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.junit.jupiter.api.Assertions.assertEquals;
021import static org.junit.jupiter.api.Assertions.assertFalse;
022import static org.junit.jupiter.api.Assertions.assertNull;
023import static org.junit.jupiter.api.Assertions.assertTrue;
024
025import java.io.IOException;
026import java.util.HashSet;
027import java.util.NavigableMap;
028import java.util.Set;
029import java.util.TreeMap;
030import java.util.concurrent.ThreadLocalRandom;
031import org.apache.hadoop.conf.Configuration;
032import org.apache.hadoop.fs.FileStatus;
033import org.apache.hadoop.fs.FileSystem;
034import org.apache.hadoop.fs.Path;
035import org.apache.hadoop.hbase.HBaseTestingUtil;
036import org.apache.hadoop.hbase.HConstants;
037import org.apache.hadoop.hbase.KeyValue;
038import org.apache.hadoop.hbase.ServerName;
039import org.apache.hadoop.hbase.TableName;
040import org.apache.hadoop.hbase.client.ColumnFamilyDescriptorBuilder;
041import org.apache.hadoop.hbase.client.RegionInfo;
042import org.apache.hadoop.hbase.client.RegionInfoBuilder;
043import org.apache.hadoop.hbase.client.TableDescriptor;
044import org.apache.hadoop.hbase.client.TableDescriptorBuilder;
045import org.apache.hadoop.hbase.regionserver.MultiVersionConcurrencyControl;
046import org.apache.hadoop.hbase.testclassification.MediumTests;
047import org.apache.hadoop.hbase.testclassification.RegionServerTests;
048import org.apache.hadoop.hbase.util.Bytes;
049import org.apache.hadoop.hbase.util.CommonFSUtils;
050import org.apache.hadoop.hbase.util.EnvironmentEdgeManager;
051import org.junit.jupiter.api.AfterAll;
052import org.junit.jupiter.api.BeforeAll;
053import org.junit.jupiter.api.BeforeEach;
054import org.junit.jupiter.api.Tag;
055import org.junit.jupiter.api.Test;
056import org.junit.jupiter.api.TestInfo;
057import org.slf4j.Logger;
058import org.slf4j.LoggerFactory;
059
060@Tag(RegionServerTests.TAG)
061@Tag(MediumTests.TAG)
062public class TestFSHLogProvider {
063
064  private static final Logger LOG = LoggerFactory.getLogger(TestFSHLogProvider.class);
065
066  private static Configuration conf;
067  private static FileSystem fs;
068  private final static HBaseTestingUtil TEST_UTIL = new HBaseTestingUtil();
069  private MultiVersionConcurrencyControl mvcc;
070
071  private String currentTestName;
072
073  @BeforeEach
074  public void setUp(TestInfo testInfo) throws Exception {
075    currentTestName = testInfo.getTestMethod().get().getName();
076    mvcc = new MultiVersionConcurrencyControl();
077    FileStatus[] entries = fs.listStatus(new Path("/"));
078    for (FileStatus dir : entries) {
079      fs.delete(dir.getPath(), true);
080    }
081  }
082
083  @BeforeAll
084  public static void setUpBeforeClass() throws Exception {
085    // Make block sizes small.
086    TEST_UTIL.getConfiguration().setInt("dfs.blocksize", 1024 * 1024);
087    // quicker heartbeat interval for faster DN death notification
088    TEST_UTIL.getConfiguration().setInt("dfs.namenode.heartbeat.recheck-interval", 5000);
089    TEST_UTIL.getConfiguration().setInt("dfs.heartbeat.interval", 1);
090    TEST_UTIL.getConfiguration().setInt("dfs.client.socket-timeout", 5000);
091
092    // faster failover with cluster.shutdown();fs.close() idiom
093    TEST_UTIL.getConfiguration().setInt("hbase.ipc.client.connect.max.retries", 1);
094    TEST_UTIL.getConfiguration().setInt("dfs.client.block.recovery.retries", 1);
095    TEST_UTIL.getConfiguration().setInt("hbase.ipc.client.connection.maxidletime", 500);
096    TEST_UTIL.startMiniDFSCluster(3);
097
098    // Set up a working space for our tests.
099    TEST_UTIL.createRootDir();
100    conf = TEST_UTIL.getConfiguration();
101    fs = TEST_UTIL.getDFSCluster().getFileSystem();
102  }
103
104  @AfterAll
105  public static void tearDownAfterClass() throws Exception {
106    TEST_UTIL.shutdownMiniCluster();
107  }
108
109  @Test
110  public void testGetServerNameFromWALDirectoryName() throws IOException {
111    ServerName sn = ServerName.valueOf("hn", 450, 1398);
112    String hl = CommonFSUtils.getRootDir(conf) + "/"
113      + AbstractFSWALProvider.getWALDirectoryName(sn.toString());
114
115    // Must not throw exception
116    assertNull(AbstractFSWALProvider.getServerNameFromWALDirectoryName(conf, null));
117    assertNull(AbstractFSWALProvider.getServerNameFromWALDirectoryName(conf,
118      CommonFSUtils.getRootDir(conf).toUri().toString()));
119    assertNull(AbstractFSWALProvider.getServerNameFromWALDirectoryName(conf, ""));
120    assertNull(AbstractFSWALProvider.getServerNameFromWALDirectoryName(conf, "                  "));
121    assertNull(AbstractFSWALProvider.getServerNameFromWALDirectoryName(conf, hl));
122    assertNull(AbstractFSWALProvider.getServerNameFromWALDirectoryName(conf, hl + "qdf"));
123    assertNull(AbstractFSWALProvider.getServerNameFromWALDirectoryName(conf, "sfqf" + hl + "qdf"));
124
125    final String wals = "/WALs/";
126    ServerName parsed = AbstractFSWALProvider.getServerNameFromWALDirectoryName(conf,
127      CommonFSUtils.getRootDir(conf).toUri().toString() + wals + sn
128        + "/localhost%2C32984%2C1343316388997.1343316390417");
129    assertEquals(sn, parsed, "standard");
130
131    parsed = AbstractFSWALProvider.getServerNameFromWALDirectoryName(conf, hl + "/qdf");
132    assertEquals(sn, parsed, "subdir");
133
134    parsed = AbstractFSWALProvider.getServerNameFromWALDirectoryName(conf,
135      CommonFSUtils.getRootDir(conf).toUri().toString() + wals + sn
136        + "-splitting/localhost%3A57020.1340474893931");
137    assertEquals(sn, parsed, "split");
138  }
139
140  private void addEdits(WAL log, RegionInfo hri, TableDescriptor htd, int times,
141    NavigableMap<byte[], Integer> scopes) throws IOException {
142    final byte[] row = Bytes.toBytes("row");
143    for (int i = 0; i < times; i++) {
144      long timestamp = EnvironmentEdgeManager.currentTime();
145      WALEdit cols = new WALEdit();
146      cols.add(new KeyValue(row, row, row, timestamp, row));
147      log.appendData(hri,
148        getWalKey(hri.getEncodedNameAsBytes(), htd.getTableName(), timestamp, scopes), cols);
149    }
150    log.sync();
151  }
152
153  /**
154   * used by TestDefaultWALProviderWithHLogKey
155   */
156  private WALKeyImpl getWalKey(final byte[] info, final TableName tableName, final long timestamp,
157    NavigableMap<byte[], Integer> scopes) {
158    return new WALKeyImpl(info, tableName, timestamp, mvcc, scopes);
159  }
160
161  /**
162   * helper method to simulate region flush for a WAL.
163   */
164  private void flushRegion(WAL wal, byte[] regionEncodedName, Set<byte[]> flushedFamilyNames) {
165    wal.startCacheFlush(regionEncodedName, flushedFamilyNames);
166    wal.completeCacheFlush(regionEncodedName, HConstants.NO_SEQNUM);
167  }
168
169  // now we will close asynchronously and will not archive a wal file unless it is fully closed, so
170  // sometimes we need to wait a bit before asserting, especially when you want to test the removal
171  // of numRolledLogFiles
172  private void waitNumRolledLogFiles(WAL wal, int expected) {
173    TEST_UTIL.waitFor(5000, () -> AbstractFSWALProvider.getNumRolledLogFiles(wal) == expected);
174  }
175
176  private void testLogCleaning(WALFactory wals) throws IOException {
177    TableDescriptor htd = TableDescriptorBuilder.newBuilder(TableName.valueOf(currentTestName))
178      .setColumnFamily(ColumnFamilyDescriptorBuilder.of("row")).build();
179    TableDescriptor htd2 =
180      TableDescriptorBuilder.newBuilder(TableName.valueOf(currentTestName + "2"))
181        .setColumnFamily(ColumnFamilyDescriptorBuilder.of("row")).build();
182    NavigableMap<byte[], Integer> scopes1 = new TreeMap<>(Bytes.BYTES_COMPARATOR);
183    for (byte[] fam : htd.getColumnFamilyNames()) {
184      scopes1.put(fam, 0);
185    }
186    NavigableMap<byte[], Integer> scopes2 = new TreeMap<>(Bytes.BYTES_COMPARATOR);
187    for (byte[] fam : htd2.getColumnFamilyNames()) {
188      scopes2.put(fam, 0);
189    }
190    RegionInfo hri = RegionInfoBuilder.newBuilder(htd.getTableName()).build();
191    RegionInfo hri2 = RegionInfoBuilder.newBuilder(htd2.getTableName()).build();
192    // we want to mix edits from regions, so pick our own identifier.
193    WAL log = wals.getWAL(null);
194
195    // Add a single edit and make sure that rolling won't remove the file
196    // Before HBASE-3198 it used to delete it
197    addEdits(log, hri, htd, 1, scopes1);
198    log.rollWriter();
199    waitNumRolledLogFiles(log, 1);
200
201    // See if there's anything wrong with more than 1 edit
202    addEdits(log, hri, htd, 2, scopes1);
203    log.rollWriter();
204    assertEquals(2, FSHLogProvider.getNumRolledLogFiles(log));
205
206    // Now mix edits from 2 regions, still no flushing
207    addEdits(log, hri, htd, 1, scopes1);
208    addEdits(log, hri2, htd2, 1, scopes2);
209    addEdits(log, hri, htd, 1, scopes1);
210    addEdits(log, hri2, htd2, 1, scopes2);
211    log.rollWriter();
212    waitNumRolledLogFiles(log, 3);
213
214    // Flush the first region, we expect to see the first two files getting
215    // archived. We need to append something or writer won't be rolled.
216    addEdits(log, hri2, htd2, 1, scopes2);
217    log.startCacheFlush(hri.getEncodedNameAsBytes(), htd.getColumnFamilyNames());
218    log.completeCacheFlush(hri.getEncodedNameAsBytes(), HConstants.NO_SEQNUM);
219    log.rollWriter();
220    waitNumRolledLogFiles(log, 2);
221
222    // Flush the second region, which removes all the remaining output files
223    // since the oldest was completely flushed and the two others only contain
224    // flush information
225    addEdits(log, hri2, htd2, 1, scopes2);
226    log.startCacheFlush(hri2.getEncodedNameAsBytes(), htd2.getColumnFamilyNames());
227    log.completeCacheFlush(hri2.getEncodedNameAsBytes(), HConstants.NO_SEQNUM);
228    log.rollWriter();
229    waitNumRolledLogFiles(log, 0);
230  }
231
232  @Test
233  public void testLogCleaning() throws Exception {
234    LOG.info(currentTestName);
235    Configuration localConf = new Configuration(conf);
236    localConf.set(WALFactory.WAL_PROVIDER, FSHLogProvider.class.getName());
237    WALFactory wals = new WALFactory(localConf, currentTestName);
238    try {
239      testLogCleaning(wals);
240    } finally {
241      wals.close();
242    }
243  }
244
245  private void testWALArchiving(WALFactory wals) throws IOException {
246    TableDescriptor table1 =
247      TableDescriptorBuilder.newBuilder(TableName.valueOf(currentTestName + "1"))
248        .setColumnFamily(ColumnFamilyDescriptorBuilder.of("row")).build();
249    TableDescriptor table2 =
250      TableDescriptorBuilder.newBuilder(TableName.valueOf(currentTestName + "2"))
251        .setColumnFamily(ColumnFamilyDescriptorBuilder.of("row")).build();
252    NavigableMap<byte[], Integer> scopes1 = new TreeMap<>(Bytes.BYTES_COMPARATOR);
253    for (byte[] fam : table1.getColumnFamilyNames()) {
254      scopes1.put(fam, 0);
255    }
256    NavigableMap<byte[], Integer> scopes2 = new TreeMap<>(Bytes.BYTES_COMPARATOR);
257    for (byte[] fam : table2.getColumnFamilyNames()) {
258      scopes2.put(fam, 0);
259    }
260    WAL wal = wals.getWAL(null);
261    assertEquals(0, AbstractFSWALProvider.getNumRolledLogFiles(wal));
262    RegionInfo hri1 = RegionInfoBuilder.newBuilder(table1.getTableName()).build();
263    RegionInfo hri2 = RegionInfoBuilder.newBuilder(table2.getTableName()).build();
264    // variables to mock region sequenceIds.
265    // start with the testing logic: insert a waledit, and roll writer
266    addEdits(wal, hri1, table1, 1, scopes1);
267    wal.rollWriter();
268    // assert that the wal is rolled
269    waitNumRolledLogFiles(wal, 1);
270    // add edits in the second wal file, and roll writer.
271    addEdits(wal, hri1, table1, 1, scopes1);
272    wal.rollWriter();
273    // assert that the wal is rolled
274    waitNumRolledLogFiles(wal, 2);
275    // add a waledit to table1, and flush the region.
276    addEdits(wal, hri1, table1, 3, scopes1);
277    flushRegion(wal, hri1.getEncodedNameAsBytes(), table1.getColumnFamilyNames());
278    // roll log; all old logs should be archived.
279    wal.rollWriter();
280    waitNumRolledLogFiles(wal, 0);
281    // add an edit to table2, and roll writer
282    addEdits(wal, hri2, table2, 1, scopes2);
283    wal.rollWriter();
284    waitNumRolledLogFiles(wal, 1);
285    // add edits for table1, and roll writer
286    addEdits(wal, hri1, table1, 2, scopes1);
287    wal.rollWriter();
288    waitNumRolledLogFiles(wal, 2);
289    // add edits for table2, and flush hri1.
290    addEdits(wal, hri2, table2, 2, scopes2);
291    flushRegion(wal, hri1.getEncodedNameAsBytes(), table2.getColumnFamilyNames());
292    // the log : region-sequenceId map is
293    // log1: region2 (unflushed)
294    // log2: region1 (flushed)
295    // log3: region2 (unflushed)
296    // roll the writer; log2 should be archived.
297    wal.rollWriter();
298    waitNumRolledLogFiles(wal, 2);
299    // flush region2, and all logs should be archived.
300    addEdits(wal, hri2, table2, 2, scopes2);
301    flushRegion(wal, hri2.getEncodedNameAsBytes(), table2.getColumnFamilyNames());
302    wal.rollWriter();
303    waitNumRolledLogFiles(wal, 0);
304  }
305
306  /**
307   * Tests wal archiving by adding data, doing flushing/rolling and checking we archive old logs and
308   * also don't archive "live logs" (that is, a log with un-flushed entries).
309   * <p>
310   * This is what it does: It creates two regions, and does a series of inserts along with log
311   * rolling. Whenever a WAL is rolled, HLogBase checks previous wals for archiving. A wal is
312   * eligible for archiving if for all the regions which have entries in that wal file, have flushed
313   * - past their maximum sequence id in that wal file.
314   * <p>
315   */
316  @Test
317  public void testWALArchiving() throws IOException {
318    LOG.debug(currentTestName);
319
320    Configuration localConf = new Configuration(conf);
321    localConf.set(WALFactory.WAL_PROVIDER, FSHLogProvider.class.getName());
322    WALFactory wals = new WALFactory(localConf, currentTestName);
323    try {
324      testWALArchiving(wals);
325    } finally {
326      wals.close();
327    }
328  }
329
330  /**
331   * Ensure that we can use Set.add to deduplicate WALs
332   */
333  @Test
334  public void setMembershipDedups() throws IOException {
335    Configuration localConf = new Configuration(conf);
336    localConf.set(WALFactory.WAL_PROVIDER, FSHLogProvider.class.getName());
337    WALFactory wals = new WALFactory(localConf, currentTestName);
338    try {
339      final Set<WAL> seen = new HashSet<>(1);
340      assertTrue(seen.add(wals.getWAL(null)),
341        "first attempt to add WAL from default provider should work.");
342      for (int i = 0; i < 1000; i++) {
343        assertFalse(
344          seen.add(wals.getWAL(RegionInfoBuilder
345            .newBuilder(TableName.valueOf("Table-" + ThreadLocalRandom.current().nextInt()))
346            .build())),
347          "default wal provider is only supposed to return a single wal, which should "
348            + "compare as .equals itself.");
349      }
350    } finally {
351      wals.close();
352    }
353  }
354}