Preloads .vec and .vex files

LuceneFlatVectorReader uses IOContext.Random to open the read. IOContext.Random
indicates the kernel to not read ahead the pages on to physical memory.
This causes an increase in merge time due to increase of read ops at
runtime.

The preload settings signals the kernal to preload the files when the
reader is opened

Signed-off-by: Tejas Shah <[email protected]>

src/main/java/org/opensearch/knn/index/engine/KNNLibrary.java

@@ -8,7 +8,6 @@
 import org.opensearch.common.ValidationException;
 import org.opensearch.knn.index.SpaceType;
 
-import java.util.Collections;
 import java.util.List;
 
 /**
@@ -137,6 +136,6 @@ KNNLibraryIndexingContext getKNNLibraryIndexingContext(
      * @return list of file extensions that will be read/write with mmap
      */
     default List<String> mmapFileExtensions() {
-        return Collections.emptyList();
+        return List.of("vec", "vex");
     }
 }

src/main/java/org/opensearch/knn/index/engine/lucene/Lucene.java

@@ -15,7 +15,6 @@
 import org.opensearch.knn.index.engine.MethodResolver;
 import org.opensearch.knn.index.engine.ResolvedMethodContext;
 
-import java.util.List;
 import java.util.Map;
 import java.util.function.Function;
 
@@ -89,11 +88,6 @@ public Float scoreToRadialThreshold(Float score, SpaceType spaceType) {
         return score;
     }
 
-    @Override
-    public List<String> mmapFileExtensions() {
-        return List.of("vec", "vex");
-    }
-
     @Override
     public ResolvedMethodContext resolveMethod(
         KNNMethodContext knnMethodContext,

src/main/java/org/opensearch/knn/plugin/KNNPlugin.java

@@ -13,6 +13,7 @@
 import org.opensearch.index.engine.EngineFactory;
 import org.opensearch.indices.SystemIndexDescriptor;
 import org.opensearch.knn.index.KNNCircuitBreaker;
+import org.opensearch.knn.index.engine.KNNEngine;
 import org.opensearch.knn.plugin.search.KNNConcurrentSearchRequestDecider;
 import org.opensearch.knn.index.util.KNNClusterUtil;
 import org.opensearch.knn.index.query.KNNQueryBuilder;
@@ -110,6 +111,7 @@
 import java.util.Map;
 import java.util.Optional;
 import java.util.function.Supplier;
+import java.util.stream.Collectors;
 
 import static java.util.Collections.singletonList;
 import static org.opensearch.knn.common.KNNConstants.KNN_THREAD_POOL_PREFIX;
@@ -352,6 +354,20 @@ public Collection<SystemIndexDescriptor> getSystemIndexDescriptors(Settings sett
         return ImmutableList.of(new SystemIndexDescriptor(MODEL_INDEX_NAME, "Index for storing models used for k-NN indices"));
     }
 
+    /**
+     * Plugin can provide additional node settings, that includes new settings or overrides for existing one from core.
+     *
+     * @return settings that are set by plugin
+     */
+    @Override
+    public Settings additionalSettings() {
+        final List<String> mmapFileExtensions = Arrays.stream(KNNEngine.values())
+            .flatMap(engine -> engine.mmapFileExtensions().stream())
+            .collect(Collectors.toList());
+
+        return Settings.builder().putList(IndexModule.INDEX_STORE_PRE_LOAD_SETTING.getKey(), mmapFileExtensions).build();
+    }
+
     @Override
     public Optional<ConcurrentSearchRequestDecider.Factory> getConcurrentSearchRequestDeciderFactory() {
         return Optional.of(new KNNConcurrentSearchRequestDecider.Factory());

src/test/java/org/opensearch/knn/index/engine/faiss/FaissTests.java

@@ -23,6 +23,7 @@
 
 import java.io.IOException;
 import java.util.HashMap;
+import java.util.List;
 import java.util.Locale;
 import java.util.Map;
 
@@ -367,4 +368,12 @@ public void testMethodAsMapBuilder() throws IOException {
         assertEquals(expectedKNNMethodContext.getVectorValidator(), actualKNNLibraryIndexingContext.getVectorValidator());
     }
 
+    public void testMmapFileExtensions() {
+        final List<String> mMapExtensions = Faiss.INSTANCE.mmapFileExtensions();
+        assertNotNull(mMapExtensions);
+        final List<String> expectedSettings = List.of("vex", "vec");
+        assertTrue(expectedSettings.containsAll(mMapExtensions));
+        assertTrue(mMapExtensions.containsAll(expectedSettings));
+    }
+
 }

src/test/java/org/opensearch/knn/index/engine/nmslib/NMSLibTests.java

@@ -0,0 +1,22 @@
+/*
+ * Copyright OpenSearch Contributors
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+package org.opensearch.knn.index.engine.nmslib;
+
+import org.opensearch.knn.KNNTestCase;
+import org.opensearch.knn.index.engine.faiss.Faiss;
+
+import java.util.List;
+
+public class NMSLibTests extends KNNTestCase {
+
+    public void testMmapFileExtensions() {
+        final List<String> mmapExtensions = Faiss.INSTANCE.mmapFileExtensions();
+        assertNotNull(mmapExtensions);
+        final List<String> expectedSettings = List.of("vex", "vec");
+        assertTrue(expectedSettings.containsAll(mmapExtensions));
+        assertTrue(mmapExtensions.containsAll(expectedSettings));
+    }
+}

src/test/java/org/opensearch/knn/plugin/KNNPluginTests.java

@@ -0,0 +1,23 @@
+/*
+ * Copyright OpenSearch Contributors
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+package org.opensearch.knn.plugin;
+
+import org.opensearch.common.settings.Settings;
+import org.opensearch.test.OpenSearchTestCase;
+
+import java.io.IOException;
+import java.util.HashSet;
+import java.util.Set;
+
+public class KNNPluginTests extends OpenSearchTestCase {
+
+    public void testKNNPlugin_additionalSettings() throws IOException {
+        try (KNNPlugin knnPlugin = new KNNPlugin()) {
+            Settings additionalSettings = knnPlugin.additionalSettings();
+            assertEquals(Set.of("vec", "vex"), new HashSet<>(additionalSettings.getAsList("index.store.preload")));
+        }
+    }
+}