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.master;
019
020import static org.junit.jupiter.api.Assertions.assertEquals;
021import static org.junit.jupiter.api.Assertions.assertInstanceOf;
022import static org.junit.jupiter.api.Assertions.assertTrue;
023
024import org.apache.hadoop.hbase.HBaseTestingUtil;
025import org.apache.hadoop.hbase.ServerName;
026import org.apache.hadoop.hbase.TableName;
027import org.apache.hadoop.hbase.client.RegionInfo;
028import org.apache.hadoop.hbase.client.Table;
029import org.apache.hadoop.hbase.master.assignment.AssignmentTestingUtil;
030import org.apache.hadoop.hbase.testclassification.MasterTests;
031import org.apache.hadoop.hbase.testclassification.MediumTests;
032import org.apache.hadoop.hbase.util.Bytes;
033import org.apache.hadoop.metrics2.AbstractMetric;
034import org.apache.hadoop.metrics2.MetricsRecord;
035import org.apache.hadoop.metrics2.MetricsSource;
036import org.apache.hadoop.metrics2.impl.MetricsCollectorImpl;
037import org.junit.jupiter.api.AfterAll;
038import org.junit.jupiter.api.BeforeAll;
039import org.junit.jupiter.api.Tag;
040import org.junit.jupiter.api.Test;
041import org.junit.jupiter.api.TestInfo;
042
043@Tag(MasterTests.TAG)
044@Tag(MediumTests.TAG)
045public class TestAssignmentManagerRitDurationMetrics {
046
047  private static final HBaseTestingUtil TEST_UTIL = new HBaseTestingUtil();
048  private static final byte[] FAMILY = Bytes.toBytes("family");
049  private static final long WAIT_TIMEOUT_MS = 10_000L;
050
051  private static HMaster MASTER;
052  private static final String RIT_DURATION_NUM_OPS_METRIC = "RitDuration_num_ops";
053
054  @BeforeAll
055  public static void startCluster() throws Exception {
056    TEST_UTIL.startMiniCluster(2);
057    MASTER = TEST_UTIL.getMiniHBaseCluster().getMaster();
058  }
059
060  @AfterAll
061  public static void after() throws Exception {
062    TEST_UTIL.shutdownMiniCluster();
063  }
064
065  @Test
066  public void testRitDurationHistogramMetric(TestInfo testInfo) throws Exception {
067    TableName tableName = TableName.valueOf(testInfo.getTestMethod().orElseThrow().getName());
068    try (Table table = TEST_UTIL.createTable(tableName, FAMILY)) {
069      RegionInfo regionInfo =
070        MASTER.getAssignmentManager().getRegionStates().getRegionsOfTable(tableName).get(0);
071      TEST_UTIL.waitFor(WAIT_TIMEOUT_MS,
072        () -> !AssignmentTestingUtil.isRegionInTransition(regionInfo, MASTER.getAssignmentManager())
073          && MASTER.getAssignmentManager().getRegionStates().getRegionServerOfRegion(regionInfo)
074              != null);
075
076      MetricsAssignmentManagerSource amSource =
077        MASTER.getAssignmentManager().getAssignmentManagerMetrics().getMetricsProcSource();
078      long ritDurationNumOps =
079        getMetricValue(snapshotMetrics(amSource), RIT_DURATION_NUM_OPS_METRIC);
080
081      ServerName current =
082        MASTER.getAssignmentManager().getRegionStates().getRegionServerOfRegion(regionInfo);
083      ServerName target = MASTER.getServerManager().getOnlineServersList().stream()
084        .filter(sn -> !sn.equals(current)).findFirst()
085        .orElseThrow(() -> new IllegalStateException("Need at least two regionservers"));
086
087      TEST_UTIL.getAdmin().move(regionInfo.getEncodedNameAsBytes(), target);
088      TEST_UTIL.waitFor(WAIT_TIMEOUT_MS, () -> target
089        .equals(MASTER.getAssignmentManager().getRegionStates().getRegionServerOfRegion(regionInfo))
090        && !AssignmentTestingUtil.isRegionInTransition(regionInfo, MASTER.getAssignmentManager()));
091
092      // num_ops is cumulative (never reset on snapshot); an increase proves the histogram is now
093      // fed on RIT completion. Use >= not ==: background RIT may also add. _max is not asserted --
094      // snapshot() resets it on every read, racing the metrics2 sampler.
095      long ritDurationNumOpsAfter =
096        getMetricValue(snapshotMetrics(amSource), RIT_DURATION_NUM_OPS_METRIC);
097      assertTrue(ritDurationNumOpsAfter >= ritDurationNumOps + 1,
098        "RitDuration histogram num_ops should increase after a region transition");
099    }
100  }
101
102  private MetricsRecord snapshotMetrics(MetricsAssignmentManagerSource amSource) {
103    MetricsCollectorImpl collector = new MetricsCollectorImpl();
104    assertInstanceOf(MetricsSource.class, amSource,
105      "MetricsAssignmentManagerSource should also implement MetricsSource");
106    ((MetricsSource) amSource).getMetrics(collector, true);
107    assertEquals(1, collector.getRecords().size());
108    return collector.getRecords().get(0);
109  }
110
111  private long getMetricValue(MetricsRecord record, String metricName) {
112    for (AbstractMetric metric : record.metrics()) {
113      if (metricName.equals(metric.name())) {
114        return metric.value().longValue();
115      }
116    }
117    throw new AssertionError("Metric not found: " + metricName);
118  }
119}