDataSizeWithHeader() {<a name="line.1969"></a> |
| <span class="sourceLineNo">1970</span> return this.onDiskDataSizeWithHeader;<a name="line.1970"></a> |
| <span class="sourceLineNo">1971</span> }<a name="line.1971"></a> |
| <span class="sourceLineNo">1972</span><a name="line.1972"></a> |
| <span class="sourceLineNo">1973</span> /**<a name="line.1973"></a> |
| <span class="sourceLineNo">1974</span> * Calculate the number of bytes required to store all the checksums for this block. Each checksum<a name="line.1974"></a> |
| <span class="sourceLineNo">1975</span> * value is a 4 byte integer.<a name="line.1975"></a> |
| <span class="sourceLineNo">1976</span> */<a name="line.1976"></a> |
| <span class="sourceLineNo">1977</span> int totalChecksumBytes() {<a name="line.1977"></a> |
| <span class="sourceLineNo">1978</span> // If the hfile block has minorVersion 0, then there are no checksum<a name="line.1978"></a> |
| <span class="sourceLineNo">1979</span> // data to validate. Similarly, a zero value in this.bytesPerChecksum<a name="line.1979"></a> |
| <span class="sourceLineNo">1980</span> // indicates that cached blocks do not have checksum data because<a name="line.1980"></a> |
| <span class="sourceLineNo">1981</span> // checksums were already validated when the block was read from disk.<a name="line.1981"></a> |
| <span class="sourceLineNo">1982</span> if (!fileContext.isUseHBaseChecksum() || this.fileContext.getBytesPerChecksum() == 0) {<a name="line.1982"></a> |
| <span class="sourceLineNo">1983</span> return 0;<a name="line.1983"></a> |
| <span class="sourceLineNo">1984</span> }<a name="line.1984"></a> |
| <span class="sourceLineNo">1985</span> return (int) ChecksumUtil.numBytes(onDiskDataSizeWithHeader,<a name="line.1985"></a> |
| <span class="sourceLineNo">1986</span> this.fileContext.getBytesPerChecksum());<a name="line.1986"></a> |
| <span class="sourceLineNo">1987</span> }<a name="line.1987"></a> |
| <span class="sourceLineNo">1988</span><a name="line.1988"></a> |
| <span class="sourceLineNo">1989</span> /**<a name="line.1989"></a> |
| <span class="sourceLineNo">1990</span> * Returns the size of this block header.<a name="line.1990"></a> |
| <span class="sourceLineNo">1991</span> */<a name="line.1991"></a> |
| <span class="sourceLineNo">1992</span> public int headerSize() {<a name="line.1992"></a> |
| <span class="sourceLineNo">1993</span> return headerSize(this.fileContext.isUseHBaseChecksum());<a name="line.1993"></a> |
| <span class="sourceLineNo">1994</span> }<a name="line.1994"></a> |
| <span class="sourceLineNo">1995</span><a name="line.1995"></a> |
| <span class="sourceLineNo">1996</span> /**<a name="line.1996"></a> |
| <span class="sourceLineNo">1997</span> * Maps a minor version to the size of the header.<a name="line.1997"></a> |
| <span class="sourceLineNo">1998</span> */<a name="line.1998"></a> |
| <span class="sourceLineNo">1999</span> public static int headerSize(boolean usesHBaseChecksum) {<a name="line.1999"></a> |
| <span class="sourceLineNo">2000</span> return usesHBaseChecksum<a name="line.2000"></a> |
| <span class="sourceLineNo">2001</span> ? HConstants.HFILEBLOCK_HEADER_SIZE<a name="line.2001"></a> |
| <span class="sourceLineNo">2002</span> : HConstants.HFILEBLOCK_HEADER_SIZE_NO_CHECKSUM;<a name="line.2002"></a> |
| <span class="sourceLineNo">2003</span> }<a name="line.2003"></a> |
| <span class="sourceLineNo">2004</span><a name="line.2004"></a> |
| <span class="sourceLineNo">2005</span> /**<a name="line.2005"></a> |
| <span class="sourceLineNo">2006</span> * Return the appropriate DUMMY_HEADER for the minor version<a name="line.2006"></a> |
| <span class="sourceLineNo">2007</span> */<a name="line.2007"></a> |
| <span class="sourceLineNo">2008</span> // TODO: Why is this in here?<a name="line.2008"></a> |
| <span class="sourceLineNo">2009</span> byte[] getDummyHeaderForVersion() {<a name="line.2009"></a> |
| <span class="sourceLineNo">2010</span> return getDummyHeaderForVersion(this.fileContext.isUseHBaseChecksum());<a name="line.2010"></a> |
| <span class="sourceLineNo">2011</span> }<a name="line.2011"></a> |
| <span class="sourceLineNo">2012</span><a name="line.2012"></a> |
| <span class="sourceLineNo">2013</span> /**<a name="line.2013"></a> |
| <span class="sourceLineNo">2014</span> * Return the appropriate DUMMY_HEADER for the minor version<a name="line.2014"></a> |
| <span class="sourceLineNo">2015</span> */<a name="line.2015"></a> |
| <span class="sourceLineNo">2016</span> static private byte[] getDummyHeaderForVersion(boolean usesHBaseChecksum) {<a name="line.2016"></a> |
| <span class="sourceLineNo">2017</span> return usesHBaseChecksum ? HConstants.HFILEBLOCK_DUMMY_HEADER : DUMMY_HEADER_NO_CHECKSUM;<a name="line.2017"></a> |
| <span class="sourceLineNo">2018</span> }<a name="line.2018"></a> |
| <span class="sourceLineNo">2019</span><a name="line.2019"></a> |
| <span class="sourceLineNo">2020</span> /**<a name="line.2020"></a> |
| <span class="sourceLineNo">2021</span> * @return This HFileBlocks fileContext which will a derivative of the fileContext for the file<a name="line.2021"></a> |
| <span class="sourceLineNo">2022</span> * from which this block's data was originally read.<a name="line.2022"></a> |
| <span class="sourceLineNo">2023</span> */<a name="line.2023"></a> |
| <span class="sourceLineNo">2024</span> public HFileContext getHFileContext() {<a name="line.2024"></a> |
| <span class="sourceLineNo">2025</span> return this.fileContext;<a name="line.2025"></a> |
| <span class="sourceLineNo">2026</span> }<a name="line.2026"></a> |
| <span class="sourceLineNo">2027</span><a name="line.2027"></a> |
| <span class="sourceLineNo">2028</span> /**<a name="line.2028"></a> |
| <span class="sourceLineNo">2029</span> * Convert the contents of the block header into a human readable string. This is mostly helpful<a name="line.2029"></a> |
| <span class="sourceLineNo">2030</span> * for debugging. This assumes that the block has minor version > 0.<a name="line.2030"></a> |
| <span class="sourceLineNo">2031</span> */<a name="line.2031"></a> |
| <span class="sourceLineNo">2032</span> static String toStringHeader(ByteBuff buf) throws IOException {<a name="line.2032"></a> |
| <span class="sourceLineNo">2033</span> byte[] magicBuf = new byte[Math.min(buf.limit() - buf.position(), BlockType.MAGIC_LENGTH)];<a name="line.2033"></a> |
| <span class="sourceLineNo">2034</span> buf.get(magicBuf);<a name="line.2034"></a> |
| <span class="sourceLineNo">2035</span> BlockType bt = BlockType.parse(magicBuf, 0, BlockType.MAGIC_LENGTH);<a name="line.2035"></a> |
| <span class="sourceLineNo">2036</span> int compressedBlockSizeNoHeader = buf.getInt();<a name="line.2036"></a> |
| <span class="sourceLineNo">2037</span> int uncompressedBlockSizeNoHeader = buf.getInt();<a name="line.2037"></a> |
| <span class="sourceLineNo">2038</span> long prevBlockOffset = buf.getLong();<a name="line.2038"></a> |
| <span class="sourceLineNo">2039</span> byte cksumtype = buf.get();<a name="line.2039"></a> |
| <span class="sourceLineNo">2040</span> long bytesPerChecksum = buf.getInt();<a name="line.2040"></a> |
| <span class="sourceLineNo">2041</span> long onDiskDataSizeWithHeader = buf.getInt();<a name="line.2041"></a> |
| <span class="sourceLineNo">2042</span> return " Header dump: magic: " + Bytes.toString(magicBuf) + " blockType " + bt<a name="line.2042"></a> |
| <span class="sourceLineNo">2043</span> + " compressedBlockSizeNoHeader " + compressedBlockSizeNoHeader<a name="line.2043"></a> |
| <span class="sourceLineNo">2044</span> + " uncompressedBlockSizeNoHeader " + uncompressedBlockSizeNoHeader + " prevBlockOffset "<a name="line.2044"></a> |
| <span class="sourceLineNo">2045</span> + prevBlockOffset + " checksumType " + ChecksumType.codeToType(cksumtype)<a name="line.2045"></a> |
| <span class="sourceLineNo">2046</span> + " bytesPerChecksum " + bytesPerChecksum + " onDiskDataSizeWithHeader "<a name="line.2046"></a> |
| <span class="sourceLineNo">2047</span> + onDiskDataSizeWithHeader;<a name="line.2047"></a> |
| <span class="sourceLineNo">2048</span> }<a name="line.2048"></a> |
| <span class="sourceLineNo">2049</span><a name="line.2049"></a> |
| <span class="sourceLineNo">2050</span> /**<a name="line.2050"></a> |
| <span class="sourceLineNo">2051</span> * Creates a new HFileBlockBuilder from the existing block and a new ByteBuff. The builder will be<a name="line.2051"></a> |
| <span class="sourceLineNo">2052</span> * loaded with all of the original fields from blk, except now using the newBuff and setting<a name="line.2052"></a> |
| <span class="sourceLineNo">2053</span> * isSharedMem based on the source of the passed in newBuff. An existing HFileBlock may have been<a name="line.2053"></a> |
| <span class="sourceLineNo">2054</span> * an {@link ExclusiveMemHFileBlock}, but the new buffer might call for a<a name="line.2054"></a> |
| <span class="sourceLineNo">2055</span> * {@link SharedMemHFileBlock}. Or vice versa.<a name="line.2055"></a> |
| <span class="sourceLineNo">2056</span> * @param blk the block to clone from<a name="line.2056"></a> |
| <span class="sourceLineNo">2057</span> * @param newBuff the new buffer to use<a name="line.2057"></a> |
| <span class="sourceLineNo">2058</span> */<a name="line.2058"></a> |
| <span class="sourceLineNo">2059</span> private static HFileBlockBuilder createBuilder(HFileBlock blk, ByteBuff newBuff) {<a name="line.2059"></a> |
| <span class="sourceLineNo">2060</span> return new HFileBlockBuilder().withBlockType(blk.blockType)<a name="line.2060"></a> |
| <span class="sourceLineNo">2061</span> .withOnDiskSizeWithoutHeader(blk.onDiskSizeWithoutHeader)<a name="line.2061"></a> |
| <span class="sourceLineNo">2062</span> .withUncompressedSizeWithoutHeader(blk.uncompressedSizeWithoutHeader)<a name="line.2062"></a> |
| <span class="sourceLineNo">2063</span> .withPrevBlockOffset(blk.prevBlockOffset).withByteBuff(newBuff).withOffset(blk.offset)<a name="line.2063"></a> |
| <span class="sourceLineNo">2064</span> .withOnDiskDataSizeWithHeader(blk.onDiskDataSizeWithHeader)<a name="line.2064"></a> |
| <span class="sourceLineNo">2065</span> .withNextBlockOnDiskSize(blk.nextBlockOnDiskSize).withHFileContext(blk.fileContext)<a name="line.2065"></a> |
| <span class="sourceLineNo">2066</span> .withByteBuffAllocator(blk.allocator).withShared(!newBuff.hasArray());<a name="line.2066"></a> |
| <span class="sourceLineNo">2067</span> }<a name="line.2067"></a> |
| <span class="sourceLineNo">2068</span><a name="line.2068"></a> |
| <span class="sourceLineNo">2069</span> private static HFileBlock shallowClone(HFileBlock blk, ByteBuff newBuf) {<a name="line.2069"></a> |
| <span class="sourceLineNo">2070</span> return createBuilder(blk, newBuf).build();<a name="line.2070"></a> |
| <span class="sourceLineNo">2071</span> }<a name="line.2071"></a> |
| <span class="sourceLineNo">2072</span><a name="line.2072"></a> |
| <span class="sourceLineNo">2073</span> static HFileBlock deepCloneOnHeap(HFileBlock blk) {<a name="line.2073"></a> |
| <span class="sourceLineNo">2074</span> ByteBuff deepCloned = ByteBuff.wrap(ByteBuffer.wrap(blk.buf.toBytes(0, blk.buf.limit())));<a name="line.2074"></a> |
| <span class="sourceLineNo">2075</span> return createBuilder(blk, deepCloned).build();<a name="line.2075"></a> |
| <span class="sourceLineNo">2076</span> }<a name="line.2076"></a> |
| <span class="sourceLineNo">2077</span>}<a name="line.2077"></a> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| </pre> |
| </div> |
| </body> |
| </html> |
