001/*
002 * Licensed to the Apache Software Foundation (ASF) under one
003 * or more contributor license agreements.  See the NOTICE file
004 * distributed with this work for additional information
005 * regarding copyright ownership.  The ASF licenses this file
006 * to you under the Apache License, Version 2.0 (the
007 * "License"); you may not use this file except in compliance
008 * with the License.  You may obtain a copy of the License at
009 *
010 *     http://www.apache.org/licenses/LICENSE-2.0
011 *
012 * Unless required by applicable law or agreed to in writing, software
013 * distributed under the License is distributed on an "AS IS" BASIS,
014 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
015 * See the License for the specific language governing permissions and
016 * limitations under the License.
017 */
018package org.apache.hadoop.hbase.mapreduce;
019
020import java.io.IOException;
021import org.apache.hadoop.conf.Configurable;
022import org.apache.hadoop.conf.Configuration;
023import org.apache.hadoop.hbase.HBaseConfiguration;
024import org.apache.hadoop.hbase.HConstants;
025import org.apache.hadoop.hbase.TableName;
026import org.apache.hadoop.hbase.TableNotEnabledException;
027import org.apache.hadoop.hbase.TableNotFoundException;
028import org.apache.hadoop.hbase.client.Admin;
029import org.apache.hadoop.hbase.client.BufferedMutator;
030import org.apache.hadoop.hbase.client.Connection;
031import org.apache.hadoop.hbase.client.ConnectionFactory;
032import org.apache.hadoop.hbase.client.Delete;
033import org.apache.hadoop.hbase.client.Mutation;
034import org.apache.hadoop.hbase.client.Put;
035import org.apache.hadoop.mapreduce.JobContext;
036import org.apache.hadoop.mapreduce.OutputCommitter;
037import org.apache.hadoop.mapreduce.OutputFormat;
038import org.apache.hadoop.mapreduce.RecordWriter;
039import org.apache.hadoop.mapreduce.TaskAttemptContext;
040import org.apache.yetus.audience.InterfaceAudience;
041import org.slf4j.Logger;
042import org.slf4j.LoggerFactory;
043
044/**
045 * Convert Map/Reduce output and write it to an HBase table. The KEY is ignored while the output
046 * value <u>must</u> be either a {@link Put} or a {@link Delete} instance.
047 */
048@InterfaceAudience.Public
049public class TableOutputFormat<KEY> extends OutputFormat<KEY, Mutation> implements Configurable {
050
051  private static final Logger LOG = LoggerFactory.getLogger(TableOutputFormat.class);
052
053  /** Job parameter that specifies the output table. */
054  public static final String OUTPUT_TABLE = "hbase.mapred.outputtable";
055
056  /**
057   * Prefix for configuration property overrides to apply in {@link #setConf(Configuration)}. For
058   * keys matching this prefix, the prefix is stripped, and the value is set in the configuration
059   * with the resulting key, ie. the entry "hbase.mapred.output.key1 = value1" would be set in the
060   * configuration as "key1 = value1". Use this to set properties which should only be applied to
061   * the {@code TableOutputFormat} configuration and not the input configuration.
062   */
063  public static final String OUTPUT_CONF_PREFIX = "hbase.mapred.output.";
064
065  /**
066   * Optional job parameter to specify a peer cluster. Used specifying remote cluster when copying
067   * between hbase clusters (the source is picked up from <code>hbase-site.xml</code>).
068   * @see TableMapReduceUtil#initTableReducerJob(String, Class, org.apache.hadoop.mapreduce.Job,
069   *      Class, String)
070   */
071  public static final String QUORUM_ADDRESS = OUTPUT_CONF_PREFIX + "quorum";
072
073  /** Optional job parameter to specify peer cluster's ZK client port */
074  public static final String QUORUM_PORT = OUTPUT_CONF_PREFIX + "quorum.port";
075
076  /**
077   * Optional specification of the rs class name of the peer cluster.
078   * @deprecated Since 2.5.9, 2.6.1 and 2.7.0, will be removed in 4.0.0. Does not take effect from
079   *             long ago, see HBASE-6044.
080   */
081  @Deprecated
082  public static final String REGION_SERVER_CLASS = OUTPUT_CONF_PREFIX + "rs.class";
083  /**
084   * Optional specification of the rs impl name of the peer cluster
085   * @deprecated Since 2.5.9, 2.6.1 and 2.7.0, will be removed in 4.0.0. Does not take effect from
086   *             long ago, see HBASE-6044.
087   */
088  @Deprecated
089  public static final String REGION_SERVER_IMPL = OUTPUT_CONF_PREFIX + "rs.impl";
090
091  /** The configuration. */
092  private Configuration conf = null;
093
094  /**
095   * Writes the reducer output to an HBase table.
096   */
097  protected class TableRecordWriter extends RecordWriter<KEY, Mutation> {
098
099    private Connection connection;
100    private BufferedMutator mutator;
101
102    /**
103     *
104    *
105     */
106    public TableRecordWriter() throws IOException {
107      String tableName = conf.get(OUTPUT_TABLE);
108      this.connection = ConnectionFactory.createConnection(conf);
109      this.mutator = connection.getBufferedMutator(TableName.valueOf(tableName));
110      LOG.info("Created table instance for " + tableName);
111    }
112
113    /**
114     * Closes the writer, in this case flush table commits.
115     * @param context The context.
116     * @throws IOException When closing the writer fails.
117     * @see RecordWriter#close(TaskAttemptContext)
118     */
119    @Override
120    public void close(TaskAttemptContext context) throws IOException {
121      try {
122        if (mutator != null) {
123          mutator.close();
124        }
125      } finally {
126        if (connection != null) {
127          connection.close();
128        }
129      }
130    }
131
132    /**
133     * Writes a key/value pair into the table.
134     * @param key   The key.
135     * @param value The value.
136     * @throws IOException When writing fails.
137     * @see RecordWriter#write(Object, Object)
138     */
139    @Override
140    public void write(KEY key, Mutation value) throws IOException {
141      if (!(value instanceof Put) && !(value instanceof Delete)) {
142        throw new IOException("Pass a Delete or a Put");
143      }
144      mutator.mutate(value);
145    }
146  }
147
148  /**
149   * Creates a new record writer. Be aware that the baseline javadoc gives the impression that there
150   * is a single {@link RecordWriter} per job but in HBase, it is more natural if we give you a new
151   * RecordWriter per call of this method. You must close the returned RecordWriter when done.
152   * Failure to do so will drop writes.
153   * @param context The current task context.
154   * @return The newly created writer instance.
155   * @throws IOException          When creating the writer fails.
156   * @throws InterruptedException When the job is cancelled.
157   */
158  @Override
159  public RecordWriter<KEY, Mutation> getRecordWriter(TaskAttemptContext context)
160    throws IOException, InterruptedException {
161    return new TableRecordWriter();
162  }
163
164  /**
165   * Checks if the output table exists and is enabled.
166   * @param context The current context.
167   * @throws IOException          When the check fails.
168   * @throws InterruptedException When the job is aborted.
169   * @see OutputFormat#checkOutputSpecs(JobContext)
170   */
171  @Override
172  public void checkOutputSpecs(JobContext context) throws IOException, InterruptedException {
173    Configuration hConf = getConf();
174    if (hConf == null) {
175      hConf = context.getConfiguration();
176    }
177
178    try (Connection connection = ConnectionFactory.createConnection(hConf);
179      Admin admin = connection.getAdmin()) {
180      TableName tableName = TableName.valueOf(hConf.get(OUTPUT_TABLE));
181      if (!admin.tableExists(tableName)) {
182        throw new TableNotFoundException(
183          "Can't write, table does not exist:" + tableName.getNameAsString());
184      }
185
186      if (!admin.isTableEnabled(tableName)) {
187        throw new TableNotEnabledException(
188          "Can't write, table is not enabled: " + tableName.getNameAsString());
189      }
190    }
191  }
192
193  /**
194   * Returns the output committer.
195   * @param context The current context.
196   * @return The committer.
197   * @throws IOException          When creating the committer fails.
198   * @throws InterruptedException When the job is aborted.
199   * @see OutputFormat#getOutputCommitter(TaskAttemptContext)
200   */
201  @Override
202  public OutputCommitter getOutputCommitter(TaskAttemptContext context)
203    throws IOException, InterruptedException {
204    return new TableOutputCommitter();
205  }
206
207  @Override
208  public Configuration getConf() {
209    return conf;
210  }
211
212  @Override
213  public void setConf(Configuration otherConf) {
214    String tableName = otherConf.get(OUTPUT_TABLE);
215    if (tableName == null || tableName.length() <= 0) {
216      throw new IllegalArgumentException("Must specify table name");
217    }
218
219    String address = otherConf.get(QUORUM_ADDRESS);
220    int zkClientPort = otherConf.getInt(QUORUM_PORT, 0);
221
222    try {
223      this.conf = HBaseConfiguration.createClusterConf(otherConf, address, OUTPUT_CONF_PREFIX);
224      if (zkClientPort != 0) {
225        this.conf.setInt(HConstants.ZOOKEEPER_CLIENT_PORT, zkClientPort);
226      }
227    } catch (IOException e) {
228      LOG.error(e.toString(), e);
229      throw new RuntimeException(e);
230    }
231  }
232}