Examples with IndexManager org.apache.nifi.provenance.lucene.IndexManager used on opensource projects

Example 1 with IndexManager

use of org.apache.nifi.provenance.lucene.IndexManager in project nifi by apache.

the class TestPersistentProvenanceRepository method testModifyIndexWhileSearching.

@Test(timeout = 10000)
public void testModifyIndexWhileSearching() throws IOException, InterruptedException, ParseException {
    assumeFalse(isWindowsEnvironment());
    final RepositoryConfiguration config = createConfiguration();
    config.setMaxRecordLife(30, TimeUnit.SECONDS);
    config.setMaxStorageCapacity(1024L * 1024L * 10);
    config.setMaxEventFileLife(500, TimeUnit.MILLISECONDS);
    config.setMaxEventFileCapacity(1024L * 1024L * 10);
    config.setSearchableFields(new ArrayList<>(SearchableFields.getStandardFields()));
    final CountDownLatch obtainIndexSearcherLatch = new CountDownLatch(2);
    repo = new PersistentProvenanceRepository(config, DEFAULT_ROLLOVER_MILLIS) {

        private CachingIndexManager wrappedManager = null;

        // Create an IndexManager that adds a delay before returning the Index Searcher.
        @Override
        protected synchronized CachingIndexManager getIndexManager() {
            if (wrappedManager == null) {
                final IndexManager mgr = super.getIndexManager();
                final Logger logger = LoggerFactory.getLogger("IndexManager");
                wrappedManager = new CachingIndexManager() {

                    final AtomicInteger indexSearcherCount = new AtomicInteger(0);

                    @Override
                    public EventIndexSearcher borrowIndexSearcher(File indexDir) throws IOException {
                        final EventIndexSearcher searcher = mgr.borrowIndexSearcher(indexDir);
                        final int idx = indexSearcherCount.incrementAndGet();
                        obtainIndexSearcherLatch.countDown();
                        // second thread is still holding the searcher
                        try {
                            if (idx == 1) {
                                Thread.sleep(3000L);
                            } else {
                                Thread.sleep(5000L);
                            }
                        } catch (InterruptedException e) {
                            throw new IOException("Interrupted", e);
                        }
                        logger.info("Releasing index searcher");
                        return searcher;
                    }

                    @Override
                    public EventIndexWriter borrowIndexWriter(File indexingDirectory) throws IOException {
                        return mgr.borrowIndexWriter(indexingDirectory);
                    }

                    @Override
                    public void close() throws IOException {
                        mgr.close();
                    }

                    @Override
                    public boolean removeIndex(File indexDirectory) {
                        mgr.removeIndex(indexDirectory);
                        return true;
                    }

                    @Override
                    public void returnIndexSearcher(EventIndexSearcher searcher) {
                        mgr.returnIndexSearcher(searcher);
                    }

                    @Override
                    public void returnIndexWriter(EventIndexWriter writer) {
                        mgr.returnIndexWriter(writer);
                    }
                };
            }
            return wrappedManager;
        }
    };
    repo.initialize(getEventReporter(), null, null, IdentifierLookup.EMPTY);
    final String uuid = "10000000-0000-0000-0000-000000000000";
    final Map<String, String> attributes = new HashMap<>();
    attributes.put("abc", "xyz");
    attributes.put("xyz", "abc");
    attributes.put("filename", "file-" + uuid);
    final ProvenanceEventBuilder builder = new StandardProvenanceEventRecord.Builder();
    builder.setEventTime(System.currentTimeMillis());
    builder.setEventType(ProvenanceEventType.RECEIVE);
    builder.setTransitUri("nifi://unit-test");
    attributes.put("uuid", uuid);
    builder.fromFlowFile(createFlowFile(3L, 3000L, attributes));
    builder.setComponentId("1234");
    builder.setComponentType("dummy processor");
    for (int i = 0; i < 10; i++) {
        builder.fromFlowFile(createFlowFile(i, 3000L, attributes));
        attributes.put("uuid", "00000000-0000-0000-0000-00000000000" + i);
        repo.registerEvent(builder.build());
    }
    repo.waitForRollover();
    // Perform a query. This will ensure that an IndexSearcher is created and cached.
    final Query query = new Query(UUID.randomUUID().toString());
    query.addSearchTerm(SearchTerms.newSearchTerm(SearchableFields.Filename, "file-*"));
    query.addSearchTerm(SearchTerms.newSearchTerm(SearchableFields.ComponentID, "12?4"));
    query.addSearchTerm(SearchTerms.newSearchTerm(SearchableFields.TransitURI, "nifi://*"));
    query.setMaxResults(100);
    // Run a query in a background thread. When this thread goes to obtain the IndexSearcher, it will have a 5 second delay.
    // That delay will occur as the main thread is updating the index. This should result in the search creating a new Index Reader
    // that can properly query the index.
    final int numThreads = 2;
    final CountDownLatch performSearchLatch = new CountDownLatch(numThreads);
    final Runnable searchRunnable = new Runnable() {

        @Override
        public void run() {
            QueryResult result;
            try {
                result = repo.queryEvents(query, createUser());
            } catch (IOException e) {
                e.printStackTrace();
                Assert.fail(e.toString());
                return;
            }
            System.out.println("Finished search: " + result);
            performSearchLatch.countDown();
        }
    };
    // Kick off the searcher threads
    for (int i = 0; i < numThreads; i++) {
        final Thread searchThread = new Thread(searchRunnable);
        searchThread.start();
    }
    // Wait until we've obtained the Index Searchers before modifying the index.
    obtainIndexSearcherLatch.await();
    // add more events to the repo
    for (int i = 0; i < 10; i++) {
        builder.fromFlowFile(createFlowFile(i, 3000L, attributes));
        attributes.put("uuid", "00000000-0000-0000-0000-00000000000" + i);
        repo.registerEvent(builder.build());
    }
    // Force a rollover to occur. This will modify the index.
    repo.rolloverWithLock(true);
    // Wait for the repository to roll over.
    repo.waitForRollover();
    // Wait for the searches to complete.
    performSearchLatch.await();
}

Example 2 with IndexManager

use of org.apache.nifi.provenance.lucene.IndexManager in project nifi by apache.

the class TestEventIndexTask method testIndexWriterCommittedWhenAppropriate.

@Test(timeout = 5000)
public void testIndexWriterCommittedWhenAppropriate() throws IOException, InterruptedException {
    final BlockingQueue<StoredDocument> docQueue = new LinkedBlockingQueue<>();
    final RepositoryConfiguration repoConfig = new RepositoryConfiguration();
    final File storageDir = new File("target/storage/TestEventIndexTask/1");
    repoConfig.addStorageDirectory("1", storageDir);
    final AtomicInteger commitCount = new AtomicInteger(0);
    // Mock out an IndexWriter and keep track of the number of events that are indexed.
    final IndexWriter indexWriter = Mockito.mock(IndexWriter.class);
    final EventIndexWriter eventIndexWriter = new LuceneEventIndexWriter(indexWriter, storageDir);
    final IndexManager indexManager = Mockito.mock(IndexManager.class);
    Mockito.when(indexManager.borrowIndexWriter(Mockito.any(File.class))).thenReturn(eventIndexWriter);
    final IndexDirectoryManager directoryManager = new IndexDirectoryManager(repoConfig);
    // Create an EventIndexTask and override the commit(IndexWriter) method so that we can keep track of how
    // many times the index writer gets committed.
    final EventIndexTask task = new EventIndexTask(docQueue, repoConfig, indexManager, directoryManager, 201, EventReporter.NO_OP) {

        @Override
        protected void commit(EventIndexWriter indexWriter) throws IOException {
            commitCount.incrementAndGet();
        }
    };
    // Create 4 threads, each one a daemon thread running the EventIndexTask
    for (int i = 0; i < 4; i++) {
        final Thread t = new Thread(task);
        t.setDaemon(true);
        t.start();
    }
    assertEquals(0, commitCount.get());
    // Index 100 documents with a storage filename of "0.0.prov"
    for (int i = 0; i < 100; i++) {
        final Document document = new Document();
        document.add(new LongField(SearchableFields.EventTime.getSearchableFieldName(), System.currentTimeMillis(), Store.NO));
        final StorageSummary location = new StorageSummary(1L, "0.0.prov", "1", 0, 1000L, 1000L);
        final StoredDocument storedDoc = new StoredDocument(document, location);
        docQueue.add(storedDoc);
    }
    assertEquals(0, commitCount.get());
    // Index 100 documents
    for (int i = 0; i < 100; i++) {
        final Document document = new Document();
        document.add(new LongField(SearchableFields.EventTime.getSearchableFieldName(), System.currentTimeMillis(), Store.NO));
        final StorageSummary location = new StorageSummary(1L, "0.0.prov", "1", 0, 1000L, 1000L);
        final StoredDocument storedDoc = new StoredDocument(document, location);
        docQueue.add(storedDoc);
    }
    // Wait until we've indexed all 200 events
    while (eventIndexWriter.getEventsIndexed() < 200) {
        Thread.sleep(10L);
    }
    // Wait a bit and make sure that we still haven't committed the index writer.
    Thread.sleep(100L);
    assertEquals(0, commitCount.get());
    // Add another document.
    final Document document = new Document();
    document.add(new LongField(SearchableFields.EventTime.getSearchableFieldName(), System.currentTimeMillis(), Store.NO));
    final StorageSummary location = new StorageSummary(1L, "0.0.prov", "1", 0, 1000L, 1000L);
    StoredDocument storedDoc = new StoredDocument(document, location);
    docQueue.add(storedDoc);
    // Wait until index writer is committed.
    while (commitCount.get() == 0) {
        Thread.sleep(10L);
    }
    assertEquals(1, commitCount.get());
    // Add a new IndexableDocument with a count of 1 to ensure that the writer is committed again.
    storedDoc = new StoredDocument(document, location);
    docQueue.add(storedDoc);
    Thread.sleep(100L);
    assertEquals(1, commitCount.get());
    // Add a new IndexableDocument with a count of 3. Index writer should not be committed again.
    storedDoc = new StoredDocument(document, location);
    docQueue.add(storedDoc);
    Thread.sleep(100L);
    assertEquals(1, commitCount.get());
}

Example 3 with IndexManager

use of org.apache.nifi.provenance.lucene.IndexManager in project nifi by apache.

the class TestLuceneEventIndex method testUnauthorizedEventsGetFilteredForQuery.

@Test(timeout = 60000)
public void testUnauthorizedEventsGetFilteredForQuery() throws InterruptedException {
    assumeFalse(isWindowsEnvironment());
    final RepositoryConfiguration repoConfig = createConfig(1);
    repoConfig.setDesiredIndexSize(1L);
    final IndexManager indexManager = new SimpleIndexManager(repoConfig);
    final ArrayListEventStore eventStore = new ArrayListEventStore();
    final LuceneEventIndex index = new LuceneEventIndex(repoConfig, indexManager, 3, EventReporter.NO_OP);
    index.initialize(eventStore);
    for (int i = 0; i < 3; i++) {
        final ProvenanceEventRecord event = createEvent("1234");
        final StorageResult storageResult = eventStore.addEvent(event);
        index.addEvents(storageResult.getStorageLocations());
    }
    final Query query = new Query(UUID.randomUUID().toString());
    final EventAuthorizer authorizer = new EventAuthorizer() {

        @Override
        public boolean isAuthorized(ProvenanceEventRecord event) {
            return event.getEventId() % 2 == 0;
        }

        @Override
        public void authorize(ProvenanceEventRecord event) throws AccessDeniedException {
            throw new AccessDeniedException();
        }
    };
    List<ProvenanceEventRecord> events = Collections.emptyList();
    while (events.size() < 2) {
        final QuerySubmission submission = index.submitQuery(query, authorizer, "unit test");
        assertTrue(submission.getResult().awaitCompletion(5, TimeUnit.SECONDS));
        events = submission.getResult().getMatchingEvents();
        Thread.sleep(25L);
    }
    assertEquals(2, events.size());
}

Example 4 with IndexManager

use of org.apache.nifi.provenance.lucene.IndexManager in project nifi by apache.

the class TestLuceneEventIndex method testUnauthorizedEventsGetPlaceholdersForFindParents.

@Test(timeout = 60000)
public void testUnauthorizedEventsGetPlaceholdersForFindParents() throws InterruptedException {
    assumeFalse(isWindowsEnvironment());
    final RepositoryConfiguration repoConfig = createConfig(1);
    repoConfig.setDesiredIndexSize(1L);
    final IndexManager indexManager = new SimpleIndexManager(repoConfig);
    final ArrayListEventStore eventStore = new ArrayListEventStore();
    final LuceneEventIndex index = new LuceneEventIndex(repoConfig, indexManager, 3, EventReporter.NO_OP);
    index.initialize(eventStore);
    final ProvenanceEventRecord firstEvent = createEvent("4444");
    final Map<String, String> previousAttributes = new HashMap<>();
    previousAttributes.put("uuid", "4444");
    final Map<String, String> updatedAttributes = new HashMap<>();
    updatedAttributes.put("updated", "true");
    final ProvenanceEventRecord join = new StandardProvenanceEventRecord.Builder().setEventType(ProvenanceEventType.JOIN).setAttributes(previousAttributes, updatedAttributes).addParentUuid("4444").addChildFlowFile("1234").setComponentId("component-1").setComponentType("unit test").setEventId(idGenerator.getAndIncrement()).setEventTime(System.currentTimeMillis()).setFlowFileEntryDate(System.currentTimeMillis()).setFlowFileUUID("1234").setLineageStartDate(System.currentTimeMillis()).setCurrentContentClaim("container", "section", "unit-test-id", 0L, 1024L).build();
    index.addEvents(eventStore.addEvent(firstEvent).getStorageLocations());
    index.addEvents(eventStore.addEvent(join).getStorageLocations());
    for (int i = 0; i < 3; i++) {
        final ProvenanceEventRecord event = createEvent("1234");
        final StorageResult storageResult = eventStore.addEvent(event);
        index.addEvents(storageResult.getStorageLocations());
    }
    final NiFiUser user = createUser();
    final EventAuthorizer allowJoinEvents = new EventAuthorizer() {

        @Override
        public boolean isAuthorized(ProvenanceEventRecord event) {
            return event.getEventType() == ProvenanceEventType.JOIN;
        }

        @Override
        public void authorize(ProvenanceEventRecord event) throws AccessDeniedException {
        }
    };
    List<LineageNode> nodes = Collections.emptyList();
    while (nodes.size() < 2) {
        final ComputeLineageSubmission submission = index.submitExpandParents(1L, user, allowJoinEvents);
        assertTrue(submission.getResult().awaitCompletion(5, TimeUnit.SECONDS));
        nodes = submission.getResult().getNodes();
        Thread.sleep(25L);
    }
    assertEquals(2, nodes.size());
    final Map<ProvenanceEventType, List<LineageNode>> eventMap = nodes.stream().filter(n -> n.getNodeType() == LineageNodeType.PROVENANCE_EVENT_NODE).collect(Collectors.groupingBy(n -> ((ProvenanceEventLineageNode) n).getEventType()));
    assertEquals(2, eventMap.size());
    assertEquals(1, eventMap.get(ProvenanceEventType.JOIN).size());
    assertEquals(1, eventMap.get(ProvenanceEventType.UNKNOWN).size());
    assertEquals("4444", eventMap.get(ProvenanceEventType.UNKNOWN).get(0).getFlowFileUuid());
}

Example 5 with IndexManager

use of org.apache.nifi.provenance.lucene.IndexManager in project nifi by apache.

the class TestLuceneEventIndex method testUnauthorizedEventsGetPlaceholdersForLineage.

@Test(timeout = 60000)
public void testUnauthorizedEventsGetPlaceholdersForLineage() throws InterruptedException {
    assumeFalse(isWindowsEnvironment());
    final RepositoryConfiguration repoConfig = createConfig(1);
    repoConfig.setDesiredIndexSize(1L);
    final IndexManager indexManager = new SimpleIndexManager(repoConfig);
    final ArrayListEventStore eventStore = new ArrayListEventStore();
    final LuceneEventIndex index = new LuceneEventIndex(repoConfig, indexManager, 3, EventReporter.NO_OP);
    index.initialize(eventStore);
    for (int i = 0; i < 3; i++) {
        final ProvenanceEventRecord event = createEvent("1234");
        final StorageResult storageResult = eventStore.addEvent(event);
        index.addEvents(storageResult.getStorageLocations());
    }
    final NiFiUser user = createUser();
    List<LineageNode> nodes = Collections.emptyList();
    while (nodes.size() < 3) {
        final ComputeLineageSubmission submission = index.submitLineageComputation(1L, user, EventAuthorizer.DENY_ALL);
        assertTrue(submission.getResult().awaitCompletion(5, TimeUnit.SECONDS));
        nodes = submission.getResult().getNodes();
        Thread.sleep(25L);
    }
    assertEquals(3, nodes.size());
    for (final LineageNode node : nodes) {
        assertEquals(LineageNodeType.PROVENANCE_EVENT_NODE, node.getNodeType());
        final ProvenanceEventLineageNode eventNode = (ProvenanceEventLineageNode) node;
        assertEquals(ProvenanceEventType.UNKNOWN, eventNode.getEventType());
    }
}