Search in sources :

Example 1 with CompletedJobFuture

use of org.janusgraph.diskstorage.keycolumnvalue.scan.CompletedJobFuture in project janusgraph by JanusGraph.

the class MapReduceIndexManagement method updateIndex.

/**
 * Updates the provided index according to the given {@link SchemaAction}.
 * Only {@link SchemaAction#REINDEX} and {@link SchemaAction#REMOVE_INDEX} are supported.
 *
 * @param index the index to process
 * @param updateAction either {@code REINDEX} or {@code REMOVE_INDEX}
 * @param hadoopConf
 * @return a future that returns immediately;
 *         this method blocks until the Hadoop MapReduce job completes
 */
// TODO make this future actually async and update javadoc @return accordingly
public ScanJobFuture updateIndex(Index index, SchemaAction updateAction, Configuration hadoopConf) throws BackendException {
    Preconditions.checkNotNull(index, "Index parameter must not be null", index);
    Preconditions.checkNotNull(updateAction, "%s parameter must not be null", SchemaAction.class.getSimpleName());
    Preconditions.checkArgument(SUPPORTED_ACTIONS.contains(updateAction), "Only these %s parameters are supported: %s (was given %s)", SchemaAction.class.getSimpleName(), SUPPORTED_ACTIONS_STRING, updateAction);
    Preconditions.checkArgument(RelationTypeIndex.class.isAssignableFrom(index.getClass()) || JanusGraphIndex.class.isAssignableFrom(index.getClass()), "Index %s has class %s: must be a %s or %s (or subtype)", index.getClass(), RelationTypeIndex.class.getSimpleName(), JanusGraphIndex.class.getSimpleName());
    ModifiableHadoopConfiguration janusGraphMapReduceConfiguration = ModifiableHadoopConfiguration.of(JanusGraphHadoopConfiguration.MAPRED_NS, hadoopConf);
    // The job we'll execute to either REINDEX or REMOVE_INDEX
    final Class<? extends IndexUpdateJob> indexJobClass;
    final Class<? extends Mapper> mapperClass;
    // The class of the IndexUpdateJob and the Mapper that will be used to run it (VertexScanJob vs ScanJob)
    if (updateAction.equals(SchemaAction.REINDEX)) {
        indexJobClass = IndexRepairJob.class;
        mapperClass = HadoopVertexScanMapper.class;
    } else {
        assert updateAction.equals(SchemaAction.REMOVE_INDEX);
        indexJobClass = IndexRemoveJob.class;
        mapperClass = HadoopScanMapper.class;
    }
    // The column family that serves as input to the IndexUpdateJob
    final String readCF;
    if (RelationTypeIndex.class.isAssignableFrom(index.getClass())) {
        readCF = Backend.EDGESTORE_NAME;
    } else {
        JanusGraphIndex graphIndex = (JanusGraphIndex) index;
        if (graphIndex.isMixedIndex() && !updateAction.equals(SchemaAction.REINDEX))
            throw new UnsupportedOperationException("External mixed indexes must be removed in the indexing system directly.");
        Preconditions.checkState(JanusGraphIndex.class.isAssignableFrom(index.getClass()));
        if (updateAction.equals(SchemaAction.REMOVE_INDEX))
            readCF = Backend.INDEXSTORE_NAME;
        else
            readCF = Backend.EDGESTORE_NAME;
    }
    janusGraphMapReduceConfiguration.set(JanusGraphHadoopConfiguration.COLUMN_FAMILY_NAME, readCF);
    // The MapReduce InputFormat class based on the open graph's store manager
    HadoopStoreManager storeManager = (HadoopStoreManager) graph.getBackend().getStoreManager().getHadoopManager();
    if (storeManager == null) {
        throw new IllegalArgumentException("Store manager class " + graph.getBackend().getStoreManagerClass() + "is not supported");
    }
    final Class<? extends InputFormat> inputFormat = storeManager.getInputFormat(hadoopConf);
    // The index name and relation type name (if the latter is applicable)
    final String indexName = index.name();
    final RelationType relationType = RelationTypeIndex.class.isAssignableFrom(index.getClass()) ? ((RelationTypeIndex) index).getType() : null;
    final String relationTypeName = relationType == null ? StringUtils.EMPTY : relationType.name();
    Preconditions.checkNotNull(indexName);
    // Set the class of the IndexUpdateJob
    janusGraphMapReduceConfiguration.set(JanusGraphHadoopConfiguration.SCAN_JOB_CLASS, indexJobClass.getName());
    // Set the configuration of the IndexUpdateJob
    copyIndexJobKeys(hadoopConf, indexName, relationTypeName);
    janusGraphMapReduceConfiguration.set(JanusGraphHadoopConfiguration.SCAN_JOB_CONFIG_ROOT, GraphDatabaseConfiguration.class.getName() + "#JOB_NS");
    // Copy the StandardJanusGraph configuration under JanusGraphHadoopConfiguration.GRAPH_CONFIG_KEYS
    org.apache.commons.configuration2.Configuration localConfiguration = graph.getConfiguration().getConfigurationAtOpen();
    localConfiguration.clearProperty(Graph.GRAPH);
    copyInputKeys(hadoopConf, localConfiguration);
    String jobName = HadoopScanMapper.class.getSimpleName() + "[" + indexJobClass.getSimpleName() + "]";
    try {
        return new CompletedJobFuture(HadoopScanRunner.runJob(hadoopConf, inputFormat, jobName, mapperClass));
    } catch (Exception e) {
        return new FailedJobFuture(e);
    }
}
Also used : HadoopScanMapper(org.janusgraph.hadoop.scan.HadoopScanMapper) SchemaAction(org.janusgraph.core.schema.SchemaAction) FailedJobFuture(org.janusgraph.diskstorage.keycolumnvalue.scan.FailedJobFuture) ModifiableHadoopConfiguration(org.janusgraph.hadoop.config.ModifiableHadoopConfiguration) RelationTypeIndex(org.janusgraph.core.schema.RelationTypeIndex) BackendException(org.janusgraph.diskstorage.BackendException) CompletedJobFuture(org.janusgraph.diskstorage.keycolumnvalue.scan.CompletedJobFuture) RelationType(org.janusgraph.core.RelationType) JanusGraphIndex(org.janusgraph.core.schema.JanusGraphIndex)

Aggregations

RelationType (org.janusgraph.core.RelationType)1 JanusGraphIndex (org.janusgraph.core.schema.JanusGraphIndex)1 RelationTypeIndex (org.janusgraph.core.schema.RelationTypeIndex)1 SchemaAction (org.janusgraph.core.schema.SchemaAction)1 BackendException (org.janusgraph.diskstorage.BackendException)1 CompletedJobFuture (org.janusgraph.diskstorage.keycolumnvalue.scan.CompletedJobFuture)1 FailedJobFuture (org.janusgraph.diskstorage.keycolumnvalue.scan.FailedJobFuture)1 ModifiableHadoopConfiguration (org.janusgraph.hadoop.config.ModifiableHadoopConfiguration)1 HadoopScanMapper (org.janusgraph.hadoop.scan.HadoopScanMapper)1