path: root/test/framework/src/main/java/org/elasticsearch/test/ESIntegTestCase.java

/*
 * Licensed to Elasticsearch under one or more contributor
 * license agreements. See the NOTICE file distributed with
 * this work for additional information regarding copyright
 * ownership. Elasticsearch licenses this file to you under
 * the Apache License, Version 2.0 (the "License"); you may
 * not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *    http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing,
 * software distributed under the License is distributed on an
 * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
 * KIND, either express or implied.  See the License for the
 * specific language governing permissions and limitations
 * under the License.
 */
package org.elasticsearch.test;

import com.carrotsearch.randomizedtesting.RandomizedContext;
import com.carrotsearch.randomizedtesting.RandomizedTest;
import com.carrotsearch.randomizedtesting.annotations.TestGroup;
import com.carrotsearch.randomizedtesting.generators.RandomInts;
import com.carrotsearch.randomizedtesting.generators.RandomPicks;
import org.apache.http.impl.client.HttpClients;
import org.apache.lucene.util.IOUtils;
import org.apache.lucene.util.LuceneTestCase;
import org.apache.lucene.util.TestUtil;
import org.elasticsearch.ElasticsearchException;
import org.elasticsearch.ExceptionsHelper;
import org.elasticsearch.action.ActionListener;
import org.elasticsearch.action.ShardOperationFailedException;
import org.elasticsearch.action.admin.cluster.health.ClusterHealthRequest;
import org.elasticsearch.action.admin.cluster.health.ClusterHealthResponse;
import org.elasticsearch.action.admin.cluster.node.info.NodeInfo;
import org.elasticsearch.action.admin.cluster.node.info.NodesInfoResponse;
import org.elasticsearch.action.admin.cluster.state.ClusterStateResponse;
import org.elasticsearch.action.admin.cluster.tasks.PendingClusterTasksResponse;
import org.elasticsearch.action.admin.indices.create.CreateIndexRequestBuilder;
import org.elasticsearch.action.admin.indices.exists.indices.IndicesExistsResponse;
import org.elasticsearch.action.admin.indices.flush.FlushResponse;
import org.elasticsearch.action.admin.indices.flush.SyncedFlushResponse;
import org.elasticsearch.action.admin.indices.forcemerge.ForceMergeResponse;
import org.elasticsearch.action.admin.indices.mapping.get.GetMappingsResponse;
import org.elasticsearch.action.admin.indices.refresh.RefreshResponse;
import org.elasticsearch.action.admin.indices.segments.IndicesSegmentResponse;
import org.elasticsearch.action.admin.indices.template.put.PutIndexTemplateRequestBuilder;
import org.elasticsearch.action.bulk.BulkRequestBuilder;
import org.elasticsearch.action.bulk.BulkResponse;
import org.elasticsearch.action.delete.DeleteResponse;
import org.elasticsearch.action.get.GetResponse;
import org.elasticsearch.action.index.IndexRequestBuilder;
import org.elasticsearch.action.index.IndexResponse;
import org.elasticsearch.action.search.ClearScrollResponse;
import org.elasticsearch.action.search.SearchResponse;
import org.elasticsearch.action.support.IndicesOptions;
import org.elasticsearch.client.AdminClient;
import org.elasticsearch.client.Client;
import org.elasticsearch.client.Requests;
import org.elasticsearch.cluster.ClusterService;
import org.elasticsearch.cluster.ClusterState;
import org.elasticsearch.cluster.health.ClusterHealthStatus;
import org.elasticsearch.cluster.metadata.IndexMetaData;
import org.elasticsearch.cluster.metadata.MappingMetaData;
import org.elasticsearch.cluster.metadata.MetaData;
import org.elasticsearch.cluster.routing.IndexRoutingTable;
import org.elasticsearch.cluster.routing.IndexShardRoutingTable;
import org.elasticsearch.cluster.routing.ShardRouting;
import org.elasticsearch.cluster.routing.UnassignedInfo;
import org.elasticsearch.cluster.routing.allocation.decider.DiskThresholdDecider;
import org.elasticsearch.cluster.routing.allocation.decider.EnableAllocationDecider;
import org.elasticsearch.common.Nullable;
import org.elasticsearch.common.Priority;
import org.elasticsearch.common.Strings;
import org.elasticsearch.common.collect.ImmutableOpenMap;
import org.elasticsearch.common.collect.Tuple;
import org.elasticsearch.common.network.NetworkAddress;
import org.elasticsearch.common.regex.Regex;
import org.elasticsearch.common.settings.Settings;
import org.elasticsearch.common.transport.InetSocketTransportAddress;
import org.elasticsearch.common.transport.TransportAddress;
import org.elasticsearch.common.unit.ByteSizeUnit;
import org.elasticsearch.common.unit.ByteSizeValue;
import org.elasticsearch.common.unit.TimeValue;
import org.elasticsearch.common.util.concurrent.EsRejectedExecutionException;
import org.elasticsearch.common.xcontent.ToXContent;
import org.elasticsearch.common.xcontent.XContentBuilder;
import org.elasticsearch.common.xcontent.XContentFactory;
import org.elasticsearch.common.xcontent.XContentHelper;
import org.elasticsearch.common.xcontent.json.JsonXContent;
import org.elasticsearch.common.xcontent.support.XContentMapValues;
import org.elasticsearch.discovery.Discovery;
import org.elasticsearch.discovery.zen.ZenDiscovery;
import org.elasticsearch.discovery.zen.elect.ElectMasterService;
import org.elasticsearch.env.Environment;
import org.elasticsearch.index.IndexService;
import org.elasticsearch.index.MockEngineFactoryPlugin;
import org.elasticsearch.index.codec.CodecService;
import org.elasticsearch.index.mapper.DocumentMapper;
import org.elasticsearch.index.mapper.MappedFieldType;
import org.elasticsearch.index.mapper.MappedFieldType.Loading;
import org.elasticsearch.index.mapper.internal.TimestampFieldMapper;
import org.elasticsearch.index.shard.IndexShard;
import org.elasticsearch.index.shard.MergePolicyConfig;
import org.elasticsearch.index.shard.MergeSchedulerConfig;
import org.elasticsearch.index.translog.Translog;
import org.elasticsearch.index.translog.TranslogConfig;
import org.elasticsearch.index.translog.TranslogWriter;
import org.elasticsearch.indices.IndicesService;
import org.elasticsearch.indices.cache.request.IndicesRequestCache;
import org.elasticsearch.indices.flush.SyncedFlushService;
import org.elasticsearch.indices.store.IndicesStore;
import org.elasticsearch.node.Node;
import org.elasticsearch.node.NodeMocksPlugin;
import org.elasticsearch.plugins.Plugin;
import org.elasticsearch.rest.RestStatus;
import org.elasticsearch.search.MockSearchService;
import org.elasticsearch.search.SearchHit;
import org.elasticsearch.search.SearchService;
import org.elasticsearch.test.client.RandomizingClient;
import org.elasticsearch.test.disruption.ServiceDisruptionScheme;
import org.elasticsearch.test.rest.client.http.HttpRequestBuilder;
import org.elasticsearch.test.store.MockFSIndexStore;
import org.elasticsearch.test.transport.AssertingLocalTransport;
import org.elasticsearch.test.transport.MockTransportService;
import org.hamcrest.Matchers;
import org.joda.time.DateTimeZone;
import org.junit.After;
import org.junit.AfterClass;
import org.junit.Before;
import org.junit.BeforeClass;

import java.io.IOException;
import java.io.InputStream;
import java.lang.annotation.Annotation;
import java.lang.annotation.Documented;
import java.lang.annotation.ElementType;
import java.lang.annotation.Inherited;
import java.lang.annotation.Retention;
import java.lang.annotation.RetentionPolicy;
import java.lang.annotation.Target;
import java.net.InetAddress;
import java.net.InetSocketAddress;
import java.net.URL;
import java.nio.file.DirectoryStream;
import java.nio.file.Files;
import java.nio.file.Path;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collection;
import java.util.Collections;
import java.util.HashSet;
import java.util.IdentityHashMap;
import java.util.List;
import java.util.Map;
import java.util.Random;
import java.util.Set;
import java.util.concurrent.Callable;
import java.util.concurrent.CopyOnWriteArrayList;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutionException;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicInteger;
import java.util.concurrent.atomic.AtomicLong;
import java.util.function.BooleanSupplier;

import static org.elasticsearch.client.Requests.syncedFlushRequest;
import static org.elasticsearch.cluster.metadata.IndexMetaData.SETTING_NUMBER_OF_REPLICAS;
import static org.elasticsearch.cluster.metadata.IndexMetaData.SETTING_NUMBER_OF_SHARDS;
import static org.elasticsearch.common.settings.Settings.settingsBuilder;
import static org.elasticsearch.common.util.CollectionUtils.eagerPartition;
import static org.elasticsearch.index.query.QueryBuilders.matchAllQuery;
import static org.elasticsearch.test.XContentTestUtils.convertToMap;
import static org.elasticsearch.test.XContentTestUtils.differenceBetweenMapsIgnoringArrayOrder;
import static org.elasticsearch.test.hamcrest.ElasticsearchAssertions.assertAcked;
import static org.elasticsearch.test.hamcrest.ElasticsearchAssertions.assertNoFailures;
import static org.elasticsearch.test.hamcrest.ElasticsearchAssertions.assertNoTimeout;
import static org.hamcrest.Matchers.emptyArray;
import static org.hamcrest.Matchers.emptyIterable;
import static org.hamcrest.Matchers.equalTo;
import static org.hamcrest.Matchers.is;
import static org.hamcrest.Matchers.notNullValue;
import static org.hamcrest.Matchers.startsWith;

/**
 * {@link ESIntegTestCase} is an abstract base class to run integration
 * tests against a JVM private Elasticsearch Cluster. The test class supports 2 different
 * cluster scopes.
 * <ul>
 * <li>{@link Scope#TEST} - uses a new cluster for each individual test method.</li>
 * <li>{@link Scope#SUITE} - uses a cluster shared across all test methods in the same suite</li>
 * </ul>
 * <p>
 * The most common test scope is {@link Scope#SUITE} which shares a cluster per test suite.
 * <p>
 * If the test methods need specific node settings or change persistent and/or transient cluster settings {@link Scope#TEST}
 * should be used. To configure a scope for the test cluster the {@link ClusterScope} annotation
 * should be used, here is an example:
 * <pre>
 *
 * {@literal @}ClusterScope(scope=Scope.TEST) public class SomeIT extends ESIntegTestCase {
 * public void testMethod() {}
 * }
 * </pre>
 * <p>
 * If no {@link ClusterScope} annotation is present on an integration test the default scope is {@link Scope#SUITE}
 * <p>
 * A test cluster creates a set of nodes in the background before the test starts. The number of nodes in the cluster is
 * determined at random and can change across tests. The {@link ClusterScope} allows configuring the initial number of nodes
 * that are created before the tests start.
 *  <pre>
 * {@literal @}ClusterScope(scope=Scope.SUITE, numDataNodes=3)
 * public class SomeIT extends ESIntegTestCase {
 * public void testMethod() {}
 * }
 * </pre>
 * <p>
 * Note, the {@link ESIntegTestCase} uses randomized settings on a cluster and index level. For instance
 * each test might use different directory implementation for each test or will return a random client to one of the
 * nodes in the cluster for each call to {@link #client()}. Test failures might only be reproducible if the correct
 * system properties are passed to the test execution environment.
 * <p>
 * This class supports the following system properties (passed with -Dkey=value to the application)
 * <ul>
 * <li>-D{@value #TESTS_CLIENT_RATIO} - a double value in the interval [0..1] which defines the ration between node and transport clients used</li>
 * <li>-D{@value #TESTS_ENABLE_MOCK_MODULES} - a boolean value to enable or disable mock modules. This is
 * useful to test the system without asserting modules that to make sure they don't hide any bugs in production.</li>
 * <li> - a random seed used to initialize the index random context.
 * </ul>
 */
@LuceneTestCase.SuppressFileSystems("ExtrasFS") // doesn't work with potential multi data path from test cluster yet
public abstract class ESIntegTestCase extends ESTestCase {

    /**
     * Property that controls whether ThirdParty Integration tests are run (not the default).
     */
    public static final String SYSPROP_THIRDPARTY = "tests.thirdparty";

    /**
     * Annotation for third-party integration tests.
     * <p>
     * These are tests the require a third-party service in order to run. They
     * may require the user to manually configure an external process (such as rabbitmq),
     * or may additionally require some external configuration (e.g. AWS credentials)
     * via the {@code tests.config} system property.
     */
    @Inherited
    @Retention(RetentionPolicy.RUNTIME)
    @Target(ElementType.TYPE)
    @TestGroup(enabled = false, sysProperty = ESIntegTestCase.SYSPROP_THIRDPARTY)
    public @interface ThirdParty {
    }

    /** node names of the corresponding clusters will start with these prefixes */
    public static final String SUITE_CLUSTER_NODE_PREFIX = "node_s";
    public static final String TEST_CLUSTER_NODE_PREFIX = "node_t";

    /**
     * Key used to set the transport client ratio via the commandline -D{@value #TESTS_CLIENT_RATIO}
     */
    public static final String TESTS_CLIENT_RATIO = "tests.client.ratio";

    /**
     * Key used to eventually switch to using an external cluster and provide its transport addresses
     */
    public static final String TESTS_CLUSTER = "tests.cluster";

    /**
     * Key used to retrieve the index random seed from the index settings on a running node.
     * The value of this seed can be used to initialize a random context for a specific index.
     * It's set once per test via a generic index template.
     */
    public static final String SETTING_INDEX_SEED = "index.tests.seed";

    /**
     * A boolean value to enable or disable mock modules. This is useful to test the
     * system without asserting modules that to make sure they don't hide any bugs in
     * production.
     *
     * @see ESIntegTestCase
     */
    public static final String TESTS_ENABLE_MOCK_MODULES = "tests.enable_mock_modules";

    /**
     * Threshold at which indexing switches from frequently async to frequently bulk.
     */
    private static final int FREQUENT_BULK_THRESHOLD = 300;

    /**
     * Threshold at which bulk indexing will always be used.
     */
    private static final int ALWAYS_BULK_THRESHOLD = 3000;

    /**
     * Maximum number of async operations that indexRandom will kick off at one time.
     */
    private static final int MAX_IN_FLIGHT_ASYNC_INDEXES = 150;

    /**
     * Maximum number of documents in a single bulk index request.
     */
    private static final int MAX_BULK_INDEX_REQUEST_SIZE = 1000;

    /**
     * Default minimum number of shards for an index
     */
    protected static final int DEFAULT_MIN_NUM_SHARDS = 1;

    /**
     * Default maximum number of shards for an index
     */
    protected static final int DEFAULT_MAX_NUM_SHARDS = 10;

    /**
     * The current cluster depending on the configured {@link Scope}.
     * By default if no {@link ClusterScope} is configured this will hold a reference to the suite cluster.
     */
    private static TestCluster currentCluster;

    private static final double TRANSPORT_CLIENT_RATIO = transportClientRatio();

    private static final Map<Class<?>, TestCluster> clusters = new IdentityHashMap<>();

    private static ESIntegTestCase INSTANCE = null; // see @SuiteScope
    private static Long SUITE_SEED = null;

    @BeforeClass
    public static void beforeClass() throws Exception {
        SUITE_SEED = randomLong();
        initializeSuiteScope();
    }

    protected final void beforeInternal() throws Exception {
        final Scope currentClusterScope = getCurrentClusterScope();
        switch (currentClusterScope) {
            case SUITE:
                assert SUITE_SEED != null : "Suite seed was not initialized";
                currentCluster = buildAndPutCluster(currentClusterScope, SUITE_SEED);
                break;
            case TEST:
                currentCluster = buildAndPutCluster(currentClusterScope, randomLong());
                break;
            default:
                fail("Unknown Scope: [" + currentClusterScope + "]");
        }
        cluster().beforeTest(getRandom(), getPerTestTransportClientRatio());
        cluster().wipe(excludeTemplates());
        randomIndexTemplate();
    }

    private void printTestMessage(String message) {
        if (isSuiteScopedTest(getClass()) && (getTestName().equals("<unknown>"))) {
            logger.info("[{}]: {} suite", getTestClass().getSimpleName(), message);
        } else {
            logger.info("[{}#{}]: {} test", getTestClass().getSimpleName(), getTestName(), message);
        }
    }

    private Loading randomLoadingValues() {
        return randomFrom(Loading.values());
    }

    /**
     * Creates a randomized index template. This template is used to pass in randomized settings on a
     * per index basis. Allows to enable/disable the randomization for number of shards and replicas
     */
    public void randomIndexTemplate() throws IOException {

        // TODO move settings for random directory etc here into the index based randomized settings.
        if (cluster().size() > 0) {
            Settings.Builder randomSettingsBuilder =
                    setRandomIndexSettings(getRandom(), Settings.builder())
                            .put(SETTING_INDEX_SEED, getRandom().nextLong());

            randomSettingsBuilder.put(SETTING_NUMBER_OF_SHARDS, numberOfShards())
                    .put(SETTING_NUMBER_OF_REPLICAS, numberOfReplicas());

            // if the test class is annotated with SuppressCodecs("*"), it means don't use lucene's codec randomization
            // otherwise, use it, it has assertions and so on that can find bugs.
            SuppressCodecs annotation = getClass().getAnnotation(SuppressCodecs.class);
            if (annotation != null && annotation.value().length == 1 && "*".equals(annotation.value()[0])) {
                randomSettingsBuilder.put("index.codec", randomFrom(CodecService.DEFAULT_CODEC, CodecService.BEST_COMPRESSION_CODEC));
            } else {
                randomSettingsBuilder.put("index.codec", CodecService.LUCENE_DEFAULT_CODEC);
            }
            XContentBuilder mappings = null;
            if (frequently() && randomDynamicTemplates()) {
                mappings = XContentFactory.jsonBuilder().startObject().startObject("_default_");
                if (randomBoolean()) {
                    mappings.startObject(TimestampFieldMapper.NAME)
                            .field("enabled", randomBoolean());
                    mappings.endObject();
                }
                mappings.startArray("dynamic_templates")
                        .startObject()
                        .startObject("template-strings")
                        .field("match_mapping_type", "string")
                        .startObject("mapping")
                        .startObject("fielddata")
                        .field(Loading.KEY, randomLoadingValues())
                        .endObject()
                        .endObject()
                        .endObject()
                        .endObject()
                        .startObject()
                        .startObject("template-longs")
                        .field("match_mapping_type", "long")
                        .startObject("mapping")
                        .startObject("fielddata")
                        .field(Loading.KEY, randomFrom(Loading.LAZY, Loading.EAGER))
                        .endObject()
                        .endObject()
                        .endObject()
                        .endObject()
                        .startObject()
                        .startObject("template-doubles")
                        .field("match_mapping_type", "double")
                        .startObject("mapping")
                        .startObject("fielddata")
                        .field(Loading.KEY, randomFrom(Loading.LAZY, Loading.EAGER))
                        .endObject()
                        .endObject()
                        .endObject()
                        .endObject()
                        .startObject()
                        .startObject("template-geo_points")
                        .field("match_mapping_type", "geo_point")
                        .startObject("mapping")
                        .startObject("fielddata")
                        .field(Loading.KEY, randomFrom(Loading.LAZY, Loading.EAGER))
                        .endObject()
                        .endObject()
                        .endObject()
                        .endObject()
                        .startObject()
                        .startObject("template-booleans")
                        .field("match_mapping_type", "boolean")
                        .startObject("mapping")
                        .startObject("fielddata")
                        .field(Loading.KEY, randomFrom(Loading.LAZY, Loading.EAGER))
                        .endObject()
                        .endObject()
                        .endObject()
                        .endObject()
                        .endArray();
                mappings.endObject().endObject();
            }

            for (String setting : randomSettingsBuilder.internalMap().keySet()) {
                assertThat("non index. prefix setting set on index template, its a node setting...", setting, startsWith("index."));
            }

            PutIndexTemplateRequestBuilder putTemplate = client().admin().indices()
                    .preparePutTemplate("random_index_template")
                    .setTemplate("*")
                    .setOrder(0)
                    .setSettings(randomSettingsBuilder);
            if (mappings != null) {
                logger.info("test using _default_ mappings: [{}]", mappings.bytes().toUtf8());
                putTemplate.addMapping("_default_", mappings);
            }
            assertAcked(putTemplate.execute().actionGet());
        }
    }

    protected Settings.Builder setRandomIndexSettings(Random random, Settings.Builder builder) {
        setRandomIndexMergeSettings(random, builder);
        setRandomIndexTranslogSettings(random, builder);
        setRandomIndexNormsLoading(random, builder);

        if (random.nextBoolean()) {
            builder.put(MergeSchedulerConfig.AUTO_THROTTLE, false);
        }

        if (random.nextBoolean()) {
            builder.put(IndicesRequestCache.INDEX_CACHE_REQUEST_ENABLED, random.nextBoolean());
        }

        if (random.nextBoolean()) {
            builder.put("index.shard.check_on_startup", randomFrom(random, "false", "checksum", "true"));
        }

        if (randomBoolean()) {
            // keep this low so we don't stall tests
            builder.put(UnassignedInfo.INDEX_DELAYED_NODE_LEFT_TIMEOUT_SETTING, RandomInts.randomIntBetween(random, 1, 15) + "ms");
        }

        return builder;
    }

    private static Settings.Builder setRandomIndexMergeSettings(Random random, Settings.Builder builder) {
        if (random.nextBoolean()) {
            builder.put(MergePolicyConfig.INDEX_COMPOUND_FORMAT,
                    random.nextBoolean() ? random.nextDouble() : random.nextBoolean());
        }
        switch (random.nextInt(4)) {
            case 3:
                final int maxThreadCount = RandomInts.randomIntBetween(random, 1, 4);
                final int maxMergeCount = RandomInts.randomIntBetween(random, maxThreadCount, maxThreadCount + 4);
                builder.put(MergeSchedulerConfig.MAX_MERGE_COUNT, maxMergeCount);
                builder.put(MergeSchedulerConfig.MAX_THREAD_COUNT, maxThreadCount);
                break;
        }

        return builder;
    }

    private static Settings.Builder setRandomIndexNormsLoading(Random random, Settings.Builder builder) {
        if (random.nextBoolean()) {
            builder.put(SearchService.NORMS_LOADING_KEY, RandomPicks.randomFrom(random, Arrays.asList(MappedFieldType.Loading.EAGER, MappedFieldType.Loading.LAZY)));
        }
        return builder;
    }

    private static Settings.Builder setRandomIndexTranslogSettings(Random random, Settings.Builder builder) {
        if (random.nextBoolean()) {
            builder.put(IndexShard.INDEX_TRANSLOG_FLUSH_THRESHOLD_OPS, RandomInts.randomIntBetween(random, 1, 10000));
        }
        if (random.nextBoolean()) {
            builder.put(IndexShard.INDEX_TRANSLOG_FLUSH_THRESHOLD_SIZE, new ByteSizeValue(RandomInts.randomIntBetween(random, 1, 300), ByteSizeUnit.MB));
        }
        if (random.nextBoolean()) {
            builder.put(IndexShard.INDEX_TRANSLOG_DISABLE_FLUSH, random.nextBoolean());
        }
        if (random.nextBoolean()) {
            builder.put(TranslogConfig.INDEX_TRANSLOG_DURABILITY, RandomPicks.randomFrom(random, Translog.Durabilty.values()));
        }

        if (random.nextBoolean()) {
            builder.put(TranslogConfig.INDEX_TRANSLOG_FS_TYPE, RandomPicks.randomFrom(random, TranslogWriter.Type.values()));
            if (rarely(random)) {
                builder.put(TranslogConfig.INDEX_TRANSLOG_SYNC_INTERVAL, 0); // 0 has special meaning to sync each op
            } else {
                builder.put(TranslogConfig.INDEX_TRANSLOG_SYNC_INTERVAL, RandomInts.randomIntBetween(random, 100, 5000), TimeUnit.MILLISECONDS);
            }
        }

        return builder;
    }

    private TestCluster buildWithPrivateContext(final Scope scope, final long seed) throws Exception {
        return RandomizedContext.current().runWithPrivateRandomness(new com.carrotsearch.randomizedtesting.Randomness(seed), new Callable<TestCluster>() {
            @Override
            public TestCluster call() throws Exception {
                return buildTestCluster(scope, seed);
            }
        });
    }

    private TestCluster buildAndPutCluster(Scope currentClusterScope, long seed) throws Exception {
        final Class<?> clazz = this.getClass();
        TestCluster testCluster = clusters.remove(clazz); // remove this cluster first
        clearClusters(); // all leftovers are gone by now... this is really just a double safety if we miss something somewhere
        switch (currentClusterScope) {
            case SUITE:
                if (testCluster == null) { // only build if it's not there yet
                    testCluster = buildWithPrivateContext(currentClusterScope, seed);
                }
                break;
            case TEST:
                // close the previous one and create a new one
                IOUtils.closeWhileHandlingException(testCluster);
                testCluster = buildTestCluster(currentClusterScope, seed);
                break;
        }
        clusters.put(clazz, testCluster);
        return testCluster;
    }

    private static void clearClusters() throws IOException {
        if (!clusters.isEmpty()) {
            IOUtils.close(clusters.values());
            clusters.clear();
        }
    }

    protected final void afterInternal(boolean afterClass) throws Exception {
        boolean success = false;
        try {
            final Scope currentClusterScope = getCurrentClusterScope();
            clearDisruptionScheme();
            try {
                if (cluster() != null) {
                    if (currentClusterScope != Scope.TEST) {
                        MetaData metaData = client().admin().cluster().prepareState().execute().actionGet().getState().getMetaData();
                        assertThat("test leaves persistent cluster metadata behind: " + metaData.persistentSettings().getAsMap(), metaData
                                .persistentSettings().getAsMap().size(), equalTo(0));
                        assertThat("test leaves transient cluster metadata behind: " + metaData.transientSettings().getAsMap(), metaData
                                .transientSettings().getAsMap().size(), equalTo(0));
                    }
                    ensureClusterSizeConsistency();
                    ensureClusterStateConsistency();
                    if (isInternalCluster()) {
                        // check no pending cluster states are leaked
                        for (Discovery discovery : internalCluster().getInstances(Discovery.class)) {
                            if (discovery instanceof ZenDiscovery) {
                                final ZenDiscovery zenDiscovery = (ZenDiscovery) discovery;
                                assertBusy(new Runnable() {
                                    @Override
                                    public void run() {
                                        assertThat("still having pending states: " + Strings.arrayToDelimitedString(zenDiscovery.pendingClusterStates(), "\n"),
                                                zenDiscovery.pendingClusterStates(), emptyArray());
                                    }
                                });
                            }
                        }
                    }
                    beforeIndexDeletion();
                    cluster().wipe(excludeTemplates()); // wipe after to make sure we fail in the test that didn't ack the delete
                    if (afterClass || currentClusterScope == Scope.TEST) {
                        cluster().close();
                    }
                    cluster().assertAfterTest();
                }
            } finally {
                if (currentClusterScope == Scope.TEST) {
                    clearClusters(); // it is ok to leave persistent / transient cluster state behind if scope is TEST
                }
            }
            success = true;
        } finally {
            if (!success) {
                // if we failed here that means that something broke horribly so we should clear all clusters
                // TODO: just let the exception happen, WTF is all this horseshit
                // afterTestRule.forceFailure();
            }
        }
    }

    /**
     * @return An exclude set of index templates that will not be removed in between tests.
     */
    protected Set<String> excludeTemplates() {
        return Collections.emptySet();
    }

    protected void beforeIndexDeletion() {
        cluster().beforeIndexDeletion();
    }

    public static TestCluster cluster() {
        return currentCluster;
    }

    public static boolean isInternalCluster() {
        return (currentCluster instanceof InternalTestCluster);
    }

    public static InternalTestCluster internalCluster() {
        if (!isInternalCluster()) {
            throw new UnsupportedOperationException("current test cluster is immutable");
        }
        return (InternalTestCluster) currentCluster;
    }

    public ClusterService clusterService() {
        return internalCluster().clusterService();
    }

    public static Client client() {
        return client(null);
    }

    public static Client client(@Nullable String node) {
        if (node != null) {
            return internalCluster().client(node);
        }
        Client client = cluster().client();
        if (frequently()) {
            client = new RandomizingClient(client, getRandom());
        }
        return client;
    }

    public static Client dataNodeClient() {
        Client client = internalCluster().dataNodeClient();
        if (frequently()) {
            client = new RandomizingClient(client, getRandom());
        }
        return client;
    }

    public static Iterable<Client> clients() {
        return cluster();
    }

    protected int minimumNumberOfShards() {
        return DEFAULT_MIN_NUM_SHARDS;
    }

    protected int maximumNumberOfShards() {
        return DEFAULT_MAX_NUM_SHARDS;
    }

    protected int numberOfShards() {
        return between(minimumNumberOfShards(), maximumNumberOfShards());
    }

    protected int minimumNumberOfReplicas() {
        return 0;
    }

    protected int maximumNumberOfReplicas() {
        //use either 0 or 1 replica, yet a higher amount when possible, but only rarely
        int maxNumReplicas = Math.max(0, cluster().numDataNodes() - 1);
        return frequently() ? Math.min(1, maxNumReplicas) : maxNumReplicas;
    }

    protected int numberOfReplicas() {
        return between(minimumNumberOfReplicas(), maximumNumberOfReplicas());
    }


    public void setDisruptionScheme(ServiceDisruptionScheme scheme) {
        internalCluster().setDisruptionScheme(scheme);
    }

    public void clearDisruptionScheme() {
        if (isInternalCluster()) {
            internalCluster().clearDisruptionScheme();
        }
    }

    /**
     * Returns a settings object used in {@link #createIndex(String...)} and {@link #prepareCreate(String)} and friends.
     * This method can be overwritten by subclasses to set defaults for the indices that are created by the test.
     * By default it returns a settings object that sets a random number of shards. Number of shards and replicas
     * can be controlled through specific methods.
     */
    public Settings indexSettings() {
        Settings.Builder builder = Settings.builder();
        int numberOfShards = numberOfShards();
        if (numberOfShards > 0) {
            builder.put(SETTING_NUMBER_OF_SHARDS, numberOfShards).build();
        }
        int numberOfReplicas = numberOfReplicas();
        if (numberOfReplicas >= 0) {
            builder.put(SETTING_NUMBER_OF_REPLICAS, numberOfReplicas).build();
        }
        // 30% of the time
        if (randomInt(9) < 3) {
            final String dataPath = randomAsciiOfLength(10);
            logger.info("using custom data_path for index: [{}]", dataPath);
            builder.put(IndexMetaData.SETTING_DATA_PATH, dataPath);
        }
        return builder.build();
    }

    /**
     * Creates one or more indices and asserts that the indices are acknowledged. If one of the indices
     * already exists this method will fail and wipe all the indices created so far.
     */
    public final void createIndex(String... names) {

        List<String> created = new ArrayList<>();
        for (String name : names) {
            boolean success = false;
            try {
                assertAcked(prepareCreate(name));
                created.add(name);
                success = true;
            } finally {
                if (!success && !created.isEmpty()) {
                    cluster().wipeIndices(created.toArray(new String[created.size()]));
                }
            }
        }
    }

    /**
     * Creates a new {@link CreateIndexRequestBuilder} with the settings obtained from {@link #indexSettings()}.
     */
    public final CreateIndexRequestBuilder prepareCreate(String index) {
        return client().admin().indices().prepareCreate(index).setSettings(indexSettings());
    }

    /**
     * Creates a new {@link CreateIndexRequestBuilder} with the settings obtained from {@link #indexSettings()}.
     * The index that is created with this builder will only be allowed to allocate on the number of nodes passed to this
     * method.
     * <p>
     * This method uses allocation deciders to filter out certain nodes to allocate the created index on. It defines allocation
     * rules based on <code>index.routing.allocation.exclude._name</code>.
     * </p>
     */
    public final CreateIndexRequestBuilder prepareCreate(String index, int numNodes) {
        return prepareCreate(index, numNodes, Settings.builder());
    }

    /**
     * Creates a new {@link CreateIndexRequestBuilder} with the settings obtained from {@link #indexSettings()}.
     * The index that is created with this builder will only be allowed to allocate on the number of nodes passed to this
     * method.
     * <p>
     * This method uses allocation deciders to filter out certain nodes to allocate the created index on. It defines allocation
     * rules based on <code>index.routing.allocation.exclude._name</code>.
     * </p>
     */
    public CreateIndexRequestBuilder prepareCreate(String index, int numNodes, Settings.Builder settingsBuilder) {
        internalCluster().ensureAtLeastNumDataNodes(numNodes);

        Settings.Builder builder = Settings.builder().put(indexSettings()).put(settingsBuilder.build());

        if (numNodes > 0) {
            getExcludeSettings(index, numNodes, builder);
        }
        return client().admin().indices().prepareCreate(index).setSettings(builder.build());
    }

    private Settings.Builder getExcludeSettings(String index, int num, Settings.Builder builder) {
        String exclude = String.join(",", internalCluster().allDataNodesButN(num));
        builder.put("index.routing.allocation.exclude._name", exclude);
        return builder;
    }

    /**
     * Waits until all nodes have no pending tasks.
     */
    public void waitNoPendingTasksOnAll() throws Exception {
        assertNoTimeout(client().admin().cluster().prepareHealth().setWaitForEvents(Priority.LANGUID).get());
        assertBusy(new Runnable() {
            @Override
            public void run() {
                for (Client client : clients()) {
                    ClusterHealthResponse clusterHealth = client.admin().cluster().prepareHealth().setLocal(true).get();
                    assertThat("client " + client + " still has in flight fetch", clusterHealth.getNumberOfInFlightFetch(), equalTo(0));
                    PendingClusterTasksResponse pendingTasks = client.admin().cluster().preparePendingClusterTasks().setLocal(true).get();
                    assertThat("client " + client + " still has pending tasks " + pendingTasks.prettyPrint(), pendingTasks, Matchers.emptyIterable());
                    clusterHealth = client.admin().cluster().prepareHealth().setLocal(true).get();
                    assertThat("client " + client + " still has in flight fetch", clusterHealth.getNumberOfInFlightFetch(), equalTo(0));
                }
            }
        });
        assertNoTimeout(client().admin().cluster().prepareHealth().setWaitForEvents(Priority.LANGUID).get());
    }

    /**
     * Waits till a (pattern) field name mappings concretely exists on all nodes. Note, this waits for the current
     * started shards and checks for concrete mappings.
     */
    public void assertConcreteMappingsOnAll(final String index, final String type, final String... fieldNames) throws Exception {
        Set<String> nodes = internalCluster().nodesInclude(index);
        assertThat(nodes, Matchers.not(Matchers.emptyIterable()));
        for (String node : nodes) {
            IndicesService indicesService = internalCluster().getInstance(IndicesService.class, node);
            IndexService indexService = indicesService.indexService(index);
            assertThat("index service doesn't exists on " + node, indexService, notNullValue());
            DocumentMapper documentMapper = indexService.mapperService().documentMapper(type);
            assertThat("document mapper doesn't exists on " + node, documentMapper, notNullValue());
            for (String fieldName : fieldNames) {
                Collection<String> matches = documentMapper.mappers().simpleMatchToFullName(fieldName);
                assertThat("field " + fieldName + " doesn't exists on " + node, matches, Matchers.not(emptyIterable()));
            }
        }
        assertMappingOnMaster(index, type, fieldNames);
    }

    /**
     * Waits for the given mapping type to exists on the master node.
     */
    public void assertMappingOnMaster(final String index, final String type, final String... fieldNames) throws Exception {
        GetMappingsResponse response = client().admin().indices().prepareGetMappings(index).setTypes(type).get();
        ImmutableOpenMap<String, MappingMetaData> mappings = response.getMappings().get(index);
        assertThat(mappings, notNullValue());
        MappingMetaData mappingMetaData = mappings.get(type);
        assertThat(mappingMetaData, notNullValue());

        Map<String, Object> mappingSource = mappingMetaData.getSourceAsMap();
        assertFalse(mappingSource.isEmpty());
        assertTrue(mappingSource.containsKey("properties"));

        for (String fieldName : fieldNames) {
            Map<String, Object> mappingProperties = (Map<String, Object>) mappingSource.get("properties");
            if (fieldName.indexOf('.') != -1) {
                fieldName = fieldName.replace(".", ".properties.");
            }
            assertThat("field " + fieldName + " doesn't exists in mapping " + mappingMetaData.source().string(), XContentMapValues.extractValue(fieldName, mappingProperties), notNullValue());
        }
    }

    /** Ensures the result counts are as expected, and logs the results if different */
    public void assertResultsAndLogOnFailure(long expectedResults, SearchResponse searchResponse) {
        if (searchResponse.getHits().getTotalHits() != expectedResults) {
            StringBuilder sb = new StringBuilder("search result contains [");
            sb.append(searchResponse.getHits().getTotalHits()).append("] results. expected [").append(expectedResults).append("]");
            String failMsg = sb.toString();
            for (SearchHit hit : searchResponse.getHits().getHits()) {
                sb.append("\n-> _index: [").append(hit.getIndex()).append("] type [").append(hit.getType())
                        .append("] id [").append(hit.id()).append("]");
            }
            logger.warn(sb.toString());
            fail(failMsg);
        }
    }

    /**
     * Restricts the given index to be allocated on <code>n</code> nodes using the allocation deciders.
     * Yet if the shards can't be allocated on any other node shards for this index will remain allocated on
     * more than <code>n</code> nodes.
     */
    public void allowNodes(String index, int n) {
        assert index != null;
        internalCluster().ensureAtLeastNumDataNodes(n);
        Settings.Builder builder = Settings.builder();
        if (n > 0) {
            getExcludeSettings(index, n, builder);
        }
        Settings build = builder.build();
        if (!build.getAsMap().isEmpty()) {
            logger.debug("allowNodes: updating [{}]'s setting to [{}]", index, build.toDelimitedString(';'));
            client().admin().indices().prepareUpdateSettings(index).setSettings(build).execute().actionGet();
        }
    }

    /**
     * Ensures the cluster has a green state via the cluster health API. This method will also wait for relocations.
     * It is useful to ensure that all action on the cluster have finished and all shards that were currently relocating
     * are now allocated and started.
     */
    public ClusterHealthStatus ensureGreen(String... indices) {
        return ensureGreen(TimeValue.timeValueSeconds(30), indices);
    }

    /**
     * Ensures the cluster has a green state via the cluster health API. This method will also wait for relocations.
     * It is useful to ensure that all action on the cluster have finished and all shards that were currently relocating
     * are now allocated and started.
     *
     * @param timeout time out value to set on {@link org.elasticsearch.action.admin.cluster.health.ClusterHealthRequest}
     */
    public ClusterHealthStatus ensureGreen(TimeValue timeout, String... indices) {
        ClusterHealthResponse actionGet = client().admin().cluster()
                .health(Requests.clusterHealthRequest(indices).timeout(timeout).waitForGreenStatus().waitForEvents(Priority.LANGUID).waitForRelocatingShards(0)).actionGet();
        if (actionGet.isTimedOut()) {
            logger.info("ensureGreen timed out, cluster state:\n{}\n{}", client().admin().cluster().prepareState().get().getState().prettyPrint(), client().admin().cluster().preparePendingClusterTasks().get().prettyPrint());
            fail("timed out waiting for green state");
        }
        assertThat(actionGet.getStatus(), equalTo(ClusterHealthStatus.GREEN));
        logger.debug("indices {} are green", indices.length == 0 ? "[_all]" : indices);
        return actionGet.getStatus();
    }

    /**
     * Waits for all relocating shards to become active using the cluster health API.
     */
    public ClusterHealthStatus waitForRelocation() {
        return waitForRelocation(null);
    }

    /**
     * Waits for all relocating shards to become active and the cluster has reached the given health status
     * using the cluster health API.
     */
    public ClusterHealthStatus waitForRelocation(ClusterHealthStatus status) {
        ClusterHealthRequest request = Requests.clusterHealthRequest().waitForRelocatingShards(0);
        if (status != null) {
            request.waitForStatus(status);
        }
        ClusterHealthResponse actionGet = client().admin().cluster()
                .health(request).actionGet();
        if (actionGet.isTimedOut()) {
            logger.info("waitForRelocation timed out (status={}), cluster state:\n{}\n{}", status, client().admin().cluster().prepareState().get().getState().prettyPrint(), client().admin().cluster().preparePendingClusterTasks().get().prettyPrint());
            assertThat("timed out waiting for relocation", actionGet.isTimedOut(), equalTo(false));
        }
        if (status != null) {
            assertThat(actionGet.getStatus(), equalTo(status));
        }
        return actionGet.getStatus();
    }

    /**
     * Waits until at least a give number of document is visible for searchers
     *
     * @param numDocs number of documents to wait for.
     * @return the actual number of docs seen.
     */
    public long waitForDocs(final long numDocs) throws InterruptedException {
        return waitForDocs(numDocs, null);
    }

    /**
     * Waits until at least a give number of document is visible for searchers
     *
     * @param numDocs number of documents to wait for
     * @param indexer a {@link org.elasticsearch.test.BackgroundIndexer}. If supplied it will be first checked for documents indexed.
     *                This saves on unneeded searches.
     * @return the actual number of docs seen.
     */
    public long waitForDocs(final long numDocs, final @Nullable BackgroundIndexer indexer) throws InterruptedException {
        // indexing threads can wait for up to ~1m before retrying when they first try to index into a shard which is not STARTED.
        return waitForDocs(numDocs, 90, TimeUnit.SECONDS, indexer);
    }

    /**
     * Waits until at least a give number of document is visible for searchers
     *
     * @param numDocs         number of documents to wait for
     * @param maxWaitTime     if not progress have been made during this time, fail the test
     * @param maxWaitTimeUnit the unit in which maxWaitTime is specified
     * @param indexer         a {@link org.elasticsearch.test.BackgroundIndexer}. If supplied it will be first checked for documents indexed.
     *                        This saves on unneeded searches.
     * @return the actual number of docs seen.
     */
    public long waitForDocs(final long numDocs, int maxWaitTime, TimeUnit maxWaitTimeUnit, final @Nullable BackgroundIndexer indexer)
            throws InterruptedException {
        final AtomicLong lastKnownCount = new AtomicLong(-1);
        long lastStartCount = -1;
        BooleanSupplier testDocs = () -> {
            if (indexer != null) {
                lastKnownCount.set(indexer.totalIndexedDocs());
            }
            if (lastKnownCount.get() >= numDocs) {
                try {
                    long count = client().prepareSearch().setSize(0).setQuery(matchAllQuery()).execute().actionGet().getHits().totalHits();
                    if (count == lastKnownCount.get()) {
                        // no progress - try to refresh for the next time
                        client().admin().indices().prepareRefresh().get();
                    }
                    lastKnownCount.set(count);
                } catch (Throwable e) { // count now acts like search and barfs if all shards failed...
                    logger.debug("failed to executed count", e);
                    return false;
                }
                logger.debug("[{}] docs visible for search. waiting for [{}]", lastKnownCount.get(), numDocs);
            } else {
                logger.debug("[{}] docs indexed. waiting for [{}]", lastKnownCount.get(), numDocs);
            }
            return lastKnownCount.get() >= numDocs;
        };

        while (!awaitBusy(testDocs, maxWaitTime, maxWaitTimeUnit)) {
            if (lastStartCount == lastKnownCount.get()) {
                // we didn't make any progress
                fail("failed to reach " + numDocs + "docs");
            }
            lastStartCount = lastKnownCount.get();
        }
        return lastKnownCount.get();
    }


    /**
     * Sets the cluster's minimum master node and make sure the response is acknowledge.
     * Note: this doesn't guarantee that the new setting has taken effect, just that it has been received by all nodes.
     */
    public void setMinimumMasterNodes(int n) {
        assertTrue(client().admin().cluster().prepareUpdateSettings().setTransientSettings(
                settingsBuilder().put(ElectMasterService.DISCOVERY_ZEN_MINIMUM_MASTER_NODES_SETTING.getKey(), n))
                .get().isAcknowledged());
    }

    /**
     * Ensures the cluster has a yellow state via the cluster health API.
     */
    public ClusterHealthStatus ensureYellow(String... indices) {
        ClusterHealthResponse actionGet = client().admin().cluster()
                .health(Requests.clusterHealthRequest(indices).waitForRelocatingShards(0).waitForYellowStatus().waitForEvents(Priority.LANGUID)).actionGet();
        if (actionGet.isTimedOut()) {
            logger.info("ensureYellow timed out, cluster state:\n{}\n{}", client().admin().cluster().prepareState().get().getState().prettyPrint(), client().admin().cluster().preparePendingClusterTasks().get().prettyPrint());
            assertThat("timed out waiting for yellow", actionGet.isTimedOut(), equalTo(false));
        }
        logger.debug("indices {} are yellow", indices.length == 0 ? "[_all]" : indices);
        return actionGet.getStatus();
    }

    /**
     * Prints the current cluster state as debug logging.
     */
    public void logClusterState() {
        logger.debug("cluster state:\n{}\n{}", client().admin().cluster().prepareState().get().getState().prettyPrint(), client().admin().cluster().preparePendingClusterTasks().get().prettyPrint());
    }

    /**
     * Prints the segments info for the given indices as debug logging.
     */
    public void logSegmentsState(String... indices) throws Exception {
        IndicesSegmentResponse segsRsp = client().admin().indices().prepareSegments(indices).get();
        logger.debug("segments {} state: \n{}", indices.length == 0 ? "[_all]" : indices,
                segsRsp.toXContent(JsonXContent.contentBuilder().prettyPrint(), ToXContent.EMPTY_PARAMS).string());
    }

    /**
     * Prints current memory stats as info logging.
     */
    public void logMemoryStats() {
        logger.info("memory: {}", XContentHelper.toString(client().admin().cluster().prepareNodesStats().clear().setJvm(true).get()));
    }

    void ensureClusterSizeConsistency() {
        if (cluster() != null) { // if static init fails the cluster can be null
            logger.trace("Check consistency for [{}] nodes", cluster().size());
            assertNoTimeout(client().admin().cluster().prepareHealth().setWaitForNodes(Integer.toString(cluster().size())).get());
        }
    }

    /**
     * Verifies that all nodes that have the same version of the cluster state as master have same cluster state
     */
    protected void ensureClusterStateConsistency() throws IOException {
        if (cluster() != null) {
            ClusterState masterClusterState = client().admin().cluster().prepareState().all().get().getState();
            byte[] masterClusterStateBytes = ClusterState.Builder.toBytes(masterClusterState);
            // remove local node reference
            masterClusterState = ClusterState.Builder.fromBytes(masterClusterStateBytes, null);
            Map<String, Object> masterStateMap = convertToMap(masterClusterState);
            int masterClusterStateSize = masterClusterState.toString().length();
            String masterId = masterClusterState.nodes().masterNodeId();
            for (Client client : cluster()) {
                ClusterState localClusterState = client.admin().cluster().prepareState().all().setLocal(true).get().getState();
                byte[] localClusterStateBytes = ClusterState.Builder.toBytes(localClusterState);
                // remove local node reference
                localClusterState = ClusterState.Builder.fromBytes(localClusterStateBytes, null);
                final Map<String, Object> localStateMap = convertToMap(localClusterState);
                final int localClusterStateSize = localClusterState.toString().length();
                // Check that the non-master node has the same version of the cluster state as the master and
                // that the master node matches the master (otherwise there is no requirement for the cluster state to match)
                if (masterClusterState.version() == localClusterState.version() && masterId.equals(localClusterState.nodes().masterNodeId())) {
                    try {
                        assertEquals("clusterstate UUID does not match", masterClusterState.stateUUID(), localClusterState.stateUUID());
                        // We cannot compare serialization bytes since serialization order of maps is not guaranteed
                        // but we can compare serialization sizes - they should be the same
                        assertEquals("clusterstate size does not match", masterClusterStateSize, localClusterStateSize);
                        // Compare JSON serialization
                        assertNull("clusterstate JSON serialization does not match", differenceBetweenMapsIgnoringArrayOrder(masterStateMap, localStateMap));
                    } catch (AssertionError error) {
                        logger.error("Cluster state from master:\n{}\nLocal cluster state:\n{}", masterClusterState.toString(), localClusterState.toString());
                        throw error;
                    }
                }
            }
        }

    }

    /**
     * Ensures the cluster is in a searchable state for the given indices. This means a searchable copy of each
     * shard is available on the cluster.
     */
    protected ClusterHealthStatus ensureSearchable(String... indices) {
        // this is just a temporary thing but it's easier to change if it is encapsulated.
        return ensureGreen(indices);
    }

    protected void ensureStableCluster(int nodeCount) {
        ensureStableCluster(nodeCount, TimeValue.timeValueSeconds(30));
    }

    protected void ensureStableCluster(int nodeCount, TimeValue timeValue) {
        ensureStableCluster(nodeCount, timeValue, false, null);
    }

    protected void ensureStableCluster(int nodeCount, @Nullable String viaNode) {
        ensureStableCluster(nodeCount, TimeValue.timeValueSeconds(30), false, viaNode);
    }

    protected void ensureStableCluster(int nodeCount, TimeValue timeValue, boolean local, @Nullable String viaNode) {
        if (viaNode == null) {
            viaNode = randomFrom(internalCluster().getNodeNames());
        }
        logger.debug("ensuring cluster is stable with [{}] nodes. access node: [{}]. timeout: [{}]", nodeCount, viaNode, timeValue);
        ClusterHealthResponse clusterHealthResponse = client(viaNode).admin().cluster().prepareHealth()
                .setWaitForEvents(Priority.LANGUID)
                .setWaitForNodes(Integer.toString(nodeCount))
                .setTimeout(timeValue)
                .setLocal(local)
                .setWaitForRelocatingShards(0)
                .get();
        if (clusterHealthResponse.isTimedOut()) {
            ClusterStateResponse stateResponse = client(viaNode).admin().cluster().prepareState().get();
            fail("failed to reach a stable cluster of [" + nodeCount + "] nodes. Tried via [" + viaNode + "]. last cluster state:\n"
                    + stateResponse.getState().prettyPrint());
        }
        assertThat(clusterHealthResponse.isTimedOut(), is(false));
    }

    /**
     * Syntactic sugar for:
     * <pre>
     *   client().prepareIndex(index, type).setSource(source).execute().actionGet();
     * </pre>
     */
    protected final IndexResponse index(String index, String type, XContentBuilder source) {
        return client().prepareIndex(index, type).setSource(source).execute().actionGet();
    }

    /**
     * Syntactic sugar for:
     * <pre>
     *   client().prepareIndex(index, type).setSource(source).execute().actionGet();
     * </pre>
     */
    protected final IndexResponse index(String index, String type, String id, Map<String, Object> source) {
        return client().prepareIndex(index, type, id).setSource(source).execute().actionGet();
    }

    /**
     * Syntactic sugar for:
     * <pre>
     *   client().prepareGet(index, type, id).execute().actionGet();
     * </pre>
     */
    protected final GetResponse get(String index, String type, String id) {
        return client().prepareGet(index, type, id).execute().actionGet();
    }

    /**
     * Syntactic sugar for:
     * <pre>
     *   return client().prepareIndex(index, type, id).setSource(source).execute().actionGet();
     * </pre>
     */
    protected final IndexResponse index(String index, String type, String id, XContentBuilder source) {
        return client().prepareIndex(index, type, id).setSource(source).execute().actionGet();
    }

    /**
     * Syntactic sugar for:
     * <pre>
     *   return client().prepareIndex(index, type, id).setSource(source).execute().actionGet();
     * </pre>
     */
    protected final IndexResponse index(String index, String type, String id, Object... source) {
        return client().prepareIndex(index, type, id).setSource(source).execute().actionGet();
    }

    /**
     * Syntactic sugar for:
     * <pre>
     *   return client().prepareIndex(index, type, id).setSource(source).execute().actionGet();
     * </pre>
     * <p>
     * where source is a String.
     */
    protected final IndexResponse index(String index, String type, String id, String source) {
        return client().prepareIndex(index, type, id).setSource(source).execute().actionGet();
    }

    /**
     * Waits for relocations and refreshes all indices in the cluster.
     *
     * @see #waitForRelocation()
     */
    protected final RefreshResponse refresh() {
        waitForRelocation();
        // TODO RANDOMIZE with flush?
        RefreshResponse actionGet = client().admin().indices().prepareRefresh().execute().actionGet();
        assertNoFailures(actionGet);
        return actionGet;
    }

    /**
     * Flushes and refreshes all indices in the cluster
     */
    protected final void flushAndRefresh(String... indices) {
        flush(indices);
        refresh();
    }

    /**
     * Flush some or all indices in the cluster.
     */
    protected final FlushResponse flush(String... indices) {
        waitForRelocation();
        FlushResponse actionGet = client().admin().indices().prepareFlush(indices).setWaitIfOngoing(true).execute().actionGet();
        for (ShardOperationFailedException failure : actionGet.getShardFailures()) {
            assertThat("unexpected flush failure " + failure.reason(), failure.status(), equalTo(RestStatus.SERVICE_UNAVAILABLE));
        }
        return actionGet;
    }

    /**
     * Waits for all relocations and force merge all indices in the cluster to 1 segment.
     */
    protected ForceMergeResponse forceMerge() {
        waitForRelocation();
        ForceMergeResponse actionGet = client().admin().indices().prepareForceMerge().setMaxNumSegments(1).execute().actionGet();
        assertNoFailures(actionGet);
        return actionGet;
    }

    /**
     * Returns <code>true</code> iff the given index exists otherwise <code>false</code>
     */
    protected boolean indexExists(String index) {
        IndicesExistsResponse actionGet = client().admin().indices().prepareExists(index).execute().actionGet();
        return actionGet.isExists();
    }

    /**
     * Syntactic sugar for enabling allocation for <code>indices</code>
     */
    protected final void enableAllocation(String... indices) {
        client().admin().indices().prepareUpdateSettings(indices).setSettings(Settings.builder().put(
                EnableAllocationDecider.INDEX_ROUTING_ALLOCATION_ENABLE, "all"
        )).get();
    }

    /**
     * Syntactic sugar for disabling allocation for <code>indices</code>
     */
    protected final void disableAllocation(String... indices) {
        client().admin().indices().prepareUpdateSettings(indices).setSettings(Settings.builder().put(
                EnableAllocationDecider.INDEX_ROUTING_ALLOCATION_ENABLE, "none"
        )).get();
    }

    /**
     * Returns a random admin client. This client can either be a node or a transport client pointing to any of
     * the nodes in the cluster.
     */
    protected AdminClient admin() {
        return client().admin();
    }

    /**
     * Convenience method that forwards to {@link #indexRandom(boolean, List)}.
     */
    public void indexRandom(boolean forceRefresh, IndexRequestBuilder... builders) throws InterruptedException, ExecutionException {
        indexRandom(forceRefresh, Arrays.asList(builders));
    }

    public void indexRandom(boolean forceRefresh, boolean dummyDocuments, IndexRequestBuilder... builders) throws InterruptedException, ExecutionException {
        indexRandom(forceRefresh, dummyDocuments, Arrays.asList(builders));
    }


    private static final String RANDOM_BOGUS_TYPE = "RANDOM_BOGUS_TYPE______";

    /**
     * Indexes the given {@link IndexRequestBuilder} instances randomly. It shuffles the given builders and either
     * indexes them in a blocking or async fashion. This is very useful to catch problems that relate to internal document
     * ids or index segment creations. Some features might have bug when a given document is the first or the last in a
     * segment or if only one document is in a segment etc. This method prevents issues like this by randomizing the index
     * layout.
     *
     * @param forceRefresh if <tt>true</tt> all involved indices are refreshed once the documents are indexed. Additionally if <tt>true</tt>
     *                     some empty dummy documents are may be randomly inserted into the document list and deleted once all documents are indexed.
     *                     This is useful to produce deleted documents on the server side.
     * @param builders     the documents to index.
     * @see #indexRandom(boolean, boolean, java.util.List)
     */
    public void indexRandom(boolean forceRefresh, List<IndexRequestBuilder> builders) throws InterruptedException, ExecutionException {
        indexRandom(forceRefresh, forceRefresh, builders);
    }

    /**
     * Indexes the given {@link IndexRequestBuilder} instances randomly. It shuffles the given builders and either
     * indexes they in a blocking or async fashion. This is very useful to catch problems that relate to internal document
     * ids or index segment creations. Some features might have bug when a given document is the first or the last in a
     * segment or if only one document is in a segment etc. This method prevents issues like this by randomizing the index
     * layout.
     *
     * @param forceRefresh   if <tt>true</tt> all involved indices are refreshed once the documents are indexed.
     * @param dummyDocuments if <tt>true</tt> some empty dummy documents may be randomly inserted into the document list and deleted once
     *                       all documents are indexed. This is useful to produce deleted documents on the server side.
     * @param builders       the documents to index.
     */
    public void indexRandom(boolean forceRefresh, boolean dummyDocuments, List<IndexRequestBuilder> builders) throws InterruptedException, ExecutionException {
        indexRandom(forceRefresh, dummyDocuments, true, builders);
    }

    /**
     * Indexes the given {@link IndexRequestBuilder} instances randomly. It shuffles the given builders and either
     * indexes they in a blocking or async fashion. This is very useful to catch problems that relate to internal document
     * ids or index segment creations. Some features might have bug when a given document is the first or the last in a
     * segment or if only one document is in a segment etc. This method prevents issues like this by randomizing the index
     * layout.
     *
     * @param forceRefresh   if <tt>true</tt> all involved indices are refreshed once the documents are indexed.
     * @param dummyDocuments if <tt>true</tt> some empty dummy documents may be randomly inserted into the document list and deleted once
     *                       all documents are indexed. This is useful to produce deleted documents on the server side.
     * @param maybeFlush     if <tt>true</tt> this method may randomly execute full flushes after index operations.
     * @param builders       the documents to index.
     */
    public void indexRandom(boolean forceRefresh, boolean dummyDocuments, boolean maybeFlush, List<IndexRequestBuilder> builders) throws InterruptedException, ExecutionException {

        Random random = getRandom();
        Set<String> indicesSet = new HashSet<>();
        for (IndexRequestBuilder builder : builders) {
            indicesSet.add(builder.request().index());
        }
        Set<Tuple<String, String>> bogusIds = new HashSet<>();
        if (random.nextBoolean() && !builders.isEmpty() && dummyDocuments) {
            builders = new ArrayList<>(builders);
            final String[] indices = indicesSet.toArray(new String[indicesSet.size()]);
            // inject some bogus docs
            final int numBogusDocs = scaledRandomIntBetween(1, builders.size() * 2);
            final int unicodeLen = between(1, 10);
            for (int i = 0; i < numBogusDocs; i++) {
                String id = randomRealisticUnicodeOfLength(unicodeLen) + Integer.toString(dummmyDocIdGenerator.incrementAndGet());
                String index = RandomPicks.randomFrom(random, indices);
                bogusIds.add(new Tuple<>(index, id));
                builders.add(client().prepareIndex(index, RANDOM_BOGUS_TYPE, id).setSource("{}"));
            }
        }
        final String[] indices = indicesSet.toArray(new String[indicesSet.size()]);
        Collections.shuffle(builders, random());
        final CopyOnWriteArrayList<Tuple<IndexRequestBuilder, Throwable>> errors = new CopyOnWriteArrayList<>();
        List<CountDownLatch> inFlightAsyncOperations = new ArrayList<>();
        // If you are indexing just a few documents then frequently do it one at a time.  If many then frequently in bulk.
        if (builders.size() < FREQUENT_BULK_THRESHOLD ? frequently() : builders.size() < ALWAYS_BULK_THRESHOLD ? rarely() : false) {
            if (frequently()) {
                logger.info("Index [{}] docs async: [{}] bulk: [{}]", builders.size(), true, false);
                for (IndexRequestBuilder indexRequestBuilder : builders) {
                    indexRequestBuilder.execute(new PayloadLatchedActionListener<IndexResponse, IndexRequestBuilder>(indexRequestBuilder, newLatch(inFlightAsyncOperations), errors));
                    postIndexAsyncActions(indices, inFlightAsyncOperations, maybeFlush);
                }
            } else {
                logger.info("Index [{}] docs async: [{}] bulk: [{}]", builders.size(), false, false);
                for (IndexRequestBuilder indexRequestBuilder : builders) {
                    indexRequestBuilder.execute().actionGet();
                    postIndexAsyncActions(indices, inFlightAsyncOperations, maybeFlush);
                }
            }
        } else {
            List<List<IndexRequestBuilder>> partition = eagerPartition(builders, Math.min(MAX_BULK_INDEX_REQUEST_SIZE,
                    Math.max(1, (int) (builders.size() * randomDouble()))));
            logger.info("Index [{}] docs async: [{}] bulk: [{}] partitions [{}]", builders.size(), false, true, partition.size());
            for (List<IndexRequestBuilder> segmented : partition) {
                BulkRequestBuilder bulkBuilder = client().prepareBulk();
                for (IndexRequestBuilder indexRequestBuilder : segmented) {
                    bulkBuilder.add(indexRequestBuilder);
                }
                BulkResponse actionGet = bulkBuilder.execute().actionGet();
                assertThat(actionGet.hasFailures() ? actionGet.buildFailureMessage() : "", actionGet.hasFailures(), equalTo(false));
            }
        }
        for (CountDownLatch operation : inFlightAsyncOperations) {
            operation.await();
        }
        final List<Throwable> actualErrors = new ArrayList<>();
        for (Tuple<IndexRequestBuilder, Throwable> tuple : errors) {
            if (ExceptionsHelper.unwrapCause(tuple.v2()) instanceof EsRejectedExecutionException) {
                tuple.v1().execute().actionGet(); // re-index if rejected
            } else {
                actualErrors.add(tuple.v2());
            }
        }
        assertThat(actualErrors, emptyIterable());
        if (!bogusIds.isEmpty()) {
            // delete the bogus types again - it might trigger merges or at least holes in the segments and enforces deleted docs!
            for (Tuple<String, String> doc : bogusIds) {
                // see https://github.com/elasticsearch/elasticsearch/issues/8706
                final DeleteResponse deleteResponse = client().prepareDelete(doc.v1(), RANDOM_BOGUS_TYPE, doc.v2()).get();
                if (deleteResponse.isFound() == false) {
                    logger.warn("failed to delete a dummy doc [{}][{}]", doc.v1(), doc.v2());
                }
            }
        }
        if (forceRefresh) {
            assertNoFailures(client().admin().indices().prepareRefresh(indices).setIndicesOptions(IndicesOptions.lenientExpandOpen()).execute().get());
        }
    }

    private AtomicInteger dummmyDocIdGenerator = new AtomicInteger();

    /** Disables translog flushing for the specified index */
    public static void disableTranslogFlush(String index) {
        Settings settings = Settings.builder().put(IndexShard.INDEX_TRANSLOG_DISABLE_FLUSH, true).build();
        client().admin().indices().prepareUpdateSettings(index).setSettings(settings).get();
    }

    /** Enables translog flushing for the specified index */
    public static void enableTranslogFlush(String index) {
        Settings settings = Settings.builder().put(IndexShard.INDEX_TRANSLOG_DISABLE_FLUSH, false).build();
        client().admin().indices().prepareUpdateSettings(index).setSettings(settings).get();
    }

    /** Disables an index block for the specified index */
    public static void disableIndexBlock(String index, String block) {
        Settings settings = Settings.builder().put(block, false).build();
        client().admin().indices().prepareUpdateSettings(index).setSettings(settings).get();
    }

    /** Enables an index block for the specified index */
    public static void enableIndexBlock(String index, String block) {
        Settings settings = Settings.builder().put(block, true).build();
        client().admin().indices().prepareUpdateSettings(index).setSettings(settings).get();
    }

    /** Sets or unsets the cluster read_only mode **/
    public static void setClusterReadOnly(boolean value) {
        Settings settings = settingsBuilder().put(MetaData.SETTING_READ_ONLY_SETTING.getKey(), value).build();
        assertAcked(client().admin().cluster().prepareUpdateSettings().setTransientSettings(settings).get());
    }

    private static CountDownLatch newLatch(List<CountDownLatch> latches) {
        CountDownLatch l = new CountDownLatch(1);
        latches.add(l);
        return l;
    }

    /**
     * Maybe refresh, force merge, or flush then always make sure there aren't too many in flight async operations.
     */
    private void postIndexAsyncActions(String[] indices, List<CountDownLatch> inFlightAsyncOperations, boolean maybeFlush) throws InterruptedException {
        if (rarely()) {
            if (rarely()) {
                client().admin().indices().prepareRefresh(indices).setIndicesOptions(IndicesOptions.lenientExpandOpen()).execute(
                        new LatchedActionListener<>(newLatch(inFlightAsyncOperations)));
            } else if (maybeFlush && rarely()) {
                if (randomBoolean()) {
                    client().admin().indices().prepareFlush(indices).setIndicesOptions(IndicesOptions.lenientExpandOpen()).execute(
                            new LatchedActionListener<>(newLatch(inFlightAsyncOperations)));
                } else {
                    client().admin().indices().syncedFlush(syncedFlushRequest(indices).indicesOptions(IndicesOptions.lenientExpandOpen()),
                        new LatchedActionListener<>(newLatch(inFlightAsyncOperations)));
                }
            } else if (rarely()) {
                client().admin().indices().prepareForceMerge(indices).setIndicesOptions(IndicesOptions.lenientExpandOpen()).setMaxNumSegments(between(1, 10)).setFlush(maybeFlush && randomBoolean()).execute(
                        new LatchedActionListener<>(newLatch(inFlightAsyncOperations)));
            }
        }
        while (inFlightAsyncOperations.size() > MAX_IN_FLIGHT_ASYNC_INDEXES) {
            int waitFor = between(0, inFlightAsyncOperations.size() - 1);
            inFlightAsyncOperations.remove(waitFor).await();
        }
    }

    /**
     * The scope of a test cluster used together with
     * {@link ESIntegTestCase.ClusterScope} annotations on {@link ESIntegTestCase} subclasses.
     */
    public enum Scope {
        /**
         * A cluster shared across all method in a single test suite
         */
        SUITE,
        /**
         * A test exclusive test cluster
         */
        TEST
    }

    /**
     * Defines a cluster scope for a {@link ESIntegTestCase} subclass.
     * By default if no {@link ClusterScope} annotation is present {@link ESIntegTestCase.Scope#SUITE} is used
     * together with randomly chosen settings like number of nodes etc.
     */
    @Retention(RetentionPolicy.RUNTIME)
    @Target({ElementType.TYPE})
    public @interface ClusterScope {
        /**
         * Returns the scope. {@link ESIntegTestCase.Scope#SUITE} is default.
         */
        Scope scope() default Scope.SUITE;

        /**
         * Returns the number of nodes in the cluster. Default is <tt>-1</tt> which means
         * a random number of nodes is used, where the minimum and maximum number of nodes
         * are either the specified ones or the default ones if not specified.
         */
        int numDataNodes() default -1;

        /**
         * Returns the minimum number of nodes in the cluster. Default is <tt>-1</tt>.
         * Ignored when {@link ClusterScope#numDataNodes()} is set.
         */
        int minNumDataNodes() default -1;

        /**
         * Returns the maximum number of nodes in the cluster.  Default is <tt>-1</tt>.
         * Ignored when {@link ClusterScope#numDataNodes()} is set.
         */
        int maxNumDataNodes() default -1;

        /**
         * Returns the number of client nodes in the cluster. Default is {@link InternalTestCluster#DEFAULT_NUM_CLIENT_NODES}, a
         * negative value means that the number of client nodes will be randomized.
         */
        int numClientNodes() default InternalTestCluster.DEFAULT_NUM_CLIENT_NODES;

        /**
         * Returns the transport client ratio. By default this returns <code>-1</code> which means a random
         * ratio in the interval <code>[0..1]</code> is used.
         */
        double transportClientRatio() default -1;

        /**
         * Return whether or not to enable dynamic templates for the mappings.
         */
        boolean randomDynamicTemplates() default true;
    }

    private class LatchedActionListener<Response> implements ActionListener<Response> {
        private final CountDownLatch latch;

        public LatchedActionListener(CountDownLatch latch) {
            this.latch = latch;
        }

        @Override
        public final void onResponse(Response response) {
            latch.countDown();
        }

        @Override
        public final void onFailure(Throwable t) {
            try {
                logger.info("Action Failed", t);
                addError(t);
            } finally {
                latch.countDown();
            }
        }

        protected void addError(Throwable t) {
        }

    }

    private class PayloadLatchedActionListener<Response, T> extends LatchedActionListener<Response> {
        private final CopyOnWriteArrayList<Tuple<T, Throwable>> errors;
        private final T builder;

        public PayloadLatchedActionListener(T builder, CountDownLatch latch, CopyOnWriteArrayList<Tuple<T, Throwable>> errors) {
            super(latch);
            this.errors = errors;
            this.builder = builder;
        }

        @Override
        protected void addError(Throwable t) {
            errors.add(new Tuple<>(builder, t));
        }

    }

    /**
     * Clears the given scroll Ids
     */
    public void clearScroll(String... scrollIds) {
        ClearScrollResponse clearResponse = client().prepareClearScroll()
                .setScrollIds(Arrays.asList(scrollIds)).get();
        assertThat(clearResponse.isSucceeded(), equalTo(true));
    }

    private static <A extends Annotation> A getAnnotation(Class<?> clazz, Class<A> annotationClass) {
        if (clazz == Object.class || clazz == ESIntegTestCase.class) {
            return null;
        }
        A annotation = clazz.getAnnotation(annotationClass);
        if (annotation != null) {
            return annotation;
        }
        return getAnnotation(clazz.getSuperclass(), annotationClass);
    }


    private Scope getCurrentClusterScope() {
        return getCurrentClusterScope(this.getClass());
    }

    private static Scope getCurrentClusterScope(Class<?> clazz) {
        ClusterScope annotation = getAnnotation(clazz, ClusterScope.class);
        // if we are not annotated assume suite!
        return annotation == null ? Scope.SUITE : annotation.scope();
    }

    private int getNumDataNodes() {
        ClusterScope annotation = getAnnotation(this.getClass(), ClusterScope.class);
        return annotation == null ? -1 : annotation.numDataNodes();
    }

    private int getMinNumDataNodes() {
        ClusterScope annotation = getAnnotation(this.getClass(), ClusterScope.class);
        return annotation == null || annotation.minNumDataNodes() == -1 ? InternalTestCluster.DEFAULT_MIN_NUM_DATA_NODES : annotation.minNumDataNodes();
    }

    private int getMaxNumDataNodes() {
        ClusterScope annotation = getAnnotation(this.getClass(), ClusterScope.class);
        return annotation == null || annotation.maxNumDataNodes() == -1 ? InternalTestCluster.DEFAULT_MAX_NUM_DATA_NODES : annotation.maxNumDataNodes();
    }

    private int getNumClientNodes() {
        ClusterScope annotation = getAnnotation(this.getClass(), ClusterScope.class);
        return annotation == null ? InternalTestCluster.DEFAULT_NUM_CLIENT_NODES : annotation.numClientNodes();
    }

    private boolean randomDynamicTemplates() {
        ClusterScope annotation = getAnnotation(this.getClass(), ClusterScope.class);
        return annotation == null || annotation.randomDynamicTemplates();
    }

    /**
     * This method is used to obtain settings for the <tt>Nth</tt> node in the cluster.
     * Nodes in this cluster are associated with an ordinal number such that nodes can
     * be started with specific configurations. This method might be called multiple
     * times with the same ordinal and is expected to return the same value for each invocation.
     * In other words subclasses must ensure this method is idempotent.
     */
    protected Settings nodeSettings(int nodeOrdinal) {
        Settings.Builder builder = settingsBuilder()
                // Default the watermarks to absurdly low to prevent the tests
                // from failing on nodes without enough disk space
                .put(DiskThresholdDecider.CLUSTER_ROUTING_ALLOCATION_LOW_DISK_WATERMARK_SETTING.getKey(), "1b")
                .put(DiskThresholdDecider.CLUSTER_ROUTING_ALLOCATION_HIGH_DISK_WATERMARK_SETTING.getKey(), "1b")
                .put("script.indexed", "on")
                .put("script.inline", "on")
                        // wait short time for other active shards before actually deleting, default 30s not needed in tests
                .put(IndicesStore.INDICES_STORE_DELETE_SHARD_TIMEOUT, new TimeValue(1, TimeUnit.SECONDS));
        return builder.build();
    }

    /**
     * Returns a collection of plugins that should be loaded on each node.
     */
    protected Collection<Class<? extends Plugin>> nodePlugins() {
        return Collections.emptyList();
    }

    /**
     * Returns a collection of plugins that should be loaded when creating a transport client.
     */
    protected Collection<Class<? extends Plugin>> transportClientPlugins() {
        return Collections.emptyList();
    }

    /** Helper method to create list of plugins without specifying generic types. */
    @SafeVarargs
    @SuppressWarnings("varargs") // due to type erasure, the varargs type is non-reifiable, which casues this warning
    protected final Collection<Class<? extends Plugin>> pluginList(Class<? extends Plugin>... plugins) {
        return Arrays.asList(plugins);
    }

    /**
     * This method is used to obtain additional settings for clients created by the internal cluster.
     * These settings will be applied on the client in addition to some randomized settings defined in
     * the cluster. These setttings will also override any other settings the internal cluster might
     * add by default.
     */
    protected Settings transportClientSettings() {
        return Settings.EMPTY;
    }

    private ExternalTestCluster buildExternalCluster(String clusterAddresses) throws IOException {
        String[] stringAddresses = clusterAddresses.split(",");
        TransportAddress[] transportAddresses = new TransportAddress[stringAddresses.length];
        int i = 0;
        for (String stringAddress : stringAddresses) {
            URL url = new URL("http://" + stringAddress);
            InetAddress inetAddress = InetAddress.getByName(url.getHost());
            transportAddresses[i++] = new InetSocketTransportAddress(new InetSocketAddress(inetAddress, url.getPort()));
        }
        return new ExternalTestCluster(createTempDir(), externalClusterClientSettings(), transportClientPlugins(), transportAddresses);
    }

    protected Settings externalClusterClientSettings() {
        return Settings.EMPTY;
    }

    protected TestCluster buildTestCluster(Scope scope, long seed) throws IOException {
        String clusterAddresses = System.getProperty(TESTS_CLUSTER);
        if (Strings.hasLength(clusterAddresses)) {
            if (scope == Scope.TEST) {
                throw new IllegalArgumentException("Cannot run TEST scope test with " + TESTS_CLUSTER);
            }
            return buildExternalCluster(clusterAddresses);
        }

        final String nodePrefix;
        switch (scope) {
            case TEST:
                nodePrefix = TEST_CLUSTER_NODE_PREFIX;
                break;
            case SUITE:
                nodePrefix = SUITE_CLUSTER_NODE_PREFIX;
                break;
            default:
                throw new ElasticsearchException("Scope not supported: " + scope);
        }
        NodeConfigurationSource nodeConfigurationSource = new NodeConfigurationSource() {
            @Override
            public Settings nodeSettings(int nodeOrdinal) {
                return Settings.builder().put(Node.HTTP_ENABLED, false).
                        put(ESIntegTestCase.this.nodeSettings(nodeOrdinal)).build();
            }

            @Override
            public Collection<Class<? extends Plugin>> nodePlugins() {
                return ESIntegTestCase.this.nodePlugins();
            }

            @Override
            public Settings transportClientSettings() {
                return ESIntegTestCase.this.transportClientSettings();
            }

            @Override
            public Collection<Class<? extends Plugin>> transportClientPlugins() {
                return ESIntegTestCase.this.transportClientPlugins();
            }
        };

        int numDataNodes = getNumDataNodes();
        int minNumDataNodes;
        int maxNumDataNodes;
        if (numDataNodes >= 0) {
            minNumDataNodes = maxNumDataNodes = numDataNodes;
        } else {
            minNumDataNodes = getMinNumDataNodes();
            maxNumDataNodes = getMaxNumDataNodes();
        }
        SuppressLocalMode noLocal = getAnnotation(this.getClass(), SuppressLocalMode.class);
        SuppressNetworkMode noNetwork = getAnnotation(this.getClass(), SuppressNetworkMode.class);
        String nodeMode = InternalTestCluster.configuredNodeMode();
        if (noLocal != null && noNetwork != null) {
            throw new IllegalStateException("Can't suppress both network and local mode");
        } else if (noLocal != null) {
            nodeMode = "network";
        } else if (noNetwork != null) {
            nodeMode = "local";
        }

        Collection<Class<? extends Plugin>> mockPlugins = getMockPlugins();

        return new InternalTestCluster(nodeMode, seed, createTempDir(), minNumDataNodes, maxNumDataNodes,
                InternalTestCluster.clusterName(scope.name(), seed) + "-cluster", nodeConfigurationSource, getNumClientNodes(),
                InternalTestCluster.DEFAULT_ENABLE_HTTP_PIPELINING, nodePrefix, mockPlugins);
    }

    /** Return the mock plugins the cluster should use. These may be randomly omitted based on the cluster seed. */
    protected Collection<Class<? extends Plugin>> getMockPlugins() {
        return pluginList(MockTransportService.TestPlugin.class,
                          MockFSIndexStore.TestPlugin.class,
                          NodeMocksPlugin.class,
                          MockEngineFactoryPlugin.class,
                          MockSearchService.TestPlugin.class,
                          AssertingLocalTransport.TestPlugin.class);
    }

    /**
     * Returns the client ratio configured via
     */
    private static double transportClientRatio() {
        String property = System.getProperty(TESTS_CLIENT_RATIO);
        if (property == null || property.isEmpty()) {
            return Double.NaN;
        }
        return Double.parseDouble(property);
    }

    /**
     * Returns the transport client ratio from the class level annotation or via
     * {@link System#getProperty(String)} if available. If both are not available this will
     * return a random ratio in the interval <tt>[0..1]</tt>
     */
    protected double getPerTestTransportClientRatio() {
        final ClusterScope annotation = getAnnotation(this.getClass(), ClusterScope.class);
        double perTestRatio = -1;
        if (annotation != null) {
            perTestRatio = annotation.transportClientRatio();
        }
        if (perTestRatio == -1) {
            return Double.isNaN(TRANSPORT_CLIENT_RATIO) ? randomDouble() : TRANSPORT_CLIENT_RATIO;
        }
        assert perTestRatio >= 0.0 && perTestRatio <= 1.0;
        return perTestRatio;
    }

    /**
     * Returns a random JODA Time Zone based on Java Time Zones
     */
    public static DateTimeZone randomDateTimeZone() {
        DateTimeZone timeZone;

        // It sounds like some Java Time Zones are unknown by JODA. For example: Asia/Riyadh88
        // We need to fallback in that case to a known time zone
        try {
            timeZone = DateTimeZone.forTimeZone(RandomizedTest.randomTimeZone());
        } catch (IllegalArgumentException e) {
            timeZone = DateTimeZone.forOffsetHours(randomIntBetween(-12, 12));
        }

        return timeZone;
    }

    /**
     * Returns path to a random directory that can be used to create a temporary file system repo
     */
    public Path randomRepoPath() {
        if (currentCluster instanceof InternalTestCluster) {
            return randomRepoPath(((InternalTestCluster) currentCluster).getDefaultSettings());
        } else if (currentCluster instanceof CompositeTestCluster) {
            return randomRepoPath(((CompositeTestCluster) currentCluster).internalCluster().getDefaultSettings());
        }
        throw new UnsupportedOperationException("unsupported cluster type");
    }

    /**
     * Returns path to a random directory that can be used to create a temporary file system repo
     */
    public static Path randomRepoPath(Settings settings) {
        Environment environment = new Environment(settings);
        Path[] repoFiles = environment.repoFiles();
        assert repoFiles.length > 0;
        Path path;
        do {
            path = repoFiles[0].resolve(randomAsciiOfLength(10));
        } while (Files.exists(path));
        return path;
    }

    protected NumShards getNumShards(String index) {
        MetaData metaData = client().admin().cluster().prepareState().get().getState().metaData();
        assertThat(metaData.hasIndex(index), equalTo(true));
        int numShards = Integer.valueOf(metaData.index(index).getSettings().get(SETTING_NUMBER_OF_SHARDS));
        int numReplicas = Integer.valueOf(metaData.index(index).getSettings().get(SETTING_NUMBER_OF_REPLICAS));
        return new NumShards(numShards, numReplicas);
    }

    /**
     * Asserts that all shards are allocated on nodes matching the given node pattern.
     */
    public Set<String> assertAllShardsOnNodes(String index, String... pattern) {
        Set<String> nodes = new HashSet<>();
        ClusterState clusterState = client().admin().cluster().prepareState().execute().actionGet().getState();
        for (IndexRoutingTable indexRoutingTable : clusterState.routingTable()) {
            for (IndexShardRoutingTable indexShardRoutingTable : indexRoutingTable) {
                for (ShardRouting shardRouting : indexShardRoutingTable) {
                    if (shardRouting.currentNodeId() != null && index.equals(shardRouting.getIndex())) {
                        String name = clusterState.nodes().get(shardRouting.currentNodeId()).name();
                        nodes.add(name);
                        assertThat("Allocated on new node: " + name, Regex.simpleMatch(pattern, name), is(true));
                    }
                }
            }
        }
        return nodes;
    }

    protected static class NumShards {
        public final int numPrimaries;
        public final int numReplicas;
        public final int totalNumShards;
        public final int dataCopies;

        private NumShards(int numPrimaries, int numReplicas) {
            this.numPrimaries = numPrimaries;
            this.numReplicas = numReplicas;
            this.dataCopies = numReplicas + 1;
            this.totalNumShards = numPrimaries * dataCopies;
        }
    }

    private static boolean runTestScopeLifecycle() {
        return INSTANCE == null;
    }


    @Before
    public final void before() throws Exception {

        if (runTestScopeLifecycle()) {
            printTestMessage("setup");
            beforeInternal();
        }
        printTestMessage("starting");
    }


    @After
    public final void after() throws Exception {
        printTestMessage("finished");
        // Deleting indices is going to clear search contexts implicitely so we
        // need to check that there are no more in-flight search contexts before
        // we remove indices
        super.ensureAllSearchContextsReleased();
        if (runTestScopeLifecycle()) {
            printTestMessage("cleaning up after");
            afterInternal(false);
            printTestMessage("cleaned up after");
        }
    }

    @AfterClass
    public static void afterClass() throws Exception {
        if (!runTestScopeLifecycle()) {
            try {
                INSTANCE.printTestMessage("cleaning up after");
                INSTANCE.afterInternal(true);
            } finally {
                INSTANCE = null;
            }
        } else {
            clearClusters();
        }
        SUITE_SEED = null;
        currentCluster = null;
    }

    private static void initializeSuiteScope() throws Exception {
        Class<?> targetClass = getTestClass();
        /**
         * Note we create these test class instance via reflection
         * since JUnit creates a new instance per test and that is also
         * the reason why INSTANCE is static since this entire method
         * must be executed in a static context.
         */
        assert INSTANCE == null;
        if (isSuiteScopedTest(targetClass)) {
            // note we need to do this this way to make sure this is reproducible
            INSTANCE = (ESIntegTestCase) targetClass.newInstance();
            boolean success = false;
            try {
                INSTANCE.printTestMessage("setup");
                INSTANCE.beforeInternal();
                INSTANCE.setupSuiteScopeCluster();
                success = true;
            } finally {
                if (!success) {
                    afterClass();
                }
            }
        } else {
            INSTANCE = null;
        }
    }

    /**
     * Compute a routing key that will route documents to the <code>shard</code>-th shard
     * of the provided index.
     */
    protected String routingKeyForShard(String index, String type, int shard) {
        return internalCluster().routingKeyForShard(index, type, shard, getRandom());
    }

    /**
     * Return settings that could be used to start a node that has the given zipped home directory.
     */
    protected Settings prepareBackwardsDataDir(Path backwardsIndex, Object... settings) throws IOException {
        Path indexDir = createTempDir();
        Path dataDir = indexDir.resolve("data");
        try (InputStream stream = Files.newInputStream(backwardsIndex)) {
            TestUtil.unzip(stream, indexDir);
        }
        assertTrue(Files.exists(dataDir));

        // list clusters in the datapath, ignoring anything from extrasfs
        final Path[] list;
        try (DirectoryStream<Path> stream = Files.newDirectoryStream(dataDir)) {
            List<Path> dirs = new ArrayList<>();
            for (Path p : stream) {
                if (!p.getFileName().toString().startsWith("extra")) {
                    dirs.add(p);
                }
            }
            list = dirs.toArray(new Path[0]);
        }

        if (list.length != 1) {
            StringBuilder builder = new StringBuilder("Backwards index must contain exactly one cluster\n");
            for (Path line : list) {
                builder.append(line.toString()).append('\n');
            }
            throw new IllegalStateException(builder.toString());
        }
        Path src = list[0];
        Path dest = dataDir.resolve(internalCluster().getClusterName());
        assertTrue(Files.exists(src));
        Files.move(src, dest);
        assertFalse(Files.exists(src));
        assertTrue(Files.exists(dest));
        Settings.Builder builder = Settings.builder()
                .put(settings)
                .put("path.data", dataDir.toAbsolutePath());

        Path configDir = indexDir.resolve("config");
        if (Files.exists(configDir)) {
            builder.put("path.conf", configDir.toAbsolutePath());
        }
        return builder.build();
    }

    protected HttpRequestBuilder httpClient() {
        final NodesInfoResponse nodeInfos = client().admin().cluster().prepareNodesInfo().get();
        final NodeInfo[] nodes = nodeInfos.getNodes();
        assertTrue(nodes.length > 0);
        TransportAddress publishAddress = randomFrom(nodes).getHttp().address().publishAddress();
        assertEquals(1, publishAddress.uniqueAddressTypeId());
        InetSocketAddress address = ((InetSocketTransportAddress) publishAddress).address();
        return new HttpRequestBuilder(HttpClients.createDefault()).host(NetworkAddress.formatAddress(address.getAddress())).port(address.getPort());
    }

    /**
     * This method is executed iff the test is annotated with {@link SuiteScopeTestCase}
     * before the first test of this class is executed.
     *
     * @see SuiteScopeTestCase
     */
    protected void setupSuiteScopeCluster() throws Exception {
    }

    private static boolean isSuiteScopedTest(Class<?> clazz) {
        return clazz.getAnnotation(SuiteScopeTestCase.class) != null;
    }

    /**
     * If a test is annotated with {@link SuiteScopeTestCase}
     * the checks and modifications that are applied to the used test cluster are only done after all tests
     * of this class are executed. This also has the side-effect of a suite level setup method {@link #setupSuiteScopeCluster()}
     * that is executed in a separate test instance. Variables that need to be accessible across test instances must be static.
     */
    @Retention(RetentionPolicy.RUNTIME)
    @Inherited
    @Target(ElementType.TYPE)
    public @interface SuiteScopeTestCase {
    }

    /**
     * If used the test will never run in local mode.
     */
    @Retention(RetentionPolicy.RUNTIME)
    @Inherited
    @Target(ElementType.TYPE)
    public @interface SuppressLocalMode {
    }

    /**
     * If used the test will never run in network mode
     */
    @Retention(RetentionPolicy.RUNTIME)
    @Inherited
    @Target(ElementType.TYPE)
    public @interface SuppressNetworkMode {
    }

}