hbase-server/src/test/java/org/apache/hadoop/hbase/wal/TestWALOpenAfterDNRollingStart.java

   1 /**
   2  * Licensed to the Apache Software Foundation (ASF) under one
   3  * or more contributor license agreements.  See the NOTICE file
   4  * distributed with this work for additional information
   5  * regarding copyright ownership.  The ASF licenses this file
   6  * to you under the Apache License, Version 2.0 (the
   7  * "License"); you may not use this file except in compliance
   8  * with the License.  You may obtain a copy of the License at
   9  *
  10  *     http://www.apache.org/licenses/LICENSE-2.0
  11  *
  12  * Unless required by applicable law or agreed to in writing, software
  13  * distributed under the License is distributed on an "AS IS" BASIS,
  14  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  15  * See the License for the specific language governing permissions and
  16  * limitations under the License.
  17  */
  18 package org.apache.hadoop.hbase.wal;
  19
  20 import java.io.IOException;
  21 import java.util.Arrays;
  22 import java.util.List;
  23 import org.apache.hadoop.fs.Path;
  24 import org.apache.hadoop.hbase.HBaseClassTestRule;
  25 import org.apache.hadoop.hbase.HBaseTestingUtility;
  26 import org.apache.hadoop.hbase.HConstants;
  27 import org.apache.hadoop.hbase.regionserver.HRegionServer;
  28 import org.apache.hadoop.hbase.regionserver.wal.AbstractFSWAL;
  29 import org.apache.hadoop.hbase.testclassification.LargeTests;
  30 import org.apache.hadoop.hbase.testclassification.RegionServerTests;
  31 import org.apache.hadoop.hbase.util.FSUtils;
  32 import org.junit.After;
  33 import org.junit.AfterClass;
  34 import org.junit.Before;
  35 import org.junit.BeforeClass;
  36 import org.junit.ClassRule;
  37 import org.junit.Test;
  38 import org.junit.experimental.categories.Category;
  39 import org.junit.runner.RunWith;
  40 import org.junit.runners.Parameterized;
  41 import org.junit.runners.Parameterized.Parameter;
  42 import org.junit.runners.Parameterized.Parameters;
  43
  44 @RunWith(Parameterized.class)
  45 @Category({ RegionServerTests.class, LargeTests.class })
  46 public class TestWALOpenAfterDNRollingStart {
  47
  48   @ClassRule
  49   public static final HBaseClassTestRule CLASS_RULE =
  50       HBaseClassTestRule.forClass(TestWALOpenAfterDNRollingStart.class);
  51
  52   private static final HBaseTestingUtility TEST_UTIL = new HBaseTestingUtility();
  53   // Sleep time before restart next dn, we need to wait the current dn to finish start up
  54   private static long DN_RESTART_INTERVAL = 15000;
  55
  56   // interval of checking low replication. The sleep time must smaller than
  57   // DataNodeRestartInterval
  58   // so a low replication case will be detected and the wal will be rolled
  59   private static long CHECK_LOW_REPLICATION_INTERVAL = 10000;
  60
  61   @Parameter
  62   public String walProvider;
  63
  64   @Parameters(name = "{index}: wal={0}")
  65   public static List<Object[]> data() {
  66     return Arrays.asList(new Object[] { "asyncfs" }, new Object[] { "filesystem" });
  67   }
  68
  69   @BeforeClass
  70   public static void setUpBeforeClass() throws Exception {
  71     // don't let hdfs client to choose a new replica when dn down
  72     TEST_UTIL.getConfiguration()
  73         .setBoolean("dfs.client.block.write.replace-datanode-on-failure.enable", false);
  74     TEST_UTIL.getConfiguration().setLong("hbase.regionserver.hlog.check.lowreplication.interval",
  75       CHECK_LOW_REPLICATION_INTERVAL);
  76     TEST_UTIL.startMiniDFSCluster(3);
  77     TEST_UTIL.startMiniZKCluster();
  78   }
  79
  80   @Before
  81   public void setUp() throws IOException, InterruptedException {
  82     TEST_UTIL.getConfiguration().set("hbase.wal.provider", walProvider);
  83     TEST_UTIL.startMiniHBaseCluster();
  84   }
  85
  86   @After
  87   public void tearDown() throws Exception {
  88     TEST_UTIL.shutdownMiniHBaseCluster();
  89   }
  90
  91   @AfterClass
  92   public static void tearDownAfterClass() throws Exception {
  93     TEST_UTIL.shutdownMiniCluster();
  94   }
  95
  96   /**
  97    * see HBASE-18132 This is a test case of failing open a wal(for replication for example) after
  98    * all datanode restarted (rolling upgrade, for example). Before this patch, low replication
  99    * detection is only used when syncing wal. But if the wal haven't had any entry whiten, it will
 100    * never know all the replica of the wal is broken(because of dn restarting). And this wal can
 101    * never be open
 102    * @throws Exception
 103    */
 104   @Test
 105   public void test() throws Exception {
 106     HRegionServer server = TEST_UTIL.getHBaseCluster().getRegionServer(0);
 107     AbstractFSWAL<?> wal = (AbstractFSWAL<?>) server.getWAL(null);
 108     Path currentFile = wal.getCurrentFileName();
 109     // restart every dn to simulate a dn rolling upgrade
 110     for (int i = 0, n = TEST_UTIL.getDFSCluster().getDataNodes().size(); i < n; i++) {
 111       // This is NOT a bug, when restart dn in miniDFSCluster, it will remove the stopped dn from
 112       // the dn list and then add to the tail of this list, we need to always restart the first one
 113       // to simulate rolling upgrade of every dn.
 114       TEST_UTIL.getDFSCluster().restartDataNode(0);
 115       // sleep enough time so log roller can detect the pipeline break and roll log
 116       Thread.sleep(DN_RESTART_INTERVAL);
 117     }
 118
 119     if (!server.getFileSystem().exists(currentFile)) {
 120       Path walRootDir = FSUtils.getWALRootDir(TEST_UTIL.getConfiguration());
 121       final Path oldLogDir = new Path(walRootDir, HConstants.HREGION_OLDLOGDIR_NAME);
 122       currentFile = new Path(oldLogDir, currentFile.getName());
 123     }
 124     // if the log is not rolled, then we can never open this wal forever.
 125     try (WAL.Reader reader = WALFactory.createReader(TEST_UTIL.getTestFileSystem(), currentFile,
 126       TEST_UTIL.getConfiguration())) {
 127       reader.next();
 128     }
 129   }
 130 }