Reindex subset of vertices

Signed-off-by: ntisseyre <[email protected]>

Author: ntisseyre

janusgraph-backend-testutils/src/main/java/org/janusgraph/graphdb/JanusGraphBaseTest.java

@@ -747,7 +747,7 @@ public static void evaluateQuery(JanusGraphQuery query, ElementCategory resultTy
     }
 
     protected ScanMetrics executeScanJob(VertexScanJob job) throws Exception {
-        return executeScanJob(VertexJobConverter.convert(graph,job));
+        return executeScanJob(VertexJobConverter.convert(graph, job, null));
     }
 
     protected ScanMetrics executeScanJob(ScanJob job) throws Exception {

janusgraph-backend-testutils/src/main/java/org/janusgraph/graphdb/JanusGraphIndexTest.java

@@ -1451,6 +1451,88 @@ public void testCompositeVsMixedIndexing() {
         assertTrue(tx.traversal().V().has("intId2", 234).hasNext());
     }
 
+    @Test
+    public void testSubsetReindex() throws Exception {
+
+        clopen(option(FORCE_INDEX_USAGE), true);
+
+        mgmt.makeVertexLabel("cat").make();
+        mgmt.makeVertexLabel("dog").make();
+
+        makeKey("id", Integer.class);
+        makeKey("name", String.class);
+        final PropertyKey typeKey = makeKey("type", String.class);
+
+        String typeIndex = "searchByType";
+        mgmt.buildIndex(typeIndex, Vertex.class)
+            .addKey(typeKey)
+            .buildCompositeIndex();
+        mgmt.commit();
+
+        //Cats
+        int catsCount = 3;
+        for (int i = 0; i < catsCount; i++) {
+            Vertex v = tx.addVertex("cat");
+            v.property("id", i);
+            v.property("name", "cat_" + i);
+            v.property("type", "cat");
+        }
+
+        //Dogs
+        for (int i = 0; i < 5; i++) {
+            Vertex v = tx.addVertex("dog");
+            v.property("id", i);
+            v.property("name", "dog_" + i);
+            v.property("type", "dog");
+        }
+
+        tx.commit();
+
+        //Select a subset of vertices to index
+        clopen(option(FORCE_INDEX_USAGE), true);
+        List<Vertex> cats = tx.traversal().V().has("type", "cat").toList();
+        assertEquals(catsCount, cats.size());
+
+        List<Vertex> dogs = tx.traversal().V().has("type", "dog").toList();
+        assertEquals(5, dogs.size());
+        tx.rollback();
+
+        //Create new Index
+        graph.getOpenTransactions().forEach(JanusGraphTransaction::rollback);
+        mgmt = graph.openManagement();
+        mgmt.getOpenInstances().stream().filter(i -> !i.contains("current")).forEach(i -> mgmt.forceCloseInstance(i));
+        mgmt.commit();
+
+        String catsNameIndex = "searchByName_CatsOnly";
+        mgmt = graph.openManagement();
+        mgmt.buildIndex(catsNameIndex, Vertex.class)
+            .addKey(mgmt.getPropertyKey("name"))
+            .indexOnly(mgmt.getVertexLabel("cat"))
+            .buildCompositeIndex();
+        mgmt.commit();
+
+        //Make Index as REGISTERED
+        mgmt = graph.openManagement();
+        mgmt.updateIndex(mgmt.getGraphIndex(catsNameIndex), SchemaAction.REGISTER_INDEX).get();
+        mgmt.commit();
+        ManagementSystem.awaitGraphIndexStatus(graph, catsNameIndex).status(SchemaStatus.REGISTERED).call();
+
+        //Reindex a given subset
+        mgmt = graph.openManagement();
+        mgmt.updateIndex(mgmt.getGraphIndex(catsNameIndex), SchemaAction.REINDEX,
+            cats.stream().map(Element::id).collect(Collectors.toList())).get();
+        mgmt.commit();
+        ManagementSystem.awaitGraphIndexStatus(graph, catsNameIndex).status(SchemaStatus.ENABLED).call();
+
+        clopen(option(FORCE_INDEX_USAGE), true);
+
+        for (int i = 0; i < catsCount; i++) {
+            List<Vertex> catsByName = tx.traversal().V().hasLabel("cat").has("name", "cat_" + i).toList();
+            assertEquals(1, catsByName.size());
+        }
+        tx.rollback();
+    }
+
     @Test
     public void testIndexInlineProperties() throws NoSuchMethodException {
 

janusgraph-core/src/main/java/org/janusgraph/core/schema/JanusGraphManagement.java

@@ -25,6 +25,7 @@
 import org.janusgraph.diskstorage.keycolumnvalue.scan.ScanJobFuture;
 
 import java.time.Duration;
+import java.util.List;
 import java.util.Set;
 
 /**
@@ -341,6 +342,17 @@ interface IndexBuilder {
      */
     ScanJobFuture updateIndex(Index index, SchemaAction updateAction, int numOfThreads);
 
+    /**
+     * Updates the provided index according to the given {@link SchemaAction} for
+     * the given subset of vertices.
+     *
+     * @param index
+     * @param updateAction
+     * @param vertexOnly Set of vertexIds that only should be considered for index update
+     * @return a future that completes when the index action is done
+     */
+    ScanJobFuture updateIndex(Index index, SchemaAction updateAction, List<Object> vertexOnly);
+
     /**
      * If an index update job was triggered through {@link #updateIndex(Index, SchemaAction)} with schema actions
      * {@link org.janusgraph.core.schema.SchemaAction#REINDEX} or {@link org.janusgraph.core.schema.SchemaAction#DISCARD_INDEX}

janusgraph-core/src/main/java/org/janusgraph/diskstorage/common/DistributedStoreManager.java

@@ -28,6 +28,7 @@
 import static org.janusgraph.graphdb.configuration.GraphDatabaseConfiguration.AUTH_PASSWORD;
 import static org.janusgraph.graphdb.configuration.GraphDatabaseConfiguration.AUTH_USERNAME;
 import static org.janusgraph.graphdb.configuration.GraphDatabaseConfiguration.CONNECTION_TIMEOUT;
+import static org.janusgraph.graphdb.configuration.GraphDatabaseConfiguration.KEYS_SIZE;
 import static org.janusgraph.graphdb.configuration.GraphDatabaseConfiguration.PAGE_SIZE;
 import static org.janusgraph.graphdb.configuration.GraphDatabaseConfiguration.STORAGE_HOSTS;
 import static org.janusgraph.graphdb.configuration.GraphDatabaseConfiguration.STORAGE_PORT;
@@ -69,6 +70,7 @@ public enum Deployment {
     protected final int port;
     protected final Duration connectionTimeoutMS;
     protected final int pageSize;
+    protected final int keysSize;
 
     protected final String username;
     protected final String password;
@@ -83,6 +85,7 @@ public DistributedStoreManager(Configuration storageConfig, int portDefault) {
         else this.port = portDefault;
         this.connectionTimeoutMS = storageConfig.get(CONNECTION_TIMEOUT);
         this.pageSize = storageConfig.get(PAGE_SIZE);
+        this.keysSize = storageConfig.get(KEYS_SIZE);
         this.times = storageConfig.get(TIMESTAMP_PROVIDER);
 
         if (storageConfig.has(AUTH_USERNAME)) {
@@ -121,6 +124,15 @@ public int getPageSize() {
         return pageSize;
     }
 
+    /**
+     * Returns the default configured keys size for this storage backend. The keys size is used to determine
+     * how many keys/partitions to request from storage within single request.
+     * @return
+     */
+    public int getKeysSize() {
+        return keysSize;
+    }
+
     /*
      * TODO this should go away once we have a JanusGraphConfig that encapsulates TimestampProvider
      */

janusgraph-core/src/main/java/org/janusgraph/diskstorage/keycolumnvalue/KeyColumnValueStore.java

@@ -14,6 +14,7 @@
 
 package org.janusgraph.diskstorage.keycolumnvalue;
 
+import org.apache.commons.lang.NotImplementedException;
 import org.janusgraph.diskstorage.BackendException;
 import org.janusgraph.diskstorage.Entry;
 import org.janusgraph.diskstorage.EntryList;
@@ -181,6 +182,10 @@ default Map<SliceQuery, Map<StaticBuffer, EntryList>> getMultiSlices(MultiKeysQu
      */
     void acquireLock(StaticBuffer key, StaticBuffer column, StaticBuffer expectedValue, StoreTransaction txh) throws BackendException;
 
+    default KeyIterator getKeys(final List<StaticBuffer> keys, final SliceQuery query, final StoreTransaction txh) throws BackendException {
+        throw new NotImplementedException();
+    }
+
     /**
      * Returns a {@link KeyIterator} over all keys that fall within the key-range specified by the given query and have one or more columns matching the column-range.
      * Calling {@link KeyIterator#getEntries()} returns the list of all entries that match the column-range specified by the given query.

janusgraph-core/src/main/java/org/janusgraph/diskstorage/keycolumnvalue/scan/MultiThreadsRowsCollector.java

@@ -61,6 +61,7 @@ class MultiThreadsRowsCollector extends RowsCollector {
     private final StoreTransaction storeTx;
     private final List<SliceQuery> queries;
     private final Predicate<StaticBuffer> keyFilter;
+    private final List<StaticBuffer> keysToScan;
     private final Configuration graphConfiguration;
     private final DataPuller[] pullThreads;
     private final BlockingQueue<SliceResult>[] dataQueues;
@@ -72,6 +73,7 @@ class MultiThreadsRowsCollector extends RowsCollector {
         StoreTransaction storeTx,
         List<SliceQuery> queries,
         Predicate<StaticBuffer> keyFilter,
+        List<StaticBuffer> keysToScan,
         BlockingQueue<Row> rowQueue,
         Configuration graphConfiguration) throws BackendException {
 
@@ -80,6 +82,7 @@ class MultiThreadsRowsCollector extends RowsCollector {
         this.storeTx = storeTx;
         this.queries = queries;
         this.keyFilter = keyFilter;
+        this.keysToScan = keysToScan;
         this.graphConfiguration = graphConfiguration;
 
         this.dataQueues = new BlockingQueue[queries.size()];
@@ -189,8 +192,14 @@ private void addDataPuller(SliceQuery sq, StoreTransaction stx, int pos) throws
             this.graphConfiguration.get(GraphDatabaseConfiguration.PAGE_SIZE));
         dataQueues[pos] = queue;
 
-        DataPuller dp = new DataPuller(sq, queue,
-            KCVSUtil.getKeys(store,sq,storeFeatures, MAX_KEY_LENGTH,stx), keyFilter);
+        KeyIterator keyIterator;
+        if (keysToScan != null) {
+            keyIterator = store.getKeys(keysToScan, sq, stx);
+        } else {
+            keyIterator = KCVSUtil.getKeys(store, sq, storeFeatures, MAX_KEY_LENGTH, stx);
+        }
+
+        DataPuller dp = new DataPuller(sq, queue, keyIterator, keyFilter);
         pullThreads[pos] = dp;
         dp.setName("data-puller-" + pos); // setting the name for thread dumps!
         dp.start();

janusgraph-core/src/main/java/org/janusgraph/diskstorage/keycolumnvalue/scan/ScanJob.java

@@ -113,6 +113,14 @@ default void workerIterationEnd(ScanMetrics metrics) {}
      */
     List<SliceQuery> getQueries();
 
+    /**
+     * Get keys to scan by the job. If stream is empty, all keys will be scanned.
+     * @return
+     */
+    default List<StaticBuffer> getKeysToScan() {
+        return null;
+    }
+
     /**
      * A predicate that determines whether
      * {@link #process(org.janusgraph.diskstorage.StaticBuffer, java.util.Map, ScanMetrics)}

janusgraph-core/src/main/java/org/janusgraph/diskstorage/keycolumnvalue/scan/StandardScannerExecutor.java

@@ -167,7 +167,7 @@ private RowsCollector buildScanner(BlockingQueue<Row> processorQueue, List<Slice
                 job.getKeyFilter(), processorQueue);
         } else {
             return new MultiThreadsRowsCollector(store, storeFeatures, storeTx, queries,
-                job.getKeyFilter(), processorQueue, graphConfiguration);
+                job.getKeyFilter(), job.getKeysToScan(), processorQueue, graphConfiguration);
         }
     }
 

janusgraph-core/src/main/java/org/janusgraph/diskstorage/util/MetricInstrumentedStore.java

@@ -165,6 +165,18 @@ public void acquireLock(final StaticBuffer key,
         });
     }
 
+    @Override
+    public KeyIterator getKeys(final List<StaticBuffer> keys, final SliceQuery query, final StoreTransaction txh) throws BackendException {
+        return runWithMetrics(txh, metricsStoreName, M_GET_KEYS, () -> {
+            final KeyIterator ki = backend.getKeys(keys, query, txh);
+            if (txh.getConfiguration().hasGroupName()) {
+                return MetricInstrumentedIterator.of(ki, txh.getConfiguration().getGroupName(), metricsStoreName, M_GET_KEYS, M_ITERATOR);
+            } else {
+                return ki;
+            }
+        });
+    }
+
     @Override
     public KeyIterator getKeys(final KeyRangeQuery query, final StoreTransaction txh) throws BackendException {
         return runWithMetrics(txh, metricsStoreName, M_GET_KEYS, () -> {

janusgraph-core/src/main/java/org/janusgraph/graphdb/configuration/GraphDatabaseConfiguration.java

@@ -830,6 +830,10 @@ public boolean apply(@Nullable String s) {
             "up to this many elements.",
             ConfigOption.Type.MASKABLE, 100);
 
+    public static final ConfigOption<Integer> KEYS_SIZE = new ConfigOption<>(STORAGE_NS,"keys-size",
+        "The maximum amount of keys/partitions to retrieve from distributed storage system by JanusGraph in a single request.",
+        ConfigOption.Type.MASKABLE, 100);
+
     public static final ConfigOption<Boolean> DROP_ON_CLEAR = new ConfigOption<>(STORAGE_NS, "drop-on-clear",
             "Whether to drop the graph database (true) or delete rows (false) when clearing storage. " +
             "Note that some backends always drop the graph database when clearing storage. Also note that indices are " +