Introduce remote store path type in customData in IndexMetadata (opensearch-project#12607)

ashking94 · ashking94 · commit 7eec761012d0 · 2024-04-23T13:33:37.000+05:30
Signed-off-by: Ashish Singh &lt;ssashish@amazon.com&gt;
diff --git a/server/src/main/java/org/opensearch/cluster/metadata/IndexMetadata.java b/server/src/main/java/org/opensearch/cluster/metadata/IndexMetadata.java
@@ -635,6 +635,7 @@ public static APIBlock readFrom(StreamInput input) throws IOException {
     static final String KEY_ROLLOVER_INFOS = "rollover_info";
     static final String KEY_SYSTEM = "system";
     public static final String KEY_PRIMARY_TERMS = "primary_terms";
+    public static final String REMOTE_STORE_CUSTOM_KEY = "remote_store";
 
     public static final String INDEX_STATE_FILE_PREFIX = "state-";
 
diff --git a/server/src/main/java/org/opensearch/cluster/metadata/MetadataCreateIndexService.java b/server/src/main/java/org/opensearch/cluster/metadata/MetadataCreateIndexService.java
@@ -90,6 +90,8 @@
 import org.opensearch.index.mapper.MapperService;
 import org.opensearch.index.mapper.MapperService.MergeReason;
 import org.opensearch.index.query.QueryShardContext;
+import org.opensearch.index.remote.RemoteStorePathResolver;
+import org.opensearch.index.remote.RemoteStorePathType;
 import org.opensearch.index.shard.IndexSettingProvider;
 import org.opensearch.index.translog.Translog;
 import org.opensearch.indices.IndexCreationException;
@@ -171,6 +173,9 @@ public class MetadataCreateIndexService {
     private final ClusterManagerTaskThrottler.ThrottlingKey createIndexTaskKey;
     private AwarenessReplicaBalance awarenessReplicaBalance;
 
+    @Nullable
+    private final RemoteStorePathResolver remoteStorePathResolver;
+
     public MetadataCreateIndexService(
         final Settings settings,
         final ClusterService clusterService,
@@ -202,6 +207,9 @@ public MetadataCreateIndexService(
 
         // Task is onboarded for throttling, it will get retried from associated TransportClusterManagerNodeAction.
         createIndexTaskKey = clusterService.registerClusterManagerTask(ClusterManagerTaskKeys.CREATE_INDEX_KEY, true);
+        remoteStorePathResolver = isRemoteDataAttributePresent(settings)
+            ? new RemoteStorePathResolver(clusterService.getClusterSettings())
+            : null;
     }
 
     /**
@@ -502,7 +510,8 @@ private ClusterState applyCreateIndexWithTemporaryService(
                     temporaryIndexMeta.getSettings(),
                     temporaryIndexMeta.getRoutingNumShards(),
                     sourceMetadata,
-                    temporaryIndexMeta.isSystem()
+                    temporaryIndexMeta.isSystem(),
+                    temporaryIndexMeta.getCustomData()
                 );
             } catch (Exception e) {
                 logger.info("failed to build index metadata [{}]", request.index());
@@ -526,10 +535,11 @@ private ClusterState applyCreateIndexWithTemporaryService(
 
     /**
      * Given a state and index settings calculated after applying templates, validate metadata for
-     * the new index, returning an {@link IndexMetadata} for the new index
+     * the new index, returning an {@link IndexMetadata} for the new index.
+     * <p>
+     * The access level of the method changed to default level for visibility to test.
      */
-    private IndexMetadata buildAndValidateTemporaryIndexMetadata(
-        final ClusterState currentState,
+    IndexMetadata buildAndValidateTemporaryIndexMetadata(
         final Settings aggregatedIndexSettings,
         final CreateIndexClusterStateUpdateRequest request,
         final int routingNumShards
@@ -548,6 +558,11 @@ private IndexMetadata buildAndValidateTemporaryIndexMetadata(
         tmpImdBuilder.settings(indexSettings);
         tmpImdBuilder.system(isSystem);
 
+        if (remoteStorePathResolver != null) {
+            String pathType = remoteStorePathResolver.resolveType().toString();
+            tmpImdBuilder.putCustom(IndexMetadata.REMOTE_STORE_CUSTOM_KEY, Map.of(RemoteStorePathType.NAME, pathType));
+        }
+
         // Set up everything, now locally create the index to see that things are ok, and apply
         IndexMetadata tempMetadata = tmpImdBuilder.build();
         validateActiveShardCount(request.waitForActiveShards(), tempMetadata);
@@ -586,7 +601,7 @@ private ClusterState applyCreateIndexRequestWithV1Templates(
             clusterService.getClusterSettings()
         );
         int routingNumShards = getIndexNumberOfRoutingShards(aggregatedIndexSettings, null);
-        IndexMetadata tmpImd = buildAndValidateTemporaryIndexMetadata(currentState, aggregatedIndexSettings, request, routingNumShards);
+        IndexMetadata tmpImd = buildAndValidateTemporaryIndexMetadata(aggregatedIndexSettings, request, routingNumShards);
 
         return applyCreateIndexWithTemporaryService(
             currentState,
@@ -651,7 +666,7 @@ private ClusterState applyCreateIndexRequestWithV2Template(
             clusterService.getClusterSettings()
         );
         int routingNumShards = getIndexNumberOfRoutingShards(aggregatedIndexSettings, null);
-        IndexMetadata tmpImd = buildAndValidateTemporaryIndexMetadata(currentState, aggregatedIndexSettings, request, routingNumShards);
+        IndexMetadata tmpImd = buildAndValidateTemporaryIndexMetadata(aggregatedIndexSettings, request, routingNumShards);
 
         return applyCreateIndexWithTemporaryService(
             currentState,
@@ -732,7 +747,7 @@ private ClusterState applyCreateIndexRequestWithExistingMetadata(
             clusterService.getClusterSettings()
         );
         final int routingNumShards = getIndexNumberOfRoutingShards(aggregatedIndexSettings, sourceMetadata);
-        IndexMetadata tmpImd = buildAndValidateTemporaryIndexMetadata(currentState, aggregatedIndexSettings, request, routingNumShards);
+        IndexMetadata tmpImd = buildAndValidateTemporaryIndexMetadata(aggregatedIndexSettings, request, routingNumShards);
 
         return applyCreateIndexWithTemporaryService(
             currentState,
@@ -1203,7 +1218,8 @@ static IndexMetadata buildIndexMetadata(
         Settings indexSettings,
         int routingNumShards,
         @Nullable IndexMetadata sourceMetadata,
-        boolean isSystem
+        boolean isSystem,
+        Map<String, DiffableStringMap> customData
     ) {
         IndexMetadata.Builder indexMetadataBuilder = createIndexMetadataBuilder(indexName, sourceMetadata, indexSettings, routingNumShards);
         indexMetadataBuilder.system(isSystem);
@@ -1224,6 +1240,10 @@ static IndexMetadata buildIndexMetadata(
             indexMetadataBuilder.putAlias(aliases.get(i));
         }
 
+        for (Map.Entry<String, DiffableStringMap> entry : customData.entrySet()) {
+            indexMetadataBuilder.putCustom(entry.getKey(), entry.getValue());
+        }
+
         indexMetadataBuilder.state(IndexMetadata.State.OPEN);
         return indexMetadataBuilder.build();
     }
diff --git a/server/src/main/java/org/opensearch/common/settings/ClusterSettings.java b/server/src/main/java/org/opensearch/common/settings/ClusterSettings.java
@@ -729,9 +729,10 @@ public void apply(Settings value, Settings current, Settings previous) {
                 IoBasedAdmissionControllerSettings.IO_BASED_ADMISSION_CONTROLLER_TRANSPORT_LAYER_MODE,
                 IoBasedAdmissionControllerSettings.SEARCH_IO_USAGE_LIMIT,
                 IoBasedAdmissionControllerSettings.INDEXING_IO_USAGE_LIMIT,
-
                 RemoteStoreSettings.CLUSTER_REMOTE_INDEX_SEGMENT_METADATA_RETENTION_MAX_COUNT_SETTING,
-                RemoteStoreSettings.CLUSTER_REMOTE_TRANSLOG_BUFFER_INTERVAL_SETTING
+                RemoteStoreSettings.CLUSTER_REMOTE_TRANSLOG_BUFFER_INTERVAL_SETTING,
+                IndicesService.CLUSTER_INDEX_RESTRICT_REPLICATION_TYPE_SETTING,
+                IndicesService.CLUSTER_REMOTE_STORE_PATH_PREFIX_TYPE_SETTING
             )
         )
     );
diff --git a/server/src/main/java/org/opensearch/index/remote/RemoteStorePathResolver.java b/server/src/main/java/org/opensearch/index/remote/RemoteStorePathResolver.java
@@ -0,0 +1,30 @@
+/*
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * The OpenSearch Contributors require contributions made to
+ * this file be licensed under the Apache-2.0 license or a
+ * compatible open source license.
+ */
+
+package org.opensearch.index.remote;
+
+import org.opensearch.common.settings.ClusterSettings;
+import org.opensearch.indices.IndicesService;
+
+/**
+ * Determines the {@link RemoteStorePathType} at the time of index metadata creation.
+ *
+ * @opensearch.internal
+ */
+public class RemoteStorePathResolver {
+
+    private final ClusterSettings clusterSettings;
+
+    public RemoteStorePathResolver(ClusterSettings clusterSettings) {
+        this.clusterSettings = clusterSettings;
+    }
+
+    public RemoteStorePathType resolveType() {
+        return clusterSettings.get(IndicesService.CLUSTER_REMOTE_STORE_PATH_PREFIX_TYPE_SETTING);
+    }
+}
diff --git a/server/src/main/java/org/opensearch/index/remote/RemoteStorePathType.java b/server/src/main/java/org/opensearch/index/remote/RemoteStorePathType.java
@@ -0,0 +1,36 @@
+/*
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * The OpenSearch Contributors require contributions made to
+ * this file be licensed under the Apache-2.0 license or a
+ * compatible open source license.
+ */
+
+package org.opensearch.index.remote;
+
+import java.util.Locale;
+
+/**
+ * Enumerates the types of remote store paths resolution techniques supported by OpenSearch.
+ * For more information, see <a href="https://github.com/opensearch-project/OpenSearch/issues/12567">Github issue #12567</a>.
+ *
+ * @opensearch.internal
+ */
+public enum RemoteStorePathType {
+
+    FIXED,
+    HASHED_PREFIX;
+
+    public static RemoteStorePathType parseString(String remoteStoreBlobPathType) {
+        try {
+            return RemoteStorePathType.valueOf(remoteStoreBlobPathType.toUpperCase(Locale.ROOT));
+        } catch (IllegalArgumentException e) {
+            throw new IllegalArgumentException("Could not parse RemoteStorePathType for [" + remoteStoreBlobPathType + "]");
+        }
+    }
+
+    /**
+     * This string is used as key for storing information in the custom data in index settings.
+     */
+    public static final String NAME = "path_type";
+}
diff --git a/server/src/main/java/org/opensearch/indices/IndicesService.java b/server/src/main/java/org/opensearch/indices/IndicesService.java
@@ -126,6 +126,7 @@
 import org.opensearch.index.query.QueryRewriteContext;
 import org.opensearch.index.recovery.RecoveryStats;
 import org.opensearch.index.refresh.RefreshStats;
+import org.opensearch.index.remote.RemoteStorePathType;
 import org.opensearch.index.remote.RemoteStoreStatsTrackerFactory;
 import org.opensearch.index.search.stats.SearchStats;
 import org.opensearch.index.seqno.RetentionLeaseStats;
@@ -307,6 +308,18 @@ public class IndicesService extends AbstractLifecycleComponent
         Property.Final
     );
 
+    /**
+     * This setting is used to set the remote store blob store path prefix strategy. This setting is effective only for
+     * remote store enabled cluster.
+     */
+    public static final Setting<RemoteStorePathType> CLUSTER_REMOTE_STORE_PATH_PREFIX_TYPE_SETTING = new Setting<>(
+        "cluster.remote_store.index.path.prefix.type",
+        RemoteStorePathType.FIXED.toString(),
+        RemoteStorePathType::parseString,
+        Property.NodeScope,
+        Property.Dynamic
+    );
+
     /**
      * The node's settings.
      */
diff --git a/server/src/test/java/org/opensearch/cluster/metadata/MetadataCreateIndexServiceTests.java b/server/src/test/java/org/opensearch/cluster/metadata/MetadataCreateIndexServiceTests.java
@@ -74,6 +74,7 @@
 import org.opensearch.index.IndexSettings;
 import org.opensearch.index.mapper.MapperService;
 import org.opensearch.index.query.QueryShardContext;
+import org.opensearch.index.remote.RemoteStorePathType;
 import org.opensearch.index.translog.Translog;
 import org.opensearch.indices.IndexCreationException;
 import org.opensearch.indices.IndicesService;
@@ -1676,13 +1677,102 @@ public void testBuildIndexMetadata() {
             .put(SETTING_NUMBER_OF_SHARDS, 1)
             .build();
         List<AliasMetadata> aliases = singletonList(AliasMetadata.builder("alias1").build());
-        IndexMetadata indexMetadata = buildIndexMetadata("test", aliases, () -> null, indexSettings, 4, sourceIndexMetadata, false);
+        IndexMetadata indexMetadata = buildIndexMetadata(
+            "test",
+            aliases,
+            () -> null,
+            indexSettings,
+            4,
+            sourceIndexMetadata,
+            false,
+            new HashMap<>()
+        );
 
         assertThat(indexMetadata.getAliases().size(), is(1));
         assertThat(indexMetadata.getAliases().keySet().iterator().next(), is("alias1"));
         assertThat("The source index primary term must be used", indexMetadata.primaryTerm(0), is(3L));
     }
 
+    /**
+     * This test checks if the cluster is a remote store cluster then we populate custom data for remote settings in
+     * index metadata of the underlying index. This captures information around the resolution pattern of the path for
+     * remote segments and translog.
+     */
+    public void testRemoteCustomData() {
+        // Case 1 - Remote store is not enabled
+        IndexMetadata indexMetadata = testRemoteCustomData(false, randomFrom(RemoteStorePathType.values()));
+        assertNull(indexMetadata.getCustomData(IndexMetadata.REMOTE_STORE_CUSTOM_KEY));
+
+        // Case 2 - cluster.remote_store.index.path.prefix.optimised=fixed (default value)
+        indexMetadata = testRemoteCustomData(true, RemoteStorePathType.FIXED);
+        validateRemoteCustomData(
+            indexMetadata.getCustomData(IndexMetadata.REMOTE_STORE_CUSTOM_KEY),
+            RemoteStorePathType.NAME,
+            RemoteStorePathType.FIXED.toString()
+        );
+
+        // Case 3 - cluster.remote_store.index.path.prefix.optimised=hashed_prefix
+        indexMetadata = testRemoteCustomData(true, RemoteStorePathType.HASHED_PREFIX);
+        validateRemoteCustomData(
+            indexMetadata.getCustomData(IndexMetadata.REMOTE_STORE_CUSTOM_KEY),
+            RemoteStorePathType.NAME,
+            RemoteStorePathType.HASHED_PREFIX.toString()
+        );
+    }
+
+    private IndexMetadata testRemoteCustomData(boolean remoteStoreEnabled, RemoteStorePathType remoteStorePathType) {
+        Settings.Builder settingsBuilder = Settings.builder();
+        if (remoteStoreEnabled) {
+            settingsBuilder.put(NODE_ATTRIBUTES.getKey() + REMOTE_STORE_SEGMENT_REPOSITORY_NAME_ATTRIBUTE_KEY, "test");
+        }
+        settingsBuilder.put(IndicesService.CLUSTER_REMOTE_STORE_PATH_PREFIX_TYPE_SETTING.getKey(), remoteStorePathType.toString());
+        Settings settings = settingsBuilder.build();
+
+        ClusterService clusterService = mock(ClusterService.class);
+        Metadata metadata = Metadata.builder()
+            .transientSettings(Settings.builder().put(Metadata.DEFAULT_REPLICA_COUNT_SETTING.getKey(), 1).build())
+            .build();
+        ClusterState clusterState = ClusterState.builder(org.opensearch.cluster.ClusterName.CLUSTER_NAME_SETTING.getDefault(Settings.EMPTY))
+            .metadata(metadata)
+            .build();
+        ClusterSettings clusterSettings = new ClusterSettings(settings, ClusterSettings.BUILT_IN_CLUSTER_SETTINGS);
+        when(clusterService.getSettings()).thenReturn(settings);
+        when(clusterService.getClusterSettings()).thenReturn(clusterSettings);
+        when(clusterService.state()).thenReturn(clusterState);
+
+        ThreadPool threadPool = new TestThreadPool(getTestName());
+        MetadataCreateIndexService metadataCreateIndexService = new MetadataCreateIndexService(
+            settings,
+            clusterService,
+            null,
+            null,
+            null,
+            createTestShardLimitService(randomIntBetween(1, 1000), false, clusterService),
+            new Environment(Settings.builder().put("path.home", "dummy").build(), null),
+            IndexScopedSettings.DEFAULT_SCOPED_SETTINGS,
+            threadPool,
+            null,
+            new SystemIndices(Collections.emptyMap()),
+            true,
+            new AwarenessReplicaBalance(settings, clusterService.getClusterSettings())
+        );
+        CreateIndexClusterStateUpdateRequest request = new CreateIndexClusterStateUpdateRequest("create index", "test", "test");
+        Settings indexSettings = Settings.builder()
+            .put("index.version.created", Version.CURRENT)
+            .put(INDEX_NUMBER_OF_SHARDS_SETTING.getKey(), 3)
+            .put(IndexMetadata.SETTING_NUMBER_OF_REPLICAS, 1)
+            .build();
+
+        IndexMetadata indexMetadata = metadataCreateIndexService.buildAndValidateTemporaryIndexMetadata(indexSettings, request, 0);
+        threadPool.shutdown();
+        return indexMetadata;
+    }
+
+    private void validateRemoteCustomData(Map<String, String> customData, String expectedKey, String expectedValue) {
+        assertTrue(customData.containsKey(expectedKey));
+        assertEquals(expectedValue, customData.get(expectedKey));
+    }
+
     public void testGetIndexNumberOfRoutingShardsWithNullSourceIndex() {
         Settings indexSettings = Settings.builder()
             .put("index.version.created", Version.CURRENT)

Original file line number	Diff line number	Diff line change
`@@ -729,9 +729,10 @@ public void apply(Settings value, Settings current, Settings previous) {`
`729`	`729`	`IoBasedAdmissionControllerSettings.IO_BASED_ADMISSION_CONTROLLER_TRANSPORT_LAYER_MODE,`
`730`	`730`	`IoBasedAdmissionControllerSettings.SEARCH_IO_USAGE_LIMIT,`
`731`	`731`	`IoBasedAdmissionControllerSettings.INDEXING_IO_USAGE_LIMIT,`
`732`		`-`
`733`	`732`	`RemoteStoreSettings.CLUSTER_REMOTE_INDEX_SEGMENT_METADATA_RETENTION_MAX_COUNT_SETTING,`
`734`		`- RemoteStoreSettings.CLUSTER_REMOTE_TRANSLOG_BUFFER_INTERVAL_SETTING`
	`733`	`+ RemoteStoreSettings.CLUSTER_REMOTE_TRANSLOG_BUFFER_INTERVAL_SETTING,`
	`734`	`+ IndicesService.CLUSTER_INDEX_RESTRICT_REPLICATION_TYPE_SETTING,`
	`735`	`+ IndicesService.CLUSTER_REMOTE_STORE_PATH_PREFIX_TYPE_SETTING`
`735`	`736`	`)`
`736`	`737`	`)`
`737`	`738`	`);`