| <!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.util;<a name="line.18"></a> |
| <span class="sourceLineNo">019</span><a name="line.19"></a> |
| <span class="sourceLineNo">020</span>import java.util.ArrayList;<a name="line.20"></a> |
| <span class="sourceLineNo">021</span>import java.util.Collection;<a name="line.21"></a> |
| <span class="sourceLineNo">022</span>import java.util.Comparator;<a name="line.22"></a> |
| <span class="sourceLineNo">023</span>import java.util.List;<a name="line.23"></a> |
| <span class="sourceLineNo">024</span>import java.util.Map.Entry;<a name="line.24"></a> |
| <span class="sourceLineNo">025</span>import java.util.TreeMap;<a name="line.25"></a> |
| <span class="sourceLineNo">026</span>import java.util.TreeSet;<a name="line.26"></a> |
| <span class="sourceLineNo">027</span>import org.apache.hadoop.hbase.util.Bytes.ByteArrayComparator;<a name="line.27"></a> |
| <span class="sourceLineNo">028</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.28"></a> |
| <span class="sourceLineNo">029</span>import org.slf4j.Logger;<a name="line.29"></a> |
| <span class="sourceLineNo">030</span>import org.slf4j.LoggerFactory;<a name="line.30"></a> |
| <span class="sourceLineNo">031</span><a name="line.31"></a> |
| <span class="sourceLineNo">032</span>import org.apache.hbase.thirdparty.com.google.common.collect.ArrayListMultimap;<a name="line.32"></a> |
| <span class="sourceLineNo">033</span>import org.apache.hbase.thirdparty.com.google.common.collect.Multimap;<a name="line.33"></a> |
| <span class="sourceLineNo">034</span>import org.apache.hbase.thirdparty.com.google.common.collect.TreeMultimap;<a name="line.34"></a> |
| <span class="sourceLineNo">035</span><a name="line.35"></a> |
| <span class="sourceLineNo">036</span>/**<a name="line.36"></a> |
| <span class="sourceLineNo">037</span> * This is a generic region split calculator. It requires Ranges that provide start, end, and a<a name="line.37"></a> |
| <span class="sourceLineNo">038</span> * comparator. It works in two phases -- the first adds ranges and rejects backwards ranges. Then<a name="line.38"></a> |
| <span class="sourceLineNo">039</span> * one calls calcRegions to generate the multimap that has a start split key as a key and possibly<a name="line.39"></a> |
| <span class="sourceLineNo">040</span> * multiple Ranges as members. To traverse, one normally would get the split set, and iterate<a name="line.40"></a> |
| <span class="sourceLineNo">041</span> * through the calcRegions. Normal regions would have only one entry, holes would have zero, and any<a name="line.41"></a> |
| <span class="sourceLineNo">042</span> * overlaps would have multiple entries. The interface is a bit cumbersome currently but is exposed<a name="line.42"></a> |
| <span class="sourceLineNo">043</span> * this way so that clients can choose how to iterate through the region splits.<a name="line.43"></a> |
| <span class="sourceLineNo">044</span> * @param <R><a name="line.44"></a> |
| <span class="sourceLineNo">045</span> */<a name="line.45"></a> |
| <span class="sourceLineNo">046</span>@InterfaceAudience.Private<a name="line.46"></a> |
| <span class="sourceLineNo">047</span>public class RegionSplitCalculator<R extends KeyRange> {<a name="line.47"></a> |
| <span class="sourceLineNo">048</span> private static final Logger LOG = LoggerFactory.getLogger(RegionSplitCalculator.class);<a name="line.48"></a> |
| <span class="sourceLineNo">049</span><a name="line.49"></a> |
| <span class="sourceLineNo">050</span> private final Comparator<R> rangeCmp;<a name="line.50"></a> |
| <span class="sourceLineNo">051</span> /**<a name="line.51"></a> |
| <span class="sourceLineNo">052</span> * This contains a sorted set of all the possible split points Invariant: once populated this has<a name="line.52"></a> |
| <span class="sourceLineNo">053</span> * 0 entries if empty or at most n+1 values where n == number of added ranges.<a name="line.53"></a> |
| <span class="sourceLineNo">054</span> */<a name="line.54"></a> |
| <span class="sourceLineNo">055</span> private final TreeSet<byte[]> splits = new TreeSet<>(BYTES_COMPARATOR);<a name="line.55"></a> |
| <span class="sourceLineNo">056</span><a name="line.56"></a> |
| <span class="sourceLineNo">057</span> /**<a name="line.57"></a> |
| <span class="sourceLineNo">058</span> * This is a map from start key to regions with the same start key. Invariant: This always have n<a name="line.58"></a> |
| <span class="sourceLineNo">059</span> * values in total<a name="line.59"></a> |
| <span class="sourceLineNo">060</span> */<a name="line.60"></a> |
| <span class="sourceLineNo">061</span> private final Multimap<byte[], R> starts = ArrayListMultimap.create();<a name="line.61"></a> |
| <span class="sourceLineNo">062</span><a name="line.62"></a> |
| <span class="sourceLineNo">063</span> /**<a name="line.63"></a> |
| <span class="sourceLineNo">064</span> * SPECIAL CASE<a name="line.64"></a> |
| <span class="sourceLineNo">065</span> */<a name="line.65"></a> |
| <span class="sourceLineNo">066</span> private final static byte[] ENDKEY = null;<a name="line.66"></a> |
| <span class="sourceLineNo">067</span><a name="line.67"></a> |
| <span class="sourceLineNo">068</span> public RegionSplitCalculator(Comparator<R> cmp) {<a name="line.68"></a> |
| <span class="sourceLineNo">069</span> rangeCmp = cmp;<a name="line.69"></a> |
| <span class="sourceLineNo">070</span> }<a name="line.70"></a> |
| <span class="sourceLineNo">071</span><a name="line.71"></a> |
| <span class="sourceLineNo">072</span> public final static Comparator<byte[]> BYTES_COMPARATOR = new ByteArrayComparator() {<a name="line.72"></a> |
| <span class="sourceLineNo">073</span> @Override<a name="line.73"></a> |
| <span class="sourceLineNo">074</span> public int compare(byte[] l, byte[] r) {<a name="line.74"></a> |
| <span class="sourceLineNo">075</span> if (l == null && r == null) return 0;<a name="line.75"></a> |
| <span class="sourceLineNo">076</span> if (l == null) return 1;<a name="line.76"></a> |
| <span class="sourceLineNo">077</span> if (r == null) return -1;<a name="line.77"></a> |
| <span class="sourceLineNo">078</span> return super.compare(l, r);<a name="line.78"></a> |
| <span class="sourceLineNo">079</span> }<a name="line.79"></a> |
| <span class="sourceLineNo">080</span> };<a name="line.80"></a> |
| <span class="sourceLineNo">081</span><a name="line.81"></a> |
| <span class="sourceLineNo">082</span> /**<a name="line.82"></a> |
| <span class="sourceLineNo">083</span> * SPECIAL CASE wrapper for empty end key<a name="line.83"></a> |
| <span class="sourceLineNo">084</span> * @return ENDKEY if end key is empty, else normal endkey.<a name="line.84"></a> |
| <span class="sourceLineNo">085</span> */<a name="line.85"></a> |
| <span class="sourceLineNo">086</span> private static <R extends KeyRange> byte[] specialEndKey(R range) {<a name="line.86"></a> |
| <span class="sourceLineNo">087</span> byte[] end = range.getEndKey();<a name="line.87"></a> |
| <span class="sourceLineNo">088</span> if (end.length == 0) {<a name="line.88"></a> |
| <span class="sourceLineNo">089</span> return ENDKEY;<a name="line.89"></a> |
| <span class="sourceLineNo">090</span> }<a name="line.90"></a> |
| <span class="sourceLineNo">091</span> return end;<a name="line.91"></a> |
| <span class="sourceLineNo">092</span> }<a name="line.92"></a> |
| <span class="sourceLineNo">093</span><a name="line.93"></a> |
| <span class="sourceLineNo">094</span> /**<a name="line.94"></a> |
| <span class="sourceLineNo">095</span> * Adds an edge to the split calculator<a name="line.95"></a> |
| <span class="sourceLineNo">096</span> * @return true if is included, false if backwards/invalid<a name="line.96"></a> |
| <span class="sourceLineNo">097</span> */<a name="line.97"></a> |
| <span class="sourceLineNo">098</span> public boolean add(R range) {<a name="line.98"></a> |
| <span class="sourceLineNo">099</span> byte[] start = range.getStartKey();<a name="line.99"></a> |
| <span class="sourceLineNo">100</span> byte[] end = specialEndKey(range);<a name="line.100"></a> |
| <span class="sourceLineNo">101</span><a name="line.101"></a> |
| <span class="sourceLineNo">102</span> // No need to use Arrays.equals because ENDKEY is null<a name="line.102"></a> |
| <span class="sourceLineNo">103</span> if (end != ENDKEY && Bytes.compareTo(start, end) > 0) {<a name="line.103"></a> |
| <span class="sourceLineNo">104</span> // don't allow backwards edges<a name="line.104"></a> |
| <span class="sourceLineNo">105</span> LOG.debug(<a name="line.105"></a> |
| <span class="sourceLineNo">106</span> "attempted to add backwards edge: " + Bytes.toString(start) + " " + Bytes.toString(end));<a name="line.106"></a> |
| <span class="sourceLineNo">107</span> return false;<a name="line.107"></a> |
| <span class="sourceLineNo">108</span> }<a name="line.108"></a> |
| <span class="sourceLineNo">109</span><a name="line.109"></a> |
| <span class="sourceLineNo">110</span> splits.add(start);<a name="line.110"></a> |
| <span class="sourceLineNo">111</span> splits.add(end);<a name="line.111"></a> |
| <span class="sourceLineNo">112</span> starts.put(start, range);<a name="line.112"></a> |
| <span class="sourceLineNo">113</span> return true;<a name="line.113"></a> |
| <span class="sourceLineNo">114</span> }<a name="line.114"></a> |
| <span class="sourceLineNo">115</span><a name="line.115"></a> |
| <span class="sourceLineNo">116</span> /**<a name="line.116"></a> |
| <span class="sourceLineNo">117</span> * Generates a coverage multimap from split key to Regions that start with the split key.<a name="line.117"></a> |
| <span class="sourceLineNo">118</span> * @return coverage multimap<a name="line.118"></a> |
| <span class="sourceLineNo">119</span> */<a name="line.119"></a> |
| <span class="sourceLineNo">120</span> public Multimap<byte[], R> calcCoverage() {<a name="line.120"></a> |
| <span class="sourceLineNo">121</span> // This needs to be sorted to force the use of the comparator on the values,<a name="line.121"></a> |
| <span class="sourceLineNo">122</span> // otherwise byte array comparison isn't used<a name="line.122"></a> |
| <span class="sourceLineNo">123</span> Multimap<byte[], R> regions = TreeMultimap.create(BYTES_COMPARATOR, rangeCmp);<a name="line.123"></a> |
| <span class="sourceLineNo">124</span><a name="line.124"></a> |
| <span class="sourceLineNo">125</span> // march through all splits from the start points<a name="line.125"></a> |
| <span class="sourceLineNo">126</span> for (Entry<byte[], Collection<R>> start : starts.asMap().entrySet()) {<a name="line.126"></a> |
| <span class="sourceLineNo">127</span> byte[] key = start.getKey();<a name="line.127"></a> |
| <span class="sourceLineNo">128</span> for (R r : start.getValue()) {<a name="line.128"></a> |
| <span class="sourceLineNo">129</span> regions.put(key, r);<a name="line.129"></a> |
| <span class="sourceLineNo">130</span><a name="line.130"></a> |
| <span class="sourceLineNo">131</span> for (byte[] coveredSplit : splits.subSet(r.getStartKey(), specialEndKey(r))) {<a name="line.131"></a> |
| <span class="sourceLineNo">132</span> regions.put(coveredSplit, r);<a name="line.132"></a> |
| <span class="sourceLineNo">133</span> }<a name="line.133"></a> |
| <span class="sourceLineNo">134</span> }<a name="line.134"></a> |
| <span class="sourceLineNo">135</span> }<a name="line.135"></a> |
| <span class="sourceLineNo">136</span> return regions;<a name="line.136"></a> |
| <span class="sourceLineNo">137</span> }<a name="line.137"></a> |
| <span class="sourceLineNo">138</span><a name="line.138"></a> |
| <span class="sourceLineNo">139</span> public TreeSet<byte[]> getSplits() {<a name="line.139"></a> |
| <span class="sourceLineNo">140</span> return splits;<a name="line.140"></a> |
| <span class="sourceLineNo">141</span> }<a name="line.141"></a> |
| <span class="sourceLineNo">142</span><a name="line.142"></a> |
| <span class="sourceLineNo">143</span> public Multimap<byte[], R> getStarts() {<a name="line.143"></a> |
| <span class="sourceLineNo">144</span> return starts;<a name="line.144"></a> |
| <span class="sourceLineNo">145</span> }<a name="line.145"></a> |
| <span class="sourceLineNo">146</span><a name="line.146"></a> |
| <span class="sourceLineNo">147</span> /**<a name="line.147"></a> |
| <span class="sourceLineNo">148</span> * Find specified number of top ranges in a big overlap group. It could return less if there are<a name="line.148"></a> |
| <span class="sourceLineNo">149</span> * not that many top ranges. Once these top ranges are excluded, the big overlap group will be<a name="line.149"></a> |
| <span class="sourceLineNo">150</span> * broken into ranges with no overlapping, or smaller overlapped groups, and most likely some<a name="line.150"></a> |
| <span class="sourceLineNo">151</span> * holes.<a name="line.151"></a> |
| <span class="sourceLineNo">152</span> * @param bigOverlap a list of ranges that overlap with each other<a name="line.152"></a> |
| <span class="sourceLineNo">153</span> * @param count the max number of ranges to find<a name="line.153"></a> |
| <span class="sourceLineNo">154</span> * @return a list of ranges that overlap with most others<a name="line.154"></a> |
| <span class="sourceLineNo">155</span> */<a name="line.155"></a> |
| <span class="sourceLineNo">156</span> public static <R extends KeyRange> List<R> findBigRanges(Collection<R> bigOverlap, int count) {<a name="line.156"></a> |
| <span class="sourceLineNo">157</span> List<R> bigRanges = new ArrayList<>();<a name="line.157"></a> |
| <span class="sourceLineNo">158</span><a name="line.158"></a> |
| <span class="sourceLineNo">159</span> // The key is the count of overlaps,<a name="line.159"></a> |
| <span class="sourceLineNo">160</span> // The value is a list of ranges that have that many overlaps<a name="line.160"></a> |
| <span class="sourceLineNo">161</span> TreeMap<Integer, List<R>> overlapRangeMap = new TreeMap<>();<a name="line.161"></a> |
| <span class="sourceLineNo">162</span> for (R r : bigOverlap) {<a name="line.162"></a> |
| <span class="sourceLineNo">163</span> // Calculates the # of overlaps for each region<a name="line.163"></a> |
| <span class="sourceLineNo">164</span> // and populates rangeOverlapMap<a name="line.164"></a> |
| <span class="sourceLineNo">165</span> byte[] startKey = r.getStartKey();<a name="line.165"></a> |
| <span class="sourceLineNo">166</span> byte[] endKey = specialEndKey(r);<a name="line.166"></a> |
| <span class="sourceLineNo">167</span><a name="line.167"></a> |
| <span class="sourceLineNo">168</span> int overlappedRegions = 0;<a name="line.168"></a> |
| <span class="sourceLineNo">169</span> for (R rr : bigOverlap) {<a name="line.169"></a> |
| <span class="sourceLineNo">170</span> byte[] start = rr.getStartKey();<a name="line.170"></a> |
| <span class="sourceLineNo">171</span> byte[] end = specialEndKey(rr);<a name="line.171"></a> |
| <span class="sourceLineNo">172</span><a name="line.172"></a> |
| <span class="sourceLineNo">173</span> if (<a name="line.173"></a> |
| <span class="sourceLineNo">174</span> BYTES_COMPARATOR.compare(startKey, end) < 0 && BYTES_COMPARATOR.compare(endKey, start) > 0<a name="line.174"></a> |
| <span class="sourceLineNo">175</span> ) {<a name="line.175"></a> |
| <span class="sourceLineNo">176</span> overlappedRegions++;<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><a name="line.179"></a> |
| <span class="sourceLineNo">180</span> // One region always overlaps with itself,<a name="line.180"></a> |
| <span class="sourceLineNo">181</span> // so overlappedRegions should be more than 1<a name="line.181"></a> |
| <span class="sourceLineNo">182</span> // for actual overlaps.<a name="line.182"></a> |
| <span class="sourceLineNo">183</span> if (overlappedRegions > 1) {<a name="line.183"></a> |
| <span class="sourceLineNo">184</span> Integer key = Integer.valueOf(overlappedRegions);<a name="line.184"></a> |
| <span class="sourceLineNo">185</span> List<R> ranges = overlapRangeMap.get(key);<a name="line.185"></a> |
| <span class="sourceLineNo">186</span> if (ranges == null) {<a name="line.186"></a> |
| <span class="sourceLineNo">187</span> ranges = new ArrayList<>();<a name="line.187"></a> |
| <span class="sourceLineNo">188</span> overlapRangeMap.put(key, ranges);<a name="line.188"></a> |
| <span class="sourceLineNo">189</span> }<a name="line.189"></a> |
| <span class="sourceLineNo">190</span> ranges.add(r);<a name="line.190"></a> |
| <span class="sourceLineNo">191</span> }<a name="line.191"></a> |
| <span class="sourceLineNo">192</span> }<a name="line.192"></a> |
| <span class="sourceLineNo">193</span> int toBeAdded = count;<a name="line.193"></a> |
| <span class="sourceLineNo">194</span> for (Integer key : overlapRangeMap.descendingKeySet()) {<a name="line.194"></a> |
| <span class="sourceLineNo">195</span> List<R> chunk = overlapRangeMap.get(key);<a name="line.195"></a> |
| <span class="sourceLineNo">196</span> int chunkSize = chunk.size();<a name="line.196"></a> |
| <span class="sourceLineNo">197</span> if (chunkSize <= toBeAdded) {<a name="line.197"></a> |
| <span class="sourceLineNo">198</span> bigRanges.addAll(chunk);<a name="line.198"></a> |
| <span class="sourceLineNo">199</span> toBeAdded -= chunkSize;<a name="line.199"></a> |
| <span class="sourceLineNo">200</span> if (toBeAdded > 0) continue;<a name="line.200"></a> |
| <span class="sourceLineNo">201</span> } else {<a name="line.201"></a> |
| <span class="sourceLineNo">202</span> // Try to use the middle chunk in case the overlapping is<a name="line.202"></a> |
| <span class="sourceLineNo">203</span> // chained, for example: [a, c), [b, e), [d, g), [f h)...<a name="line.203"></a> |
| <span class="sourceLineNo">204</span> // In such a case, sideline the middle chunk will break<a name="line.204"></a> |
| <span class="sourceLineNo">205</span> // the group efficiently.<a name="line.205"></a> |
| <span class="sourceLineNo">206</span> int start = (chunkSize - toBeAdded) / 2;<a name="line.206"></a> |
| <span class="sourceLineNo">207</span> int end = start + toBeAdded;<a name="line.207"></a> |
| <span class="sourceLineNo">208</span> for (int i = start; i < end; i++) {<a name="line.208"></a> |
| <span class="sourceLineNo">209</span> bigRanges.add(chunk.get(i));<a name="line.209"></a> |
| <span class="sourceLineNo">210</span> }<a name="line.210"></a> |
| <span class="sourceLineNo">211</span> }<a name="line.211"></a> |
| <span class="sourceLineNo">212</span> break;<a name="line.212"></a> |
| <span class="sourceLineNo">213</span> }<a name="line.213"></a> |
| <span class="sourceLineNo">214</span> return bigRanges;<a name="line.214"></a> |
| <span class="sourceLineNo">215</span> }<a name="line.215"></a> |
| <span class="sourceLineNo">216</span>}<a name="line.216"></a> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| </pre> |
| </div> |
| </body> |
| </html> |
