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.io.hfile;
019
020import static org.junit.jupiter.api.Assertions.assertEquals;
021import static org.junit.jupiter.api.Assertions.assertTrue;
022
023import java.io.IOException;
024import java.util.stream.Stream;
025import org.apache.hadoop.hbase.HBaseParameterizedTestTemplate;
026import org.apache.hadoop.hbase.HBaseTestingUtil;
027import org.apache.hadoop.hbase.client.ColumnFamilyDescriptor;
028import org.apache.hadoop.hbase.client.ColumnFamilyDescriptorBuilder;
029import org.apache.hadoop.hbase.client.Get;
030import org.apache.hadoop.hbase.client.Put;
031import org.apache.hadoop.hbase.io.compress.Compression;
032import org.apache.hadoop.hbase.io.compress.Compression.Algorithm;
033import org.apache.hadoop.hbase.regionserver.BloomType;
034import org.apache.hadoop.hbase.regionserver.HRegion;
035import org.apache.hadoop.hbase.testclassification.IOTests;
036import org.apache.hadoop.hbase.testclassification.MediumTests;
037import org.apache.hadoop.hbase.util.Bytes;
038import org.junit.jupiter.api.BeforeEach;
039import org.junit.jupiter.api.Tag;
040import org.junit.jupiter.api.TestTemplate;
041import org.junit.jupiter.params.provider.Arguments;
042
043/**
044 * Make sure we always cache important block types, such as index blocks, as long as we have a block
045 * cache, even though block caching might be disabled for the column family.
046 * <p>
047 * TODO: This test writes a lot of data and only tests the most basic of metrics. Cache stats need
048 * to reveal more about what is being cached whether DATA or INDEX blocks and then we could do more
049 * verification in this test.
050 */
051@Tag(IOTests.TAG)
052@Tag(MediumTests.TAG)
053@HBaseParameterizedTestTemplate(name = "{index}: hfileVersion={0}, cfCacheEnabled={1}")
054public class TestForceCacheImportantBlocks {
055
056  private final HBaseTestingUtil TEST_UTIL = new HBaseTestingUtil();
057
058  private static final String TABLE = "myTable";
059  private static final String CF = "myCF";
060  private static final byte[] CF_BYTES = Bytes.toBytes(CF);
061  private static final int MAX_VERSIONS = 3;
062  private static final int NUM_HFILES = 5;
063
064  private static final int ROWS_PER_HFILE = 100;
065  private static final int NUM_ROWS = NUM_HFILES * ROWS_PER_HFILE;
066  private static final int NUM_COLS_PER_ROW = 50;
067  private static final int NUM_TIMESTAMPS_PER_COL = 50;
068
069  /** Extremely small block size, so that we can get some index blocks */
070  private static final int BLOCK_SIZE = 256;
071
072  private static final Algorithm COMPRESSION_ALGORITHM = Compression.Algorithm.GZ;
073  private static final BloomType BLOOM_TYPE = BloomType.ROW;
074
075  @SuppressWarnings("unused")
076  // Currently unused.
077  private final int hfileVersion;
078  private final boolean cfCacheEnabled;
079
080  public static Stream<Arguments> parameters() {
081    return Stream.of(Arguments.of(3, true), Arguments.of(3, false));
082  }
083
084  public TestForceCacheImportantBlocks(int hfileVersion, boolean cfCacheEnabled) {
085    this.hfileVersion = hfileVersion;
086    this.cfCacheEnabled = cfCacheEnabled;
087    TEST_UTIL.getConfiguration().setInt(HFile.FORMAT_VERSION_KEY, hfileVersion);
088  }
089
090  @BeforeEach
091  public void setup() {
092    HFile.DATABLOCK_READ_COUNT.reset();
093  }
094
095  @TestTemplate
096  public void testCacheBlocks() throws IOException {
097    // Set index block size to be the same as normal block size.
098    TEST_UTIL.getConfiguration().setInt(HFileBlockIndex.MAX_CHUNK_SIZE_KEY, BLOCK_SIZE);
099    BlockCache blockCache = BlockCacheFactory.createBlockCache(TEST_UTIL.getConfiguration());
100    ColumnFamilyDescriptor cfd =
101      ColumnFamilyDescriptorBuilder.newBuilder(Bytes.toBytes(CF)).setMaxVersions(MAX_VERSIONS)
102        .setCompressionType(COMPRESSION_ALGORITHM).setBloomFilterType(BLOOM_TYPE)
103        .setBlocksize(BLOCK_SIZE).setBlockCacheEnabled(cfCacheEnabled).build();
104    HRegion region = TEST_UTIL.createTestRegion(TABLE, cfd, blockCache);
105    CacheStats stats = blockCache.getStats();
106    writeTestData(region);
107    assertEquals(0, stats.getHitCount());
108    assertEquals(0, HFile.DATABLOCK_READ_COUNT.sum());
109    // Do a single get, take count of caches. If we are NOT caching DATA blocks, the miss
110    // count should go up. Otherwise, all should be cached and the miss count should not rise.
111    region.get(new Get(Bytes.toBytes("row" + 0)));
112    assertTrue(stats.getHitCount() > 0);
113    assertTrue(HFile.DATABLOCK_READ_COUNT.sum() > 0);
114    long missCount = stats.getMissCount();
115    region.get(new Get(Bytes.toBytes("row" + 0)));
116    if (this.cfCacheEnabled) assertEquals(missCount, stats.getMissCount());
117    else assertTrue(stats.getMissCount() > missCount);
118  }
119
120  private void writeTestData(HRegion region) throws IOException {
121    for (int i = 0; i < NUM_ROWS; ++i) {
122      Put put = new Put(Bytes.toBytes("row" + i));
123      for (int j = 0; j < NUM_COLS_PER_ROW; ++j) {
124        for (long ts = 1; ts < NUM_TIMESTAMPS_PER_COL; ++ts) {
125          put.addColumn(CF_BYTES, Bytes.toBytes("col" + j), ts,
126            Bytes.toBytes("value" + i + "_" + j + "_" + ts));
127        }
128      }
129      region.put(put);
130      if ((i + 1) % ROWS_PER_HFILE == 0) {
131        region.flush(true);
132      }
133    }
134  }
135}