Source code

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.regionserver;
019
020import static org.junit.Assert.assertEquals;
021import static org.junit.Assert.assertFalse;
022import static org.junit.Assert.assertTrue;
023
024import java.io.IOException;
025import java.util.ArrayList;
026import java.util.Collection;
027import java.util.Collections;
028import java.util.HashMap;
029import java.util.HashSet;
030import java.util.List;
031import java.util.Map;
032import java.util.Random;
033import java.util.Set;
034import java.util.TreeSet;
035import java.util.concurrent.ThreadLocalRandom;
036import org.apache.hadoop.hbase.Cell;
037import org.apache.hadoop.hbase.CellComparatorImpl;
038import org.apache.hadoop.hbase.CellUtil;
039import org.apache.hadoop.hbase.ExtendedCell;
040import org.apache.hadoop.hbase.HBaseTestingUtil;
041import org.apache.hadoop.hbase.KeyValue;
042import org.apache.hadoop.hbase.KeyValueTestUtil;
043import org.apache.hadoop.hbase.PrivateCellUtil;
044import org.apache.hadoop.hbase.client.ColumnFamilyDescriptorBuilder;
045import org.apache.hadoop.hbase.client.Delete;
046import org.apache.hadoop.hbase.client.Put;
047import org.apache.hadoop.hbase.client.Scan;
048import org.apache.hadoop.hbase.io.compress.Compression;
049import org.apache.hadoop.hbase.io.encoding.DataBlockEncoding;
050import org.apache.hadoop.hbase.io.hfile.BlockCacheFactory;
051import org.apache.hadoop.hbase.util.BloomFilterUtil;
052import org.apache.hadoop.hbase.util.Bytes;
053import org.junit.Test;
054import org.junit.runners.Parameterized.Parameter;
055import org.slf4j.Logger;
056import org.slf4j.LoggerFactory;
057
058/**
059 * Tests optimized scanning of multiple columns. <br>
060 * We separated the big test into several sub-class UT, because When in ROWCOL bloom type, we will
061 * test the row-col bloom filter frequently for saving HDFS seek once we switch from one column to
062 * another in our UT. It's cpu time consuming (~45s for each case), so moved the ROWCOL case into a
063 * separated LargeTests to avoid timeout failure. <br>
064 * <br>
065 * To be clear: In TestMultiColumnScanner, we will flush 10 (NUM_FLUSHES=10) HFiles here, and the
066 * table will put ~1000 cells (rows=20, ts=6, qualifiers=8, total=20*6*8 ~ 1000) . Each full table
067 * scan will check the ROWCOL bloom filter 20 (rows)* 8 (column) * 10 (hfiles)= 1600 times, beside
068 * it will scan the full table 6*2^8=1536 times, so finally will have 1600*1536=2457600 bloom filter
069 * testing. (See HBASE-21520)
070 */
071public abstract class TestMultiColumnScanner {
072
073  private static final Logger LOG = LoggerFactory.getLogger(TestMultiColumnScanner.class);
074
075  private static final String TABLE_NAME = TestMultiColumnScanner.class.getSimpleName();
076
077  static final int MAX_VERSIONS = 50;
078
079  private static final String FAMILY = "CF";
080  private static final byte[] FAMILY_BYTES = Bytes.toBytes(FAMILY);
081
082  /**
083   * The size of the column qualifier set used. Increasing this parameter exponentially increases
084   * test time.
085   */
086  private static final int NUM_COLUMNS = 8;
087
088  private static final int MAX_COLUMN_BIT_MASK = 1 << NUM_COLUMNS - 1;
089  private static final int NUM_FLUSHES = 10;
090  private static final int NUM_ROWS = 20;
091
092  /** A large value of type long for use as a timestamp */
093  private static final long BIG_LONG = 9111222333444555666L;
094
095  /**
096   * Timestamps to test with. Cannot use {@link Long#MAX_VALUE} here, because it will be replaced by
097   * an timestamp auto-generated based on the time.
098   */
099  private static final long[] TIMESTAMPS =
100    new long[] { 1, 3, 5, Integer.MAX_VALUE, BIG_LONG, Long.MAX_VALUE - 1 };
101
102  /** The probability that a column is skipped in a store file. */
103  private static final double COLUMN_SKIP_IN_STORE_FILE_PROB = 0.7;
104
105  /** The probability to delete a row/column pair */
106  private static final double DELETE_PROBABILITY = 0.02;
107
108  private final static HBaseTestingUtil TEST_UTIL = new HBaseTestingUtil();
109
110  @Parameter(0)
111  public Compression.Algorithm comprAlgo;
112
113  @Parameter(1)
114  public BloomType bloomType;
115
116  @Parameter(2)
117  public DataBlockEncoding dataBlockEncoding;
118
119  // Some static sanity-checking.
120  static {
121    assertTrue(BIG_LONG > 0.9 * Long.MAX_VALUE); // Guard against typos.
122
123    // Ensure TIMESTAMPS are sorted.
124    for (int i = 0; i < TIMESTAMPS.length - 1; ++i)
125      assertTrue(TIMESTAMPS[i] < TIMESTAMPS[i + 1]);
126  }
127
128  public static Collection<Object[]> generateParams(Compression.Algorithm algo,
129    boolean useDataBlockEncoding) {
130    List<Object[]> parameters = new ArrayList<>();
131    for (BloomType bloomType : BloomType.values()) {
132      DataBlockEncoding dataBlockEncoding =
133        useDataBlockEncoding ? DataBlockEncoding.PREFIX : DataBlockEncoding.NONE;
134      parameters.add(new Object[] { algo, bloomType, dataBlockEncoding });
135    }
136    return parameters;
137  }
138
139  @Test
140  public void testMultiColumnScanner() throws IOException {
141    TEST_UTIL.getConfiguration().setInt(BloomFilterUtil.PREFIX_LENGTH_KEY, 10);
142    HRegion region = TEST_UTIL.createTestRegion(TABLE_NAME,
143      ColumnFamilyDescriptorBuilder.newBuilder(FAMILY_BYTES).setCompressionType(comprAlgo)
144        .setBloomFilterType(bloomType).setMaxVersions(MAX_VERSIONS)
145        .setDataBlockEncoding(dataBlockEncoding).build(),
146      BlockCacheFactory.createBlockCache(TEST_UTIL.getConfiguration()));
147    List<String> rows = sequentialStrings("row", NUM_ROWS);
148    List<String> qualifiers = sequentialStrings("qual", NUM_COLUMNS);
149    List<KeyValue> kvs = new ArrayList<>();
150    Set<String> keySet = new HashSet<>();
151
152    // A map from <row>_<qualifier> to the most recent delete timestamp for
153    // that column.
154    Map<String, Long> lastDelTimeMap = new HashMap<>();
155
156    Random rand = ThreadLocalRandom.current();
157    for (int iFlush = 0; iFlush < NUM_FLUSHES; ++iFlush) {
158      for (String qual : qualifiers) {
159        // This is where we decide to include or not include this column into
160        // this store file, regardless of row and timestamp.
161        if (rand.nextDouble() < COLUMN_SKIP_IN_STORE_FILE_PROB) continue;
162
163        byte[] qualBytes = Bytes.toBytes(qual);
164        for (String row : rows) {
165          Put p = new Put(Bytes.toBytes(row));
166          for (long ts : TIMESTAMPS) {
167            String value = createValue(row, qual, ts);
168            KeyValue kv = KeyValueTestUtil.create(row, FAMILY, qual, ts, value);
169            assertEquals(kv.getTimestamp(), ts);
170            p.add(kv);
171            String keyAsString = kv.toString();
172            if (!keySet.contains(keyAsString)) {
173              keySet.add(keyAsString);
174              kvs.add(kv);
175            }
176          }
177          region.put(p);
178
179          Delete d = new Delete(Bytes.toBytes(row));
180          boolean deletedSomething = false;
181          for (long ts : TIMESTAMPS)
182            if (rand.nextDouble() < DELETE_PROBABILITY) {
183              d.addColumns(FAMILY_BYTES, qualBytes, ts);
184              String rowAndQual = row + "_" + qual;
185              Long whenDeleted = lastDelTimeMap.get(rowAndQual);
186              lastDelTimeMap.put(rowAndQual, whenDeleted == null ? ts : Math.max(ts, whenDeleted));
187              deletedSomething = true;
188            }
189          if (deletedSomething) region.delete(d);
190        }
191      }
192      region.flush(true);
193    }
194
195    Collections.sort(kvs, CellComparatorImpl.COMPARATOR);
196    for (int maxVersions = 1; maxVersions <= TIMESTAMPS.length; ++maxVersions) {
197      for (int columnBitMask = 1; columnBitMask <= MAX_COLUMN_BIT_MASK; ++columnBitMask) {
198        Scan scan = new Scan();
199        scan.readVersions(maxVersions);
200        Set<String> qualSet = new TreeSet<>();
201        {
202          int columnMaskTmp = columnBitMask;
203          for (String qual : qualifiers) {
204            if ((columnMaskTmp & 1) != 0) {
205              scan.addColumn(FAMILY_BYTES, Bytes.toBytes(qual));
206              qualSet.add(qual);
207            }
208            columnMaskTmp >>= 1;
209          }
210          assertEquals(0, columnMaskTmp);
211        }
212
213        InternalScanner scanner = region.getScanner(scan);
214        List<ExtendedCell> results = new ArrayList<>();
215
216        int kvPos = 0;
217        int numResults = 0;
218        String queryInfo = "columns queried: " + qualSet + " (columnBitMask=" + columnBitMask
219          + "), maxVersions=" + maxVersions;
220
221        while (scanner.next(results) || results.size() > 0) {
222          for (ExtendedCell kv : results) {
223            while (
224              kvPos < kvs.size()
225                && !matchesQuery(kvs.get(kvPos), qualSet, maxVersions, lastDelTimeMap)
226            ) {
227              ++kvPos;
228            }
229            String rowQual = getRowQualStr(kv);
230            String deleteInfo = "";
231            Long lastDelTS = lastDelTimeMap.get(rowQual);
232            if (lastDelTS != null) {
233              deleteInfo =
234                "; last timestamp when row/column " + rowQual + " was deleted: " + lastDelTS;
235            }
236            assertTrue(
237              "Scanner returned additional key/value: " + kv + ", " + queryInfo + deleteInfo + ";",
238              kvPos < kvs.size());
239            assertTrue("Scanner returned wrong key/value; " + queryInfo + deleteInfo + ";",
240              PrivateCellUtil.equalsIgnoreMvccVersion(kvs.get(kvPos), kv));
241            ++kvPos;
242            ++numResults;
243          }
244          results.clear();
245        }
246        for (; kvPos < kvs.size(); ++kvPos) {
247          KeyValue remainingKV = kvs.get(kvPos);
248          assertFalse(
249            "Matching column not returned by scanner: " + remainingKV + ", " + queryInfo
250              + ", results returned: " + numResults,
251            matchesQuery(remainingKV, qualSet, maxVersions, lastDelTimeMap));
252        }
253      }
254    }
255    assertTrue("This test is supposed to delete at least some row/column " + "pairs",
256      lastDelTimeMap.size() > 0);
257    LOG.info("Number of row/col pairs deleted at least once: " + lastDelTimeMap.size());
258    HBaseTestingUtil.closeRegionAndWAL(region);
259  }
260
261  private static String getRowQualStr(Cell kv) {
262    String rowStr = Bytes.toString(CellUtil.cloneRow(kv));
263    String qualStr = Bytes.toString(CellUtil.cloneQualifier(kv));
264    return rowStr + "_" + qualStr;
265  }
266
267  private static boolean matchesQuery(KeyValue kv, Set<String> qualSet, int maxVersions,
268    Map<String, Long> lastDelTimeMap) {
269    Long lastDelTS = lastDelTimeMap.get(getRowQualStr(kv));
270    long ts = kv.getTimestamp();
271    return qualSet.contains(qualStr(kv)) && ts >= TIMESTAMPS[TIMESTAMPS.length - maxVersions]
272      && (lastDelTS == null || ts > lastDelTS);
273  }
274
275  private static String qualStr(KeyValue kv) {
276    return Bytes.toString(kv.getQualifierArray(), kv.getQualifierOffset(), kv.getQualifierLength());
277  }
278
279  static String createValue(String row, String qual, long ts) {
280    return "value_for_" + row + "_" + qual + "_" + ts;
281  }
282
283  private static List<String> sequentialStrings(String prefix, int n) {
284    List<String> lst = new ArrayList<>();
285    for (int i = 0; i < n; ++i) {
286      StringBuilder sb = new StringBuilder();
287      sb.append(prefix + i);
288
289      // Make column length depend on i.
290      int iBitShifted = i;
291      while (iBitShifted != 0) {
292        sb.append((iBitShifted & 1) == 0 ? 'a' : 'b');
293        iBitShifted >>= 1;
294      }
295
296      lst.add(sb.toString());
297    }
298    return lst;
299  }
300}