blob: 14869642ec0d635897d1b66282675afb8836b6f9 [file] [log] [blame]
<!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><a name="line.18"></a>
<span class="sourceLineNo">019</span>package org.apache.hadoop.hbase.mapreduce;<a name="line.19"></a>
<span class="sourceLineNo">020</span><a name="line.20"></a>
<span class="sourceLineNo">021</span>import java.io.ByteArrayOutputStream;<a name="line.21"></a>
<span class="sourceLineNo">022</span>import java.io.DataInput;<a name="line.22"></a>
<span class="sourceLineNo">023</span>import java.io.DataOutput;<a name="line.23"></a>
<span class="sourceLineNo">024</span>import java.io.IOException;<a name="line.24"></a>
<span class="sourceLineNo">025</span>import java.lang.reflect.InvocationTargetException;<a name="line.25"></a>
<span class="sourceLineNo">026</span>import java.util.ArrayList;<a name="line.26"></a>
<span class="sourceLineNo">027</span>import java.util.List;<a name="line.27"></a>
<span class="sourceLineNo">028</span>import java.util.UUID;<a name="line.28"></a>
<span class="sourceLineNo">029</span><a name="line.29"></a>
<span class="sourceLineNo">030</span>import org.apache.hadoop.conf.Configuration;<a name="line.30"></a>
<span class="sourceLineNo">031</span>import org.apache.hadoop.fs.FileSystem;<a name="line.31"></a>
<span class="sourceLineNo">032</span>import org.apache.hadoop.fs.Path;<a name="line.32"></a>
<span class="sourceLineNo">033</span>import org.apache.hadoop.hbase.HDFSBlocksDistribution;<a name="line.33"></a>
<span class="sourceLineNo">034</span>import org.apache.hadoop.hbase.HDFSBlocksDistribution.HostAndWeight;<a name="line.34"></a>
<span class="sourceLineNo">035</span>import org.apache.hadoop.hbase.HRegionInfo;<a name="line.35"></a>
<span class="sourceLineNo">036</span>import org.apache.hadoop.hbase.PrivateCellUtil;<a name="line.36"></a>
<span class="sourceLineNo">037</span>import org.apache.hadoop.hbase.client.ClientSideRegionScanner;<a name="line.37"></a>
<span class="sourceLineNo">038</span>import org.apache.hadoop.hbase.client.IsolationLevel;<a name="line.38"></a>
<span class="sourceLineNo">039</span>import org.apache.hadoop.hbase.client.Result;<a name="line.39"></a>
<span class="sourceLineNo">040</span>import org.apache.hadoop.hbase.client.Scan;<a name="line.40"></a>
<span class="sourceLineNo">041</span>import org.apache.hadoop.hbase.client.TableDescriptor;<a name="line.41"></a>
<span class="sourceLineNo">042</span>import org.apache.hadoop.hbase.io.ImmutableBytesWritable;<a name="line.42"></a>
<span class="sourceLineNo">043</span>import org.apache.hadoop.hbase.regionserver.HRegion;<a name="line.43"></a>
<span class="sourceLineNo">044</span>import org.apache.hadoop.hbase.snapshot.RestoreSnapshotHelper;<a name="line.44"></a>
<span class="sourceLineNo">045</span>import org.apache.hadoop.hbase.snapshot.SnapshotDescriptionUtils;<a name="line.45"></a>
<span class="sourceLineNo">046</span>import org.apache.hadoop.hbase.snapshot.SnapshotManifest;<a name="line.46"></a>
<span class="sourceLineNo">047</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.47"></a>
<span class="sourceLineNo">048</span>import org.apache.hadoop.hbase.util.FSUtils;<a name="line.48"></a>
<span class="sourceLineNo">049</span>import org.apache.hadoop.hbase.util.RegionSplitter;<a name="line.49"></a>
<span class="sourceLineNo">050</span>import org.apache.hadoop.io.Writable;<a name="line.50"></a>
<span class="sourceLineNo">051</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.51"></a>
<span class="sourceLineNo">052</span>import org.slf4j.Logger;<a name="line.52"></a>
<span class="sourceLineNo">053</span>import org.slf4j.LoggerFactory;<a name="line.53"></a>
<span class="sourceLineNo">054</span><a name="line.54"></a>
<span class="sourceLineNo">055</span>import org.apache.hbase.thirdparty.com.google.common.collect.Lists;<a name="line.55"></a>
<span class="sourceLineNo">056</span>import org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil;<a name="line.56"></a>
<span class="sourceLineNo">057</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.MapReduceProtos.TableSnapshotRegionSplit;<a name="line.57"></a>
<span class="sourceLineNo">058</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.SnapshotProtos.SnapshotDescription;<a name="line.58"></a>
<span class="sourceLineNo">059</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.SnapshotProtos.SnapshotRegionManifest;<a name="line.59"></a>
<span class="sourceLineNo">060</span><a name="line.60"></a>
<span class="sourceLineNo">061</span>/**<a name="line.61"></a>
<span class="sourceLineNo">062</span> * Hadoop MR API-agnostic implementation for mapreduce over table snapshots.<a name="line.62"></a>
<span class="sourceLineNo">063</span> */<a name="line.63"></a>
<span class="sourceLineNo">064</span>@InterfaceAudience.Private<a name="line.64"></a>
<span class="sourceLineNo">065</span>public class TableSnapshotInputFormatImpl {<a name="line.65"></a>
<span class="sourceLineNo">066</span> // TODO: Snapshots files are owned in fs by the hbase user. There is no<a name="line.66"></a>
<span class="sourceLineNo">067</span> // easy way to delegate access.<a name="line.67"></a>
<span class="sourceLineNo">068</span><a name="line.68"></a>
<span class="sourceLineNo">069</span> public static final Logger LOG = LoggerFactory.getLogger(TableSnapshotInputFormatImpl.class);<a name="line.69"></a>
<span class="sourceLineNo">070</span><a name="line.70"></a>
<span class="sourceLineNo">071</span> private static final String SNAPSHOT_NAME_KEY = "hbase.TableSnapshotInputFormat.snapshot.name";<a name="line.71"></a>
<span class="sourceLineNo">072</span> // key for specifying the root dir of the restored snapshot<a name="line.72"></a>
<span class="sourceLineNo">073</span> protected static final String RESTORE_DIR_KEY = "hbase.TableSnapshotInputFormat.restore.dir";<a name="line.73"></a>
<span class="sourceLineNo">074</span><a name="line.74"></a>
<span class="sourceLineNo">075</span> /** See {@link #getBestLocations(Configuration, HDFSBlocksDistribution, int)} */<a name="line.75"></a>
<span class="sourceLineNo">076</span> private static final String LOCALITY_CUTOFF_MULTIPLIER =<a name="line.76"></a>
<span class="sourceLineNo">077</span> "hbase.tablesnapshotinputformat.locality.cutoff.multiplier";<a name="line.77"></a>
<span class="sourceLineNo">078</span> private static final float DEFAULT_LOCALITY_CUTOFF_MULTIPLIER = 0.8f;<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> * For MapReduce jobs running multiple mappers per region, determines<a name="line.81"></a>
<span class="sourceLineNo">082</span> * what split algorithm we should be using to find split points for scanners.<a name="line.82"></a>
<span class="sourceLineNo">083</span> */<a name="line.83"></a>
<span class="sourceLineNo">084</span> public static final String SPLIT_ALGO = "hbase.mapreduce.split.algorithm";<a name="line.84"></a>
<span class="sourceLineNo">085</span> /**<a name="line.85"></a>
<span class="sourceLineNo">086</span> * For MapReduce jobs running multiple mappers per region, determines<a name="line.86"></a>
<span class="sourceLineNo">087</span> * number of splits to generate per region.<a name="line.87"></a>
<span class="sourceLineNo">088</span> */<a name="line.88"></a>
<span class="sourceLineNo">089</span> public static final String NUM_SPLITS_PER_REGION = "hbase.mapreduce.splits.per.region";<a name="line.89"></a>
<span class="sourceLineNo">090</span><a name="line.90"></a>
<span class="sourceLineNo">091</span> /**<a name="line.91"></a>
<span class="sourceLineNo">092</span> * Whether to calculate the block location for splits. Default to true.<a name="line.92"></a>
<span class="sourceLineNo">093</span> * If the computing layer runs outside of HBase cluster, the block locality does not master.<a name="line.93"></a>
<span class="sourceLineNo">094</span> * Setting this value to false could skip the calculation and save some time.<a name="line.94"></a>
<span class="sourceLineNo">095</span> *<a name="line.95"></a>
<span class="sourceLineNo">096</span> * Set access modifier to "public" so that these could be accessed by test classes of<a name="line.96"></a>
<span class="sourceLineNo">097</span> * both org.apache.hadoop.hbase.mapred<a name="line.97"></a>
<span class="sourceLineNo">098</span> * and org.apache.hadoop.hbase.mapreduce.<a name="line.98"></a>
<span class="sourceLineNo">099</span> */<a name="line.99"></a>
<span class="sourceLineNo">100</span> public static final String SNAPSHOT_INPUTFORMAT_LOCALITY_ENABLED_KEY =<a name="line.100"></a>
<span class="sourceLineNo">101</span> "hbase.TableSnapshotInputFormat.locality.enabled";<a name="line.101"></a>
<span class="sourceLineNo">102</span> public static final boolean SNAPSHOT_INPUTFORMAT_LOCALITY_ENABLED_DEFAULT = true;<a name="line.102"></a>
<span class="sourceLineNo">103</span><a name="line.103"></a>
<span class="sourceLineNo">104</span> /**<a name="line.104"></a>
<span class="sourceLineNo">105</span> * Implementation class for InputSplit logic common between mapred and mapreduce.<a name="line.105"></a>
<span class="sourceLineNo">106</span> */<a name="line.106"></a>
<span class="sourceLineNo">107</span> public static class InputSplit implements Writable {<a name="line.107"></a>
<span class="sourceLineNo">108</span><a name="line.108"></a>
<span class="sourceLineNo">109</span> private TableDescriptor htd;<a name="line.109"></a>
<span class="sourceLineNo">110</span> private HRegionInfo regionInfo;<a name="line.110"></a>
<span class="sourceLineNo">111</span> private String[] locations;<a name="line.111"></a>
<span class="sourceLineNo">112</span> private String scan;<a name="line.112"></a>
<span class="sourceLineNo">113</span> private String restoreDir;<a name="line.113"></a>
<span class="sourceLineNo">114</span><a name="line.114"></a>
<span class="sourceLineNo">115</span> // constructor for mapreduce framework / Writable<a name="line.115"></a>
<span class="sourceLineNo">116</span> public InputSplit() {}<a name="line.116"></a>
<span class="sourceLineNo">117</span><a name="line.117"></a>
<span class="sourceLineNo">118</span> public InputSplit(TableDescriptor htd, HRegionInfo regionInfo, List&lt;String&gt; locations,<a name="line.118"></a>
<span class="sourceLineNo">119</span> Scan scan, Path restoreDir) {<a name="line.119"></a>
<span class="sourceLineNo">120</span> this.htd = htd;<a name="line.120"></a>
<span class="sourceLineNo">121</span> this.regionInfo = regionInfo;<a name="line.121"></a>
<span class="sourceLineNo">122</span> if (locations == null || locations.isEmpty()) {<a name="line.122"></a>
<span class="sourceLineNo">123</span> this.locations = new String[0];<a name="line.123"></a>
<span class="sourceLineNo">124</span> } else {<a name="line.124"></a>
<span class="sourceLineNo">125</span> this.locations = locations.toArray(new String[locations.size()]);<a name="line.125"></a>
<span class="sourceLineNo">126</span> }<a name="line.126"></a>
<span class="sourceLineNo">127</span> try {<a name="line.127"></a>
<span class="sourceLineNo">128</span> this.scan = scan != null ? TableMapReduceUtil.convertScanToString(scan) : "";<a name="line.128"></a>
<span class="sourceLineNo">129</span> } catch (IOException e) {<a name="line.129"></a>
<span class="sourceLineNo">130</span> LOG.warn("Failed to convert Scan to String", e);<a name="line.130"></a>
<span class="sourceLineNo">131</span> }<a name="line.131"></a>
<span class="sourceLineNo">132</span><a name="line.132"></a>
<span class="sourceLineNo">133</span> this.restoreDir = restoreDir.toString();<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> public TableDescriptor getHtd() {<a name="line.136"></a>
<span class="sourceLineNo">137</span> return htd;<a name="line.137"></a>
<span class="sourceLineNo">138</span> }<a name="line.138"></a>
<span class="sourceLineNo">139</span><a name="line.139"></a>
<span class="sourceLineNo">140</span> public String getScan() {<a name="line.140"></a>
<span class="sourceLineNo">141</span> return scan;<a name="line.141"></a>
<span class="sourceLineNo">142</span> }<a name="line.142"></a>
<span class="sourceLineNo">143</span><a name="line.143"></a>
<span class="sourceLineNo">144</span> public String getRestoreDir() {<a name="line.144"></a>
<span class="sourceLineNo">145</span> return restoreDir;<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> public long getLength() {<a name="line.148"></a>
<span class="sourceLineNo">149</span> //TODO: We can obtain the file sizes of the snapshot here.<a name="line.149"></a>
<span class="sourceLineNo">150</span> return 0;<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> public String[] getLocations() {<a name="line.153"></a>
<span class="sourceLineNo">154</span> return locations;<a name="line.154"></a>
<span class="sourceLineNo">155</span> }<a name="line.155"></a>
<span class="sourceLineNo">156</span><a name="line.156"></a>
<span class="sourceLineNo">157</span> public TableDescriptor getTableDescriptor() {<a name="line.157"></a>
<span class="sourceLineNo">158</span> return htd;<a name="line.158"></a>
<span class="sourceLineNo">159</span> }<a name="line.159"></a>
<span class="sourceLineNo">160</span><a name="line.160"></a>
<span class="sourceLineNo">161</span> public HRegionInfo getRegionInfo() {<a name="line.161"></a>
<span class="sourceLineNo">162</span> return regionInfo;<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> // TODO: We should have ProtobufSerialization in Hadoop, and directly use PB objects instead of<a name="line.165"></a>
<span class="sourceLineNo">166</span> // doing this wrapping with Writables.<a name="line.166"></a>
<span class="sourceLineNo">167</span> @Override<a name="line.167"></a>
<span class="sourceLineNo">168</span> public void write(DataOutput out) throws IOException {<a name="line.168"></a>
<span class="sourceLineNo">169</span> TableSnapshotRegionSplit.Builder builder = TableSnapshotRegionSplit.newBuilder()<a name="line.169"></a>
<span class="sourceLineNo">170</span> .setTable(ProtobufUtil.toTableSchema(htd))<a name="line.170"></a>
<span class="sourceLineNo">171</span> .setRegion(HRegionInfo.convert(regionInfo));<a name="line.171"></a>
<span class="sourceLineNo">172</span><a name="line.172"></a>
<span class="sourceLineNo">173</span> for (String location : locations) {<a name="line.173"></a>
<span class="sourceLineNo">174</span> builder.addLocations(location);<a name="line.174"></a>
<span class="sourceLineNo">175</span> }<a name="line.175"></a>
<span class="sourceLineNo">176</span><a name="line.176"></a>
<span class="sourceLineNo">177</span> TableSnapshotRegionSplit split = builder.build();<a name="line.177"></a>
<span class="sourceLineNo">178</span><a name="line.178"></a>
<span class="sourceLineNo">179</span> ByteArrayOutputStream baos = new ByteArrayOutputStream();<a name="line.179"></a>
<span class="sourceLineNo">180</span> split.writeTo(baos);<a name="line.180"></a>
<span class="sourceLineNo">181</span> baos.close();<a name="line.181"></a>
<span class="sourceLineNo">182</span> byte[] buf = baos.toByteArray();<a name="line.182"></a>
<span class="sourceLineNo">183</span> out.writeInt(buf.length);<a name="line.183"></a>
<span class="sourceLineNo">184</span> out.write(buf);<a name="line.184"></a>
<span class="sourceLineNo">185</span><a name="line.185"></a>
<span class="sourceLineNo">186</span> Bytes.writeByteArray(out, Bytes.toBytes(scan));<a name="line.186"></a>
<span class="sourceLineNo">187</span> Bytes.writeByteArray(out, Bytes.toBytes(restoreDir));<a name="line.187"></a>
<span class="sourceLineNo">188</span><a name="line.188"></a>
<span class="sourceLineNo">189</span> }<a name="line.189"></a>
<span class="sourceLineNo">190</span><a name="line.190"></a>
<span class="sourceLineNo">191</span> @Override<a name="line.191"></a>
<span class="sourceLineNo">192</span> public void readFields(DataInput in) throws IOException {<a name="line.192"></a>
<span class="sourceLineNo">193</span> int len = in.readInt();<a name="line.193"></a>
<span class="sourceLineNo">194</span> byte[] buf = new byte[len];<a name="line.194"></a>
<span class="sourceLineNo">195</span> in.readFully(buf);<a name="line.195"></a>
<span class="sourceLineNo">196</span> TableSnapshotRegionSplit split = TableSnapshotRegionSplit.PARSER.parseFrom(buf);<a name="line.196"></a>
<span class="sourceLineNo">197</span> this.htd = ProtobufUtil.toTableDescriptor(split.getTable());<a name="line.197"></a>
<span class="sourceLineNo">198</span> this.regionInfo = HRegionInfo.convert(split.getRegion());<a name="line.198"></a>
<span class="sourceLineNo">199</span> List&lt;String&gt; locationsList = split.getLocationsList();<a name="line.199"></a>
<span class="sourceLineNo">200</span> this.locations = locationsList.toArray(new String[locationsList.size()]);<a name="line.200"></a>
<span class="sourceLineNo">201</span><a name="line.201"></a>
<span class="sourceLineNo">202</span> this.scan = Bytes.toString(Bytes.readByteArray(in));<a name="line.202"></a>
<span class="sourceLineNo">203</span> this.restoreDir = Bytes.toString(Bytes.readByteArray(in));<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><a name="line.206"></a>
<span class="sourceLineNo">207</span> /**<a name="line.207"></a>
<span class="sourceLineNo">208</span> * Implementation class for RecordReader logic common between mapred and mapreduce.<a name="line.208"></a>
<span class="sourceLineNo">209</span> */<a name="line.209"></a>
<span class="sourceLineNo">210</span> public static class RecordReader {<a name="line.210"></a>
<span class="sourceLineNo">211</span> private InputSplit split;<a name="line.211"></a>
<span class="sourceLineNo">212</span> private Scan scan;<a name="line.212"></a>
<span class="sourceLineNo">213</span> private Result result = null;<a name="line.213"></a>
<span class="sourceLineNo">214</span> private ImmutableBytesWritable row = null;<a name="line.214"></a>
<span class="sourceLineNo">215</span> private ClientSideRegionScanner scanner;<a name="line.215"></a>
<span class="sourceLineNo">216</span><a name="line.216"></a>
<span class="sourceLineNo">217</span> public ClientSideRegionScanner getScanner() {<a name="line.217"></a>
<span class="sourceLineNo">218</span> return scanner;<a name="line.218"></a>
<span class="sourceLineNo">219</span> }<a name="line.219"></a>
<span class="sourceLineNo">220</span><a name="line.220"></a>
<span class="sourceLineNo">221</span> public void initialize(InputSplit split, Configuration conf) throws IOException {<a name="line.221"></a>
<span class="sourceLineNo">222</span> this.scan = TableMapReduceUtil.convertStringToScan(split.getScan());<a name="line.222"></a>
<span class="sourceLineNo">223</span> this.split = split;<a name="line.223"></a>
<span class="sourceLineNo">224</span> TableDescriptor htd = split.htd;<a name="line.224"></a>
<span class="sourceLineNo">225</span> HRegionInfo hri = this.split.getRegionInfo();<a name="line.225"></a>
<span class="sourceLineNo">226</span> FileSystem fs = FSUtils.getCurrentFileSystem(conf);<a name="line.226"></a>
<span class="sourceLineNo">227</span><a name="line.227"></a>
<span class="sourceLineNo">228</span><a name="line.228"></a>
<span class="sourceLineNo">229</span> // region is immutable, this should be fine,<a name="line.229"></a>
<span class="sourceLineNo">230</span> // otherwise we have to set the thread read point<a name="line.230"></a>
<span class="sourceLineNo">231</span> scan.setIsolationLevel(IsolationLevel.READ_UNCOMMITTED);<a name="line.231"></a>
<span class="sourceLineNo">232</span> // disable caching of data blocks<a name="line.232"></a>
<span class="sourceLineNo">233</span> scan.setCacheBlocks(false);<a name="line.233"></a>
<span class="sourceLineNo">234</span> scan.setScanMetricsEnabled(true);<a name="line.234"></a>
<span class="sourceLineNo">235</span><a name="line.235"></a>
<span class="sourceLineNo">236</span> scanner =<a name="line.236"></a>
<span class="sourceLineNo">237</span> new ClientSideRegionScanner(conf, fs, new Path(split.restoreDir), htd, hri, scan, null);<a name="line.237"></a>
<span class="sourceLineNo">238</span> }<a name="line.238"></a>
<span class="sourceLineNo">239</span><a name="line.239"></a>
<span class="sourceLineNo">240</span> public boolean nextKeyValue() throws IOException {<a name="line.240"></a>
<span class="sourceLineNo">241</span> result = scanner.next();<a name="line.241"></a>
<span class="sourceLineNo">242</span> if (result == null) {<a name="line.242"></a>
<span class="sourceLineNo">243</span> //we are done<a name="line.243"></a>
<span class="sourceLineNo">244</span> return false;<a name="line.244"></a>
<span class="sourceLineNo">245</span> }<a name="line.245"></a>
<span class="sourceLineNo">246</span><a name="line.246"></a>
<span class="sourceLineNo">247</span> if (this.row == null) {<a name="line.247"></a>
<span class="sourceLineNo">248</span> this.row = new ImmutableBytesWritable();<a name="line.248"></a>
<span class="sourceLineNo">249</span> }<a name="line.249"></a>
<span class="sourceLineNo">250</span> this.row.set(result.getRow());<a name="line.250"></a>
<span class="sourceLineNo">251</span> return true;<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> public ImmutableBytesWritable getCurrentKey() {<a name="line.254"></a>
<span class="sourceLineNo">255</span> return row;<a name="line.255"></a>
<span class="sourceLineNo">256</span> }<a name="line.256"></a>
<span class="sourceLineNo">257</span><a name="line.257"></a>
<span class="sourceLineNo">258</span> public Result getCurrentValue() {<a name="line.258"></a>
<span class="sourceLineNo">259</span> return result;<a name="line.259"></a>
<span class="sourceLineNo">260</span> }<a name="line.260"></a>
<span class="sourceLineNo">261</span><a name="line.261"></a>
<span class="sourceLineNo">262</span> public long getPos() {<a name="line.262"></a>
<span class="sourceLineNo">263</span> return 0;<a name="line.263"></a>
<span class="sourceLineNo">264</span> }<a name="line.264"></a>
<span class="sourceLineNo">265</span><a name="line.265"></a>
<span class="sourceLineNo">266</span> public float getProgress() {<a name="line.266"></a>
<span class="sourceLineNo">267</span> return 0; // TODO: use total bytes to estimate<a name="line.267"></a>
<span class="sourceLineNo">268</span> }<a name="line.268"></a>
<span class="sourceLineNo">269</span><a name="line.269"></a>
<span class="sourceLineNo">270</span> public void close() {<a name="line.270"></a>
<span class="sourceLineNo">271</span> if (this.scanner != null) {<a name="line.271"></a>
<span class="sourceLineNo">272</span> this.scanner.close();<a name="line.272"></a>
<span class="sourceLineNo">273</span> }<a name="line.273"></a>
<span class="sourceLineNo">274</span> }<a name="line.274"></a>
<span class="sourceLineNo">275</span> }<a name="line.275"></a>
<span class="sourceLineNo">276</span><a name="line.276"></a>
<span class="sourceLineNo">277</span> public static List&lt;InputSplit&gt; getSplits(Configuration conf) throws IOException {<a name="line.277"></a>
<span class="sourceLineNo">278</span> String snapshotName = getSnapshotName(conf);<a name="line.278"></a>
<span class="sourceLineNo">279</span><a name="line.279"></a>
<span class="sourceLineNo">280</span> Path rootDir = FSUtils.getRootDir(conf);<a name="line.280"></a>
<span class="sourceLineNo">281</span> FileSystem fs = rootDir.getFileSystem(conf);<a name="line.281"></a>
<span class="sourceLineNo">282</span><a name="line.282"></a>
<span class="sourceLineNo">283</span> SnapshotManifest manifest = getSnapshotManifest(conf, snapshotName, rootDir, fs);<a name="line.283"></a>
<span class="sourceLineNo">284</span><a name="line.284"></a>
<span class="sourceLineNo">285</span> List&lt;HRegionInfo&gt; regionInfos = getRegionInfosFromManifest(manifest);<a name="line.285"></a>
<span class="sourceLineNo">286</span><a name="line.286"></a>
<span class="sourceLineNo">287</span> // TODO: mapred does not support scan as input API. Work around for now.<a name="line.287"></a>
<span class="sourceLineNo">288</span> Scan scan = extractScanFromConf(conf);<a name="line.288"></a>
<span class="sourceLineNo">289</span> // the temp dir where the snapshot is restored<a name="line.289"></a>
<span class="sourceLineNo">290</span> Path restoreDir = new Path(conf.get(RESTORE_DIR_KEY));<a name="line.290"></a>
<span class="sourceLineNo">291</span><a name="line.291"></a>
<span class="sourceLineNo">292</span> RegionSplitter.SplitAlgorithm splitAlgo = getSplitAlgo(conf);<a name="line.292"></a>
<span class="sourceLineNo">293</span><a name="line.293"></a>
<span class="sourceLineNo">294</span> int numSplits = conf.getInt(NUM_SPLITS_PER_REGION, 1);<a name="line.294"></a>
<span class="sourceLineNo">295</span><a name="line.295"></a>
<span class="sourceLineNo">296</span> return getSplits(scan, manifest, regionInfos, restoreDir, conf, splitAlgo, numSplits);<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> public static RegionSplitter.SplitAlgorithm getSplitAlgo(Configuration conf) throws IOException{<a name="line.299"></a>
<span class="sourceLineNo">300</span> String splitAlgoClassName = conf.get(SPLIT_ALGO);<a name="line.300"></a>
<span class="sourceLineNo">301</span> if (splitAlgoClassName == null)<a name="line.301"></a>
<span class="sourceLineNo">302</span> return null;<a name="line.302"></a>
<span class="sourceLineNo">303</span> try {<a name="line.303"></a>
<span class="sourceLineNo">304</span> return Class.forName(splitAlgoClassName).asSubclass(RegionSplitter.SplitAlgorithm.class)<a name="line.304"></a>
<span class="sourceLineNo">305</span> .getDeclaredConstructor().newInstance();<a name="line.305"></a>
<span class="sourceLineNo">306</span> } catch (ClassNotFoundException | InstantiationException | IllegalAccessException |<a name="line.306"></a>
<span class="sourceLineNo">307</span> NoSuchMethodException | InvocationTargetException e) {<a name="line.307"></a>
<span class="sourceLineNo">308</span> throw new IOException("SplitAlgo class " + splitAlgoClassName + " is not found", e);<a name="line.308"></a>
<span class="sourceLineNo">309</span> }<a name="line.309"></a>
<span class="sourceLineNo">310</span> }<a name="line.310"></a>
<span class="sourceLineNo">311</span><a name="line.311"></a>
<span class="sourceLineNo">312</span><a name="line.312"></a>
<span class="sourceLineNo">313</span> public static List&lt;HRegionInfo&gt; getRegionInfosFromManifest(SnapshotManifest manifest) {<a name="line.313"></a>
<span class="sourceLineNo">314</span> List&lt;SnapshotRegionManifest&gt; regionManifests = manifest.getRegionManifests();<a name="line.314"></a>
<span class="sourceLineNo">315</span> if (regionManifests == null) {<a name="line.315"></a>
<span class="sourceLineNo">316</span> throw new IllegalArgumentException("Snapshot seems empty");<a name="line.316"></a>
<span class="sourceLineNo">317</span> }<a name="line.317"></a>
<span class="sourceLineNo">318</span><a name="line.318"></a>
<span class="sourceLineNo">319</span> List&lt;HRegionInfo&gt; regionInfos = Lists.newArrayListWithCapacity(regionManifests.size());<a name="line.319"></a>
<span class="sourceLineNo">320</span><a name="line.320"></a>
<span class="sourceLineNo">321</span> for (SnapshotRegionManifest regionManifest : regionManifests) {<a name="line.321"></a>
<span class="sourceLineNo">322</span> HRegionInfo hri = HRegionInfo.convert(regionManifest.getRegionInfo());<a name="line.322"></a>
<span class="sourceLineNo">323</span> if (hri.isOffline() &amp;&amp; (hri.isSplit() || hri.isSplitParent())) {<a name="line.323"></a>
<span class="sourceLineNo">324</span> continue;<a name="line.324"></a>
<span class="sourceLineNo">325</span> }<a name="line.325"></a>
<span class="sourceLineNo">326</span> regionInfos.add(hri);<a name="line.326"></a>
<span class="sourceLineNo">327</span> }<a name="line.327"></a>
<span class="sourceLineNo">328</span> return regionInfos;<a name="line.328"></a>
<span class="sourceLineNo">329</span> }<a name="line.329"></a>
<span class="sourceLineNo">330</span><a name="line.330"></a>
<span class="sourceLineNo">331</span> public static SnapshotManifest getSnapshotManifest(Configuration conf, String snapshotName,<a name="line.331"></a>
<span class="sourceLineNo">332</span> Path rootDir, FileSystem fs) throws IOException {<a name="line.332"></a>
<span class="sourceLineNo">333</span> Path snapshotDir = SnapshotDescriptionUtils.getCompletedSnapshotDir(snapshotName, rootDir);<a name="line.333"></a>
<span class="sourceLineNo">334</span> SnapshotDescription snapshotDesc = SnapshotDescriptionUtils.readSnapshotInfo(fs, snapshotDir);<a name="line.334"></a>
<span class="sourceLineNo">335</span> return SnapshotManifest.open(conf, fs, snapshotDir, snapshotDesc);<a name="line.335"></a>
<span class="sourceLineNo">336</span> }<a name="line.336"></a>
<span class="sourceLineNo">337</span><a name="line.337"></a>
<span class="sourceLineNo">338</span> public static Scan extractScanFromConf(Configuration conf) throws IOException {<a name="line.338"></a>
<span class="sourceLineNo">339</span> Scan scan = null;<a name="line.339"></a>
<span class="sourceLineNo">340</span> if (conf.get(TableInputFormat.SCAN) != null) {<a name="line.340"></a>
<span class="sourceLineNo">341</span> scan = TableMapReduceUtil.convertStringToScan(conf.get(TableInputFormat.SCAN));<a name="line.341"></a>
<span class="sourceLineNo">342</span> } else if (conf.get(org.apache.hadoop.hbase.mapred.TableInputFormat.COLUMN_LIST) != null) {<a name="line.342"></a>
<span class="sourceLineNo">343</span> String[] columns =<a name="line.343"></a>
<span class="sourceLineNo">344</span> conf.get(org.apache.hadoop.hbase.mapred.TableInputFormat.COLUMN_LIST).split(" ");<a name="line.344"></a>
<span class="sourceLineNo">345</span> scan = new Scan();<a name="line.345"></a>
<span class="sourceLineNo">346</span> for (String col : columns) {<a name="line.346"></a>
<span class="sourceLineNo">347</span> scan.addFamily(Bytes.toBytes(col));<a name="line.347"></a>
<span class="sourceLineNo">348</span> }<a name="line.348"></a>
<span class="sourceLineNo">349</span> } else {<a name="line.349"></a>
<span class="sourceLineNo">350</span> throw new IllegalArgumentException("Unable to create scan");<a name="line.350"></a>
<span class="sourceLineNo">351</span> }<a name="line.351"></a>
<span class="sourceLineNo">352</span> return scan;<a name="line.352"></a>
<span class="sourceLineNo">353</span> }<a name="line.353"></a>
<span class="sourceLineNo">354</span><a name="line.354"></a>
<span class="sourceLineNo">355</span> public static List&lt;InputSplit&gt; getSplits(Scan scan, SnapshotManifest manifest,<a name="line.355"></a>
<span class="sourceLineNo">356</span> List&lt;HRegionInfo&gt; regionManifests, Path restoreDir, Configuration conf) throws IOException {<a name="line.356"></a>
<span class="sourceLineNo">357</span> return getSplits(scan, manifest, regionManifests, restoreDir, conf, null, 1);<a name="line.357"></a>
<span class="sourceLineNo">358</span> }<a name="line.358"></a>
<span class="sourceLineNo">359</span><a name="line.359"></a>
<span class="sourceLineNo">360</span> public static List&lt;InputSplit&gt; getSplits(Scan scan, SnapshotManifest manifest,<a name="line.360"></a>
<span class="sourceLineNo">361</span> List&lt;HRegionInfo&gt; regionManifests, Path restoreDir,<a name="line.361"></a>
<span class="sourceLineNo">362</span> Configuration conf, RegionSplitter.SplitAlgorithm sa, int numSplits) throws IOException {<a name="line.362"></a>
<span class="sourceLineNo">363</span> // load table descriptor<a name="line.363"></a>
<span class="sourceLineNo">364</span> TableDescriptor htd = manifest.getTableDescriptor();<a name="line.364"></a>
<span class="sourceLineNo">365</span><a name="line.365"></a>
<span class="sourceLineNo">366</span> Path tableDir = FSUtils.getTableDir(restoreDir, htd.getTableName());<a name="line.366"></a>
<span class="sourceLineNo">367</span><a name="line.367"></a>
<span class="sourceLineNo">368</span> boolean localityEnabled = conf.getBoolean(SNAPSHOT_INPUTFORMAT_LOCALITY_ENABLED_KEY,<a name="line.368"></a>
<span class="sourceLineNo">369</span> SNAPSHOT_INPUTFORMAT_LOCALITY_ENABLED_DEFAULT);<a name="line.369"></a>
<span class="sourceLineNo">370</span><a name="line.370"></a>
<span class="sourceLineNo">371</span> List&lt;InputSplit&gt; splits = new ArrayList&lt;&gt;();<a name="line.371"></a>
<span class="sourceLineNo">372</span> for (HRegionInfo hri : regionManifests) {<a name="line.372"></a>
<span class="sourceLineNo">373</span> // load region descriptor<a name="line.373"></a>
<span class="sourceLineNo">374</span><a name="line.374"></a>
<span class="sourceLineNo">375</span> if (numSplits &gt; 1) {<a name="line.375"></a>
<span class="sourceLineNo">376</span> byte[][] sp = sa.split(hri.getStartKey(), hri.getEndKey(), numSplits, true);<a name="line.376"></a>
<span class="sourceLineNo">377</span> for (int i = 0; i &lt; sp.length - 1; i++) {<a name="line.377"></a>
<span class="sourceLineNo">378</span> if (PrivateCellUtil.overlappingKeys(scan.getStartRow(), scan.getStopRow(), sp[i],<a name="line.378"></a>
<span class="sourceLineNo">379</span> sp[i + 1])) {<a name="line.379"></a>
<span class="sourceLineNo">380</span> List&lt;String&gt; hosts =<a name="line.380"></a>
<span class="sourceLineNo">381</span> calculateLocationsForInputSplit(conf, htd, hri, tableDir, localityEnabled);<a name="line.381"></a>
<span class="sourceLineNo">382</span><a name="line.382"></a>
<span class="sourceLineNo">383</span> Scan boundedScan = new Scan(scan);<a name="line.383"></a>
<span class="sourceLineNo">384</span> if (scan.getStartRow().length == 0) {<a name="line.384"></a>
<span class="sourceLineNo">385</span> boundedScan.withStartRow(sp[i]);<a name="line.385"></a>
<span class="sourceLineNo">386</span> } else {<a name="line.386"></a>
<span class="sourceLineNo">387</span> boundedScan.withStartRow(<a name="line.387"></a>
<span class="sourceLineNo">388</span> Bytes.compareTo(scan.getStartRow(), sp[i]) &gt; 0 ? scan.getStartRow() : sp[i]);<a name="line.388"></a>
<span class="sourceLineNo">389</span> }<a name="line.389"></a>
<span class="sourceLineNo">390</span><a name="line.390"></a>
<span class="sourceLineNo">391</span> if (scan.getStopRow().length == 0) {<a name="line.391"></a>
<span class="sourceLineNo">392</span> boundedScan.withStopRow(sp[i + 1]);<a name="line.392"></a>
<span class="sourceLineNo">393</span> } else {<a name="line.393"></a>
<span class="sourceLineNo">394</span> boundedScan.withStopRow(<a name="line.394"></a>
<span class="sourceLineNo">395</span> Bytes.compareTo(scan.getStopRow(), sp[i + 1]) &lt; 0 ? scan.getStopRow() : sp[i + 1]);<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> splits.add(new InputSplit(htd, hri, hosts, boundedScan, restoreDir));<a name="line.398"></a>
<span class="sourceLineNo">399</span> }<a name="line.399"></a>
<span class="sourceLineNo">400</span> }<a name="line.400"></a>
<span class="sourceLineNo">401</span> } else {<a name="line.401"></a>
<span class="sourceLineNo">402</span> if (PrivateCellUtil.overlappingKeys(scan.getStartRow(), scan.getStopRow(),<a name="line.402"></a>
<span class="sourceLineNo">403</span> hri.getStartKey(), hri.getEndKey())) {<a name="line.403"></a>
<span class="sourceLineNo">404</span> List&lt;String&gt; hosts =<a name="line.404"></a>
<span class="sourceLineNo">405</span> calculateLocationsForInputSplit(conf, htd, hri, tableDir, localityEnabled);<a name="line.405"></a>
<span class="sourceLineNo">406</span> splits.add(new InputSplit(htd, hri, hosts, scan, restoreDir));<a name="line.406"></a>
<span class="sourceLineNo">407</span> }<a name="line.407"></a>
<span class="sourceLineNo">408</span> }<a name="line.408"></a>
<span class="sourceLineNo">409</span> }<a name="line.409"></a>
<span class="sourceLineNo">410</span><a name="line.410"></a>
<span class="sourceLineNo">411</span> return splits;<a name="line.411"></a>
<span class="sourceLineNo">412</span> }<a name="line.412"></a>
<span class="sourceLineNo">413</span><a name="line.413"></a>
<span class="sourceLineNo">414</span> /**<a name="line.414"></a>
<span class="sourceLineNo">415</span> * Compute block locations for snapshot files (which will get the locations for referred hfiles)<a name="line.415"></a>
<span class="sourceLineNo">416</span> * only when localityEnabled is true.<a name="line.416"></a>
<span class="sourceLineNo">417</span> */<a name="line.417"></a>
<span class="sourceLineNo">418</span> private static List&lt;String&gt; calculateLocationsForInputSplit(Configuration conf,<a name="line.418"></a>
<span class="sourceLineNo">419</span> TableDescriptor htd, HRegionInfo hri, Path tableDir, boolean localityEnabled)<a name="line.419"></a>
<span class="sourceLineNo">420</span> throws IOException {<a name="line.420"></a>
<span class="sourceLineNo">421</span> if (localityEnabled) { // care block locality<a name="line.421"></a>
<span class="sourceLineNo">422</span> return getBestLocations(conf,<a name="line.422"></a>
<span class="sourceLineNo">423</span> HRegion.computeHDFSBlocksDistribution(conf, htd, hri, tableDir));<a name="line.423"></a>
<span class="sourceLineNo">424</span> } else { // do not care block locality<a name="line.424"></a>
<span class="sourceLineNo">425</span> return null;<a name="line.425"></a>
<span class="sourceLineNo">426</span> }<a name="line.426"></a>
<span class="sourceLineNo">427</span> }<a name="line.427"></a>
<span class="sourceLineNo">428</span><a name="line.428"></a>
<span class="sourceLineNo">429</span> /**<a name="line.429"></a>
<span class="sourceLineNo">430</span> * This computes the locations to be passed from the InputSplit. MR/Yarn schedulers does not take<a name="line.430"></a>
<span class="sourceLineNo">431</span> * weights into account, thus will treat every location passed from the input split as equal. We<a name="line.431"></a>
<span class="sourceLineNo">432</span> * do not want to blindly pass all the locations, since we are creating one split per region, and<a name="line.432"></a>
<span class="sourceLineNo">433</span> * the region's blocks are all distributed throughout the cluster unless favorite node assignment<a name="line.433"></a>
<span class="sourceLineNo">434</span> * is used. On the expected stable case, only one location will contain most of the blocks as<a name="line.434"></a>
<span class="sourceLineNo">435</span> * local.<a name="line.435"></a>
<span class="sourceLineNo">436</span> * On the other hand, in favored node assignment, 3 nodes will contain highly local blocks. Here<a name="line.436"></a>
<span class="sourceLineNo">437</span> * we are doing a simple heuristic, where we will pass all hosts which have at least 80%<a name="line.437"></a>
<span class="sourceLineNo">438</span> * (hbase.tablesnapshotinputformat.locality.cutoff.multiplier) as much block locality as the top<a name="line.438"></a>
<span class="sourceLineNo">439</span> * host with the best locality.<a name="line.439"></a>
<span class="sourceLineNo">440</span> * Return at most numTopsAtMost locations if there are more than that.<a name="line.440"></a>
<span class="sourceLineNo">441</span> */<a name="line.441"></a>
<span class="sourceLineNo">442</span> private static List&lt;String&gt; getBestLocations(Configuration conf,<a name="line.442"></a>
<span class="sourceLineNo">443</span> HDFSBlocksDistribution blockDistribution, int numTopsAtMost) {<a name="line.443"></a>
<span class="sourceLineNo">444</span> HostAndWeight[] hostAndWeights = blockDistribution.getTopHostsWithWeights();<a name="line.444"></a>
<span class="sourceLineNo">445</span><a name="line.445"></a>
<span class="sourceLineNo">446</span> if (hostAndWeights.length == 0) { // no matter what numTopsAtMost is<a name="line.446"></a>
<span class="sourceLineNo">447</span> return null;<a name="line.447"></a>
<span class="sourceLineNo">448</span> }<a name="line.448"></a>
<span class="sourceLineNo">449</span><a name="line.449"></a>
<span class="sourceLineNo">450</span> if (numTopsAtMost &lt; 1) { // invalid if numTopsAtMost &lt; 1, correct it to be 1<a name="line.450"></a>
<span class="sourceLineNo">451</span> numTopsAtMost = 1;<a name="line.451"></a>
<span class="sourceLineNo">452</span> }<a name="line.452"></a>
<span class="sourceLineNo">453</span> int top = Math.min(numTopsAtMost, hostAndWeights.length);<a name="line.453"></a>
<span class="sourceLineNo">454</span> List&lt;String&gt; locations = new ArrayList&lt;&gt;(top);<a name="line.454"></a>
<span class="sourceLineNo">455</span> HostAndWeight topHost = hostAndWeights[0];<a name="line.455"></a>
<span class="sourceLineNo">456</span> locations.add(topHost.getHost());<a name="line.456"></a>
<span class="sourceLineNo">457</span><a name="line.457"></a>
<span class="sourceLineNo">458</span> if (top == 1) { // only care about the top host<a name="line.458"></a>
<span class="sourceLineNo">459</span> return locations;<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> // When top &gt;= 2,<a name="line.462"></a>
<span class="sourceLineNo">463</span> // do the heuristic: filter all hosts which have at least cutoffMultiplier % of block locality<a name="line.463"></a>
<span class="sourceLineNo">464</span> double cutoffMultiplier<a name="line.464"></a>
<span class="sourceLineNo">465</span> = conf.getFloat(LOCALITY_CUTOFF_MULTIPLIER, DEFAULT_LOCALITY_CUTOFF_MULTIPLIER);<a name="line.465"></a>
<span class="sourceLineNo">466</span><a name="line.466"></a>
<span class="sourceLineNo">467</span> double filterWeight = topHost.getWeight() * cutoffMultiplier;<a name="line.467"></a>
<span class="sourceLineNo">468</span><a name="line.468"></a>
<span class="sourceLineNo">469</span> for (int i = 1; i &lt;= top - 1; i++) {<a name="line.469"></a>
<span class="sourceLineNo">470</span> if (hostAndWeights[i].getWeight() &gt;= filterWeight) {<a name="line.470"></a>
<span class="sourceLineNo">471</span> locations.add(hostAndWeights[i].getHost());<a name="line.471"></a>
<span class="sourceLineNo">472</span> } else {<a name="line.472"></a>
<span class="sourceLineNo">473</span> // As hostAndWeights is in descending order,<a name="line.473"></a>
<span class="sourceLineNo">474</span> // we could break the loop as long as we meet a weight which is less than filterWeight.<a name="line.474"></a>
<span class="sourceLineNo">475</span> break;<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><a name="line.478"></a>
<span class="sourceLineNo">479</span> return locations;<a name="line.479"></a>
<span class="sourceLineNo">480</span> }<a name="line.480"></a>
<span class="sourceLineNo">481</span><a name="line.481"></a>
<span class="sourceLineNo">482</span> public static List&lt;String&gt; getBestLocations(Configuration conf,<a name="line.482"></a>
<span class="sourceLineNo">483</span> HDFSBlocksDistribution blockDistribution) {<a name="line.483"></a>
<span class="sourceLineNo">484</span> // 3 nodes will contain highly local blocks. So default to 3.<a name="line.484"></a>
<span class="sourceLineNo">485</span> return getBestLocations(conf, blockDistribution, 3);<a name="line.485"></a>
<span class="sourceLineNo">486</span> }<a name="line.486"></a>
<span class="sourceLineNo">487</span><a name="line.487"></a>
<span class="sourceLineNo">488</span> private static String getSnapshotName(Configuration conf) {<a name="line.488"></a>
<span class="sourceLineNo">489</span> String snapshotName = conf.get(SNAPSHOT_NAME_KEY);<a name="line.489"></a>
<span class="sourceLineNo">490</span> if (snapshotName == null) {<a name="line.490"></a>
<span class="sourceLineNo">491</span> throw new IllegalArgumentException("Snapshot name must be provided");<a name="line.491"></a>
<span class="sourceLineNo">492</span> }<a name="line.492"></a>
<span class="sourceLineNo">493</span> return snapshotName;<a name="line.493"></a>
<span class="sourceLineNo">494</span> }<a name="line.494"></a>
<span class="sourceLineNo">495</span><a name="line.495"></a>
<span class="sourceLineNo">496</span> /**<a name="line.496"></a>
<span class="sourceLineNo">497</span> * Configures the job to use TableSnapshotInputFormat to read from a snapshot.<a name="line.497"></a>
<span class="sourceLineNo">498</span> * @param conf the job to configuration<a name="line.498"></a>
<span class="sourceLineNo">499</span> * @param snapshotName the name of the snapshot to read from<a name="line.499"></a>
<span class="sourceLineNo">500</span> * @param restoreDir a temporary directory to restore the snapshot into. Current user should have<a name="line.500"></a>
<span class="sourceLineNo">501</span> * write permissions to this directory, and this should not be a subdirectory of rootdir.<a name="line.501"></a>
<span class="sourceLineNo">502</span> * After the job is finished, restoreDir can be deleted.<a name="line.502"></a>
<span class="sourceLineNo">503</span> * @throws IOException if an error occurs<a name="line.503"></a>
<span class="sourceLineNo">504</span> */<a name="line.504"></a>
<span class="sourceLineNo">505</span> public static void setInput(Configuration conf, String snapshotName, Path restoreDir)<a name="line.505"></a>
<span class="sourceLineNo">506</span> throws IOException {<a name="line.506"></a>
<span class="sourceLineNo">507</span> setInput(conf, snapshotName, restoreDir, null, 1);<a name="line.507"></a>
<span class="sourceLineNo">508</span> }<a name="line.508"></a>
<span class="sourceLineNo">509</span><a name="line.509"></a>
<span class="sourceLineNo">510</span> /**<a name="line.510"></a>
<span class="sourceLineNo">511</span> * Configures the job to use TableSnapshotInputFormat to read from a snapshot.<a name="line.511"></a>
<span class="sourceLineNo">512</span> * @param conf the job to configure<a name="line.512"></a>
<span class="sourceLineNo">513</span> * @param snapshotName the name of the snapshot to read from<a name="line.513"></a>
<span class="sourceLineNo">514</span> * @param restoreDir a temporary directory to restore the snapshot into. Current user should<a name="line.514"></a>
<span class="sourceLineNo">515</span> * have write permissions to this directory, and this should not be a subdirectory of rootdir.<a name="line.515"></a>
<span class="sourceLineNo">516</span> * After the job is finished, restoreDir can be deleted.<a name="line.516"></a>
<span class="sourceLineNo">517</span> * @param numSplitsPerRegion how many input splits to generate per one region<a name="line.517"></a>
<span class="sourceLineNo">518</span> * @param splitAlgo SplitAlgorithm to be used when generating InputSplits<a name="line.518"></a>
<span class="sourceLineNo">519</span> * @throws IOException if an error occurs<a name="line.519"></a>
<span class="sourceLineNo">520</span> */<a name="line.520"></a>
<span class="sourceLineNo">521</span> public static void setInput(Configuration conf, String snapshotName, Path restoreDir,<a name="line.521"></a>
<span class="sourceLineNo">522</span> RegionSplitter.SplitAlgorithm splitAlgo, int numSplitsPerRegion)<a name="line.522"></a>
<span class="sourceLineNo">523</span> throws IOException {<a name="line.523"></a>
<span class="sourceLineNo">524</span> conf.set(SNAPSHOT_NAME_KEY, snapshotName);<a name="line.524"></a>
<span class="sourceLineNo">525</span> if (numSplitsPerRegion &lt; 1) {<a name="line.525"></a>
<span class="sourceLineNo">526</span> throw new IllegalArgumentException("numSplits must be &gt;= 1, " +<a name="line.526"></a>
<span class="sourceLineNo">527</span> "illegal numSplits : " + numSplitsPerRegion);<a name="line.527"></a>
<span class="sourceLineNo">528</span> }<a name="line.528"></a>
<span class="sourceLineNo">529</span> if (splitAlgo == null &amp;&amp; numSplitsPerRegion &gt; 1) {<a name="line.529"></a>
<span class="sourceLineNo">530</span> throw new IllegalArgumentException("Split algo can't be null when numSplits &gt; 1");<a name="line.530"></a>
<span class="sourceLineNo">531</span> }<a name="line.531"></a>
<span class="sourceLineNo">532</span> if (splitAlgo != null) {<a name="line.532"></a>
<span class="sourceLineNo">533</span> conf.set(SPLIT_ALGO, splitAlgo.getClass().getName());<a name="line.533"></a>
<span class="sourceLineNo">534</span> }<a name="line.534"></a>
<span class="sourceLineNo">535</span> conf.setInt(NUM_SPLITS_PER_REGION, numSplitsPerRegion);<a name="line.535"></a>
<span class="sourceLineNo">536</span> Path rootDir = FSUtils.getRootDir(conf);<a name="line.536"></a>
<span class="sourceLineNo">537</span> FileSystem fs = rootDir.getFileSystem(conf);<a name="line.537"></a>
<span class="sourceLineNo">538</span><a name="line.538"></a>
<span class="sourceLineNo">539</span> restoreDir = new Path(restoreDir, UUID.randomUUID().toString());<a name="line.539"></a>
<span class="sourceLineNo">540</span><a name="line.540"></a>
<span class="sourceLineNo">541</span> RestoreSnapshotHelper.copySnapshotForScanner(conf, fs, rootDir, restoreDir, snapshotName);<a name="line.541"></a>
<span class="sourceLineNo">542</span> conf.set(RESTORE_DIR_KEY, restoreDir.toString());<a name="line.542"></a>
<span class="sourceLineNo">543</span> }<a name="line.543"></a>
<span class="sourceLineNo">544</span>}<a name="line.544"></a>
</pre>
</div>
</body>
</html>