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;
019
020import static org.junit.jupiter.api.Assertions.assertEquals;
021
022import org.apache.hadoop.conf.Configuration;
023import org.apache.hadoop.hbase.Waiter.ExplainingPredicate;
024import org.apache.hadoop.hbase.client.Table;
025import org.apache.hadoop.hbase.testclassification.MediumTests;
026import org.apache.hadoop.hbase.testclassification.MiscTests;
027import org.apache.hadoop.hbase.util.Bytes;
028import org.apache.hadoop.hbase.util.JVMClusterUtil.RegionServerThread;
029import org.junit.jupiter.api.AfterAll;
030import org.junit.jupiter.api.BeforeAll;
031import org.junit.jupiter.api.Tag;
032import org.junit.jupiter.api.Test;
033import org.slf4j.Logger;
034import org.slf4j.LoggerFactory;
035
036@Tag(MiscTests.TAG)
037@Tag(MediumTests.TAG)
038public class TestFullLogReconstruction {
039  private static final Logger LOG = LoggerFactory.getLogger(TestFullLogReconstruction.class);
040
041  private final static HBaseTestingUtil TEST_UTIL = new HBaseTestingUtil();
042
043  private final static TableName TABLE_NAME = TableName.valueOf("tabletest");
044  private final static byte[] FAMILY = Bytes.toBytes("family");
045
046  @BeforeAll
047  public static void setUpBeforeClass() throws Exception {
048    Configuration c = TEST_UTIL.getConfiguration();
049    // quicker heartbeat interval for faster DN death notification
050    c.setInt("dfs.namenode.heartbeat.recheck-interval", 5000);
051    c.setInt("dfs.heartbeat.interval", 1);
052    c.setInt("dfs.client.socket-timeout", 5000);
053    // faster failover with cluster.shutdown();fs.close() idiom
054    c.setInt("hbase.ipc.client.connect.max.retries", 1);
055    c.setInt("dfs.client.block.recovery.retries", 1);
056    c.setInt(HConstants.ZK_SESSION_TIMEOUT, 1000);
057    TEST_UTIL.startMiniCluster(3);
058  }
059
060  @AfterAll
061  public static void tearDownAfterClass() throws Exception {
062    TEST_UTIL.shutdownMiniCluster();
063  }
064
065  /**
066   * Test the whole reconstruction loop. Build a table with regions aaa to zzz and load every one of
067   * them multiple times with the same date and do a flush at some point. Kill one of the region
068   * servers and scan the table. We should see all the rows.
069   */
070  @Test
071  public void testReconstruction() throws Exception {
072    Table table = TEST_UTIL.createMultiRegionTable(TABLE_NAME, FAMILY);
073
074    // Load up the table with simple rows and count them
075    int initialCount = TEST_UTIL.loadTable(table, FAMILY);
076    int count = HBaseTestingUtil.countRows(table);
077
078    assertEquals(initialCount, count);
079
080    for (int i = 0; i < 4; i++) {
081      TEST_UTIL.loadTable(table, FAMILY);
082    }
083    RegionServerThread rsThread = TEST_UTIL.getHBaseCluster().getRegionServerThreads().get(0);
084    int index = 0;
085    LOG.info("Expiring {}", TEST_UTIL.getMiniHBaseCluster().getRegionServer(index));
086    TEST_UTIL.expireRegionServerSession(index);
087    // make sure that the RS is fully down before reading, so that we will read the data from other
088    // RSes.
089    TEST_UTIL.waitFor(30000, new ExplainingPredicate<Exception>() {
090
091      @Override
092      public boolean evaluate() throws Exception {
093        return !rsThread.isAlive();
094      }
095
096      @Override
097      public String explainFailure() throws Exception {
098        return rsThread.getRegionServer() + " is still alive";
099      }
100    });
101    LOG.info("Starting count");
102
103    int newCount = HBaseTestingUtil.countRows(table);
104    assertEquals(count, newCount);
105    table.close();
106  }
107}