| <!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.mapred;<a name="line.18"></a> |
| <span class="sourceLineNo">019</span><a name="line.19"></a> |
| <span class="sourceLineNo">020</span>import edu.umd.cs.findbugs.annotations.SuppressWarnings;<a name="line.20"></a> |
| <span class="sourceLineNo">021</span>import java.io.IOException;<a name="line.21"></a> |
| <span class="sourceLineNo">022</span>import java.util.Collection;<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;<a name="line.24"></a> |
| <span class="sourceLineNo">025</span>import org.apache.hadoop.conf.Configuration;<a name="line.25"></a> |
| <span class="sourceLineNo">026</span>import org.apache.hadoop.fs.Path;<a name="line.26"></a> |
| <span class="sourceLineNo">027</span>import org.apache.hadoop.hbase.client.Result;<a name="line.27"></a> |
| <span class="sourceLineNo">028</span>import org.apache.hadoop.hbase.client.Scan;<a name="line.28"></a> |
| <span class="sourceLineNo">029</span>import org.apache.hadoop.hbase.io.ImmutableBytesWritable;<a name="line.29"></a> |
| <span class="sourceLineNo">030</span>import org.apache.hadoop.hbase.mapreduce.MultiTableSnapshotInputFormatImpl;<a name="line.30"></a> |
| <span class="sourceLineNo">031</span>import org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormatImpl;<a name="line.31"></a> |
| <span class="sourceLineNo">032</span>import org.apache.hadoop.mapred.InputFormat;<a name="line.32"></a> |
| <span class="sourceLineNo">033</span>import org.apache.hadoop.mapred.InputSplit;<a name="line.33"></a> |
| <span class="sourceLineNo">034</span>import org.apache.hadoop.mapred.JobConf;<a name="line.34"></a> |
| <span class="sourceLineNo">035</span>import org.apache.hadoop.mapred.RecordReader;<a name="line.35"></a> |
| <span class="sourceLineNo">036</span>import org.apache.hadoop.mapred.Reporter;<a name="line.36"></a> |
| <span class="sourceLineNo">037</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.37"></a> |
| <span class="sourceLineNo">038</span><a name="line.38"></a> |
| <span class="sourceLineNo">039</span>/**<a name="line.39"></a> |
| <span class="sourceLineNo">040</span> * MultiTableSnapshotInputFormat generalizes<a name="line.40"></a> |
| <span class="sourceLineNo">041</span> * {@link org.apache.hadoop.hbase.mapred.TableSnapshotInputFormat} allowing a MapReduce job to run<a name="line.41"></a> |
| <span class="sourceLineNo">042</span> * over one or more table snapshots, with one or more scans configured for each. Internally, the<a name="line.42"></a> |
| <span class="sourceLineNo">043</span> * input format delegates to {@link org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormat} and<a name="line.43"></a> |
| <span class="sourceLineNo">044</span> * thus has the same performance advantages; see<a name="line.44"></a> |
| <span class="sourceLineNo">045</span> * {@link org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormat} for more details. Usage is<a name="line.45"></a> |
| <span class="sourceLineNo">046</span> * similar to TableSnapshotInputFormat, with the following exception:<a name="line.46"></a> |
| <span class="sourceLineNo">047</span> * initMultiTableSnapshotMapperJob takes in a map from snapshot name to a collection of scans. For<a name="line.47"></a> |
| <span class="sourceLineNo">048</span> * each snapshot in the map, each corresponding scan will be applied; the overall dataset for the<a name="line.48"></a> |
| <span class="sourceLineNo">049</span> * job is defined by the concatenation of the regions and tables included in each snapshot/scan<a name="line.49"></a> |
| <span class="sourceLineNo">050</span> * pair.<a name="line.50"></a> |
| <span class="sourceLineNo">051</span> * {@link TableMapReduceUtil#initMultiTableSnapshotMapperJob(Map, Class, Class, Class, JobConf, boolean, Path)}<a name="line.51"></a> |
| <span class="sourceLineNo">052</span> * can be used to configure the job.<a name="line.52"></a> |
| <span class="sourceLineNo">053</span> *<a name="line.53"></a> |
| <span class="sourceLineNo">054</span> * <pre><a name="line.54"></a> |
| <span class="sourceLineNo">055</span> * {@code<a name="line.55"></a> |
| <span class="sourceLineNo">056</span> * Job job = new Job(conf);<a name="line.56"></a> |
| <span class="sourceLineNo">057</span> * Map<String, Collection<Scan>> snapshotScans = ImmutableMap.of(<a name="line.57"></a> |
| <span class="sourceLineNo">058</span> * "snapshot1", ImmutableList.of(new Scan(Bytes.toBytes("a"), Bytes.toBytes("b"))),<a name="line.58"></a> |
| <span class="sourceLineNo">059</span> * "snapshot2", ImmutableList.of(new Scan(Bytes.toBytes("1"), Bytes.toBytes("2")))<a name="line.59"></a> |
| <span class="sourceLineNo">060</span> * );<a name="line.60"></a> |
| <span class="sourceLineNo">061</span> * Path restoreDir = new Path("/tmp/snapshot_restore_dir")<a name="line.61"></a> |
| <span class="sourceLineNo">062</span> * TableMapReduceUtil.initTableSnapshotMapperJob(<a name="line.62"></a> |
| <span class="sourceLineNo">063</span> * snapshotScans, MyTableMapper.class, MyMapKeyOutput.class,<a name="line.63"></a> |
| <span class="sourceLineNo">064</span> * MyMapOutputValueWritable.class, job, true, restoreDir);<a name="line.64"></a> |
| <span class="sourceLineNo">065</span> * }<a name="line.65"></a> |
| <span class="sourceLineNo">066</span> * </pre><a name="line.66"></a> |
| <span class="sourceLineNo">067</span> *<a name="line.67"></a> |
| <span class="sourceLineNo">068</span> * Internally, this input format restores each snapshot into a subdirectory of the given tmp<a name="line.68"></a> |
| <span class="sourceLineNo">069</span> * directory. Input splits and record readers are created as described in<a name="line.69"></a> |
| <span class="sourceLineNo">070</span> * {@link org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormat} (one per region). See<a name="line.70"></a> |
| <span class="sourceLineNo">071</span> * {@link org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormat} for more notes on<a name="line.71"></a> |
| <span class="sourceLineNo">072</span> * permissioning; the same caveats apply here.<a name="line.72"></a> |
| <span class="sourceLineNo">073</span> * @see org.apache.hadoop.hbase.mapreduce.TableSnapshotInputFormat<a name="line.73"></a> |
| <span class="sourceLineNo">074</span> * @see org.apache.hadoop.hbase.client.TableSnapshotScanner<a name="line.74"></a> |
| <span class="sourceLineNo">075</span> */<a name="line.75"></a> |
| <span class="sourceLineNo">076</span>@InterfaceAudience.Public<a name="line.76"></a> |
| <span class="sourceLineNo">077</span>public class MultiTableSnapshotInputFormat extends TableSnapshotInputFormat<a name="line.77"></a> |
| <span class="sourceLineNo">078</span> implements InputFormat<ImmutableBytesWritable, Result> {<a name="line.78"></a> |
| <span class="sourceLineNo">079</span><a name="line.79"></a> |
| <span class="sourceLineNo">080</span> private final MultiTableSnapshotInputFormatImpl delegate;<a name="line.80"></a> |
| <span class="sourceLineNo">081</span><a name="line.81"></a> |
| <span class="sourceLineNo">082</span> public MultiTableSnapshotInputFormat() {<a name="line.82"></a> |
| <span class="sourceLineNo">083</span> this.delegate = new MultiTableSnapshotInputFormatImpl();<a name="line.83"></a> |
| <span class="sourceLineNo">084</span> }<a name="line.84"></a> |
| <span class="sourceLineNo">085</span><a name="line.85"></a> |
| <span class="sourceLineNo">086</span> @Override<a name="line.86"></a> |
| <span class="sourceLineNo">087</span> public InputSplit[] getSplits(JobConf job, int numSplits) throws IOException {<a name="line.87"></a> |
| <span class="sourceLineNo">088</span> List<TableSnapshotInputFormatImpl.InputSplit> splits = delegate.getSplits(job);<a name="line.88"></a> |
| <span class="sourceLineNo">089</span> InputSplit[] results = new InputSplit[splits.size()];<a name="line.89"></a> |
| <span class="sourceLineNo">090</span> for (int i = 0; i < splits.size(); i++) {<a name="line.90"></a> |
| <span class="sourceLineNo">091</span> results[i] = new TableSnapshotRegionSplit(splits.get(i));<a name="line.91"></a> |
| <span class="sourceLineNo">092</span> }<a name="line.92"></a> |
| <span class="sourceLineNo">093</span> return results;<a name="line.93"></a> |
| <span class="sourceLineNo">094</span> }<a name="line.94"></a> |
| <span class="sourceLineNo">095</span><a name="line.95"></a> |
| <span class="sourceLineNo">096</span> @Override<a name="line.96"></a> |
| <span class="sourceLineNo">097</span> public RecordReader<ImmutableBytesWritable, Result> getRecordReader(InputSplit split, JobConf job,<a name="line.97"></a> |
| <span class="sourceLineNo">098</span> Reporter reporter) throws IOException {<a name="line.98"></a> |
| <span class="sourceLineNo">099</span> return new TableSnapshotRecordReader((TableSnapshotRegionSplit) split, job);<a name="line.99"></a> |
| <span class="sourceLineNo">100</span> }<a name="line.100"></a> |
| <span class="sourceLineNo">101</span><a name="line.101"></a> |
| <span class="sourceLineNo">102</span> @SuppressWarnings("checkstyle:linelength")<a name="line.102"></a> |
| <span class="sourceLineNo">103</span> /**<a name="line.103"></a> |
| <span class="sourceLineNo">104</span> * Configure conf to read from snapshotScans, with snapshots restored to a subdirectory of<a name="line.104"></a> |
| <span class="sourceLineNo">105</span> * restoreDir. Sets:<a name="line.105"></a> |
| <span class="sourceLineNo">106</span> * {@link org.apache.hadoop.hbase.mapreduce.MultiTableSnapshotInputFormatImpl#RESTORE_DIRS_KEY},<a name="line.106"></a> |
| <span class="sourceLineNo">107</span> * {@link org.apache.hadoop.hbase.mapreduce.MultiTableSnapshotInputFormatImpl#SNAPSHOT_TO_SCANS_KEY}<a name="line.107"></a> |
| <span class="sourceLineNo">108</span> */<a name="line.108"></a> |
| <span class="sourceLineNo">109</span> public static void setInput(Configuration conf, Map<String, Collection<Scan>> snapshotScans,<a name="line.109"></a> |
| <span class="sourceLineNo">110</span> Path restoreDir) throws IOException {<a name="line.110"></a> |
| <span class="sourceLineNo">111</span> new MultiTableSnapshotInputFormatImpl().setInput(conf, snapshotScans, restoreDir);<a name="line.111"></a> |
| <span class="sourceLineNo">112</span> }<a name="line.112"></a> |
| <span class="sourceLineNo">113</span><a name="line.113"></a> |
| <span class="sourceLineNo">114</span>}<a name="line.114"></a> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| </pre> |
| </div> |
| </body> |
| </html> |
