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.wal;
019
020import static org.apache.hadoop.hbase.TableName.META_TABLE_NAME;
021
022import java.io.IOException;
023import java.io.InterruptedIOException;
024import java.util.HashMap;
025import java.util.List;
026import java.util.Map;
027import java.util.concurrent.ConcurrentHashMap;
028import java.util.concurrent.ConcurrentMap;
029import java.util.concurrent.ExecutionException;
030import java.util.concurrent.Future;
031import java.util.concurrent.atomic.AtomicInteger;
032import org.apache.hadoop.fs.Path;
033import org.apache.hadoop.hbase.Cell;
034import org.apache.hadoop.hbase.CellComparatorImpl;
035import org.apache.hadoop.hbase.CellUtil;
036import org.apache.hadoop.hbase.MetaCellComparator;
037import org.apache.hadoop.hbase.PrivateCellUtil;
038import org.apache.hadoop.hbase.TableName;
039import org.apache.hadoop.hbase.io.hfile.CacheConfig;
040import org.apache.hadoop.hbase.io.hfile.HFileContext;
041import org.apache.hadoop.hbase.io.hfile.HFileContextBuilder;
042import org.apache.hadoop.hbase.regionserver.CellSet;
043import org.apache.hadoop.hbase.regionserver.StoreFileWriter;
044import org.apache.hadoop.hbase.regionserver.StoreUtils;
045import org.apache.hadoop.hbase.util.Bytes;
046import org.apache.hadoop.hbase.wal.EntryBuffers.RegionEntryBuffer;
047import org.apache.hadoop.hbase.wal.WAL.Entry;
048import org.apache.yetus.audience.InterfaceAudience;
049import org.slf4j.Logger;
050import org.slf4j.LoggerFactory;
051
052/**
053 * A WALSplitter sink that outputs {@link org.apache.hadoop.hbase.io.hfile.HFile}s. Runs with a
054 * bounded number of HFile writers at any one time rather than let the count run up.
055 * @see BoundedRecoveredEditsOutputSink for a sink implementation that writes intermediate
056 *      recovered.edits files.
057 */
058@InterfaceAudience.Private
059public class BoundedRecoveredHFilesOutputSink extends OutputSink {
060  private static final Logger LOG = LoggerFactory.getLogger(BoundedRecoveredHFilesOutputSink.class);
061
062  private final WALSplitter walSplitter;
063
064  // Since the splitting process may create multiple output files, we need a map
065  // to track the output count of each region.
066  private ConcurrentMap<String, Long> regionEditsWrittenMap = new ConcurrentHashMap<>();
067  // Need a counter to track the opening writers.
068  private final AtomicInteger openingWritersNum = new AtomicInteger(0);
069
070  public BoundedRecoveredHFilesOutputSink(WALSplitter walSplitter,
071    WALSplitter.PipelineController controller, EntryBuffers entryBuffers, int numWriters) {
072    super(controller, entryBuffers, numWriters);
073    this.walSplitter = walSplitter;
074  }
075
076  @Override
077  public void append(RegionEntryBuffer buffer) throws IOException {
078    Map<String, CellSet> familyCells = new HashMap<>();
079    Map<String, Long> familySeqIds = new HashMap<>();
080    boolean isMetaTable = buffer.tableName.equals(META_TABLE_NAME);
081    // First iterate all Cells to find which column families are present and to stamp Cell with
082    // sequence id.
083    for (WAL.Entry entry : buffer.entries) {
084      long seqId = entry.getKey().getSequenceId();
085      List<Cell> cells = entry.getEdit().getCells();
086      for (Cell cell : cells) {
087        if (CellUtil.matchingFamily(cell, WALEdit.METAFAMILY)) {
088          continue;
089        }
090        PrivateCellUtil.setSequenceId(cell, seqId);
091        String familyName = Bytes.toString(CellUtil.cloneFamily(cell));
092        // comparator need to be specified for meta
093        familyCells
094          .computeIfAbsent(familyName,
095            key -> new CellSet(
096              isMetaTable ? MetaCellComparator.META_COMPARATOR : CellComparatorImpl.COMPARATOR))
097          .add(cell);
098        familySeqIds.compute(familyName, (k, v) -> v == null ? seqId : Math.max(v, seqId));
099      }
100    }
101
102    // Create a new hfile writer for each column family, write edits then close writer.
103    String regionName = Bytes.toString(buffer.encodedRegionName);
104    for (Map.Entry<String, CellSet> cellsEntry : familyCells.entrySet()) {
105      String familyName = cellsEntry.getKey();
106      StoreFileWriter writer = createRecoveredHFileWriter(buffer.tableName, regionName,
107        familySeqIds.get(familyName), familyName, isMetaTable);
108      LOG.trace("Created {}", writer.getPath());
109      openingWritersNum.incrementAndGet();
110      try {
111        for (Cell cell : cellsEntry.getValue()) {
112          writer.append(cell);
113        }
114        // Append the max seqid to hfile, used when recovery.
115        writer.appendMetadata(familySeqIds.get(familyName), false);
116        regionEditsWrittenMap.compute(Bytes.toString(buffer.encodedRegionName),
117          (k, v) -> v == null ? buffer.entries.size() : v + buffer.entries.size());
118        splits.add(writer.getPath());
119        openingWritersNum.decrementAndGet();
120      } finally {
121        writer.close();
122        LOG.trace("Closed {}, edits={}", writer.getPath(), familyCells.size());
123      }
124    }
125  }
126
127  @Override
128  public List<Path> close() throws IOException {
129    boolean isSuccessful = true;
130    try {
131      isSuccessful = finishWriterThreads(false);
132    } finally {
133      isSuccessful &= writeRemainingEntryBuffers();
134    }
135    return isSuccessful ? splits : null;
136  }
137
138  /**
139   * Write out the remaining RegionEntryBuffers and close the writers.
140   * @return true when there is no error.
141   */
142  private boolean writeRemainingEntryBuffers() throws IOException {
143    for (EntryBuffers.RegionEntryBuffer buffer : entryBuffers.buffers.values()) {
144      closeCompletionService.submit(() -> {
145        append(buffer);
146        return null;
147      });
148    }
149    boolean progressFailed = false;
150    try {
151      for (int i = 0, n = entryBuffers.buffers.size(); i < n; i++) {
152        Future<Void> future = closeCompletionService.take();
153        future.get();
154        if (!progressFailed && reporter != null && !reporter.progress()) {
155          progressFailed = true;
156        }
157      }
158    } catch (InterruptedException e) {
159      IOException iie = new InterruptedIOException();
160      iie.initCause(e);
161      throw iie;
162    } catch (ExecutionException e) {
163      throw new IOException(e.getCause());
164    } finally {
165      closeThreadPool.shutdownNow();
166    }
167    return !progressFailed;
168  }
169
170  @Override
171  public Map<String, Long> getOutputCounts() {
172    return regionEditsWrittenMap;
173  }
174
175  @Override
176  public int getNumberOfRecoveredRegions() {
177    return regionEditsWrittenMap.size();
178  }
179
180  @Override
181  public int getNumOpenWriters() {
182    return openingWritersNum.get();
183  }
184
185  @Override
186  public boolean keepRegionEvent(Entry entry) {
187    return false;
188  }
189
190  /**
191   * @return Returns a base HFile without compressions or encodings; good enough for recovery given
192   *         hfile has metadata on how it was written.
193   */
194  private StoreFileWriter createRecoveredHFileWriter(TableName tableName, String regionName,
195    long seqId, String familyName, boolean isMetaTable) throws IOException {
196    Path outputDir = WALSplitUtil.tryCreateRecoveredHFilesDir(walSplitter.rootFS, walSplitter.conf,
197      tableName, regionName, familyName);
198    StoreFileWriter.Builder writerBuilder =
199      new StoreFileWriter.Builder(walSplitter.conf, CacheConfig.DISABLED, walSplitter.rootFS)
200        .withOutputDir(outputDir);
201    HFileContext hFileContext =
202      new HFileContextBuilder().withChecksumType(StoreUtils.getChecksumType(walSplitter.conf))
203        .withBytesPerCheckSum(StoreUtils.getBytesPerChecksum(walSplitter.conf)).withCellComparator(
204          isMetaTable ? MetaCellComparator.META_COMPARATOR : CellComparatorImpl.COMPARATOR)
205        .build();
206    return writerBuilder.withFileContext(hFileContext).build();
207  }
208}