use of co.cask.cdap.app.preview.PreviewManager in project cdap by caskdata.
the class PreviewDataPipelineTest method testPreviewFailedRun.
private void testPreviewFailedRun(Engine engine) throws Exception {
PreviewManager previewManager = getPreviewManager();
String sourceTableName = "singleInput";
String sinkTableName = "singleOutput";
Schema schema = Schema.recordOf("testRecord", Schema.Field.of("name", Schema.of(Schema.Type.STRING)));
/*
* source --> transform -> sink
*/
ETLBatchConfig etlConfig = ETLBatchConfig.builder("* * * * *").addStage(new ETLStage("source", MockSource.getPlugin(sourceTableName, schema))).addStage(new ETLStage("transform", ExceptionTransform.getPlugin("name", "samuel"))).addStage(new ETLStage("sink", MockSink.getPlugin(sinkTableName))).addConnection("source", "transform").addConnection("transform", "sink").setNumOfRecordsPreview(100).setEngine(engine).build();
// Construct the preview config with the program name and program type.
PreviewConfig previewConfig = new PreviewConfig(SmartWorkflow.NAME, ProgramType.WORKFLOW, Collections.<String, String>emptyMap(), 10);
// Create the table for the mock source
addDatasetInstance(Table.class.getName(), sourceTableName, DatasetProperties.of(ImmutableMap.of("schema", schema.toString())));
DataSetManager<Table> inputManager = getDataset(NamespaceId.DEFAULT.dataset(sourceTableName));
StructuredRecord recordSamuel = StructuredRecord.builder(schema).set("name", "samuel").build();
StructuredRecord recordBob = StructuredRecord.builder(schema).set("name", "bob").build();
MockSource.writeInput(inputManager, "1", recordSamuel);
MockSource.writeInput(inputManager, "2", recordBob);
AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig, previewConfig);
// Start the preview and get the corresponding PreviewRunner.
ApplicationId previewId = previewManager.start(NamespaceId.DEFAULT, appRequest);
final PreviewRunner previewRunner = previewManager.getRunner(previewId);
// Wait for the preview status go into FAILED.
Tasks.waitFor(PreviewStatus.Status.RUN_FAILED, new Callable<PreviewStatus.Status>() {
@Override
public PreviewStatus.Status call() throws Exception {
PreviewStatus status = previewRunner.getStatus();
return status == null ? null : status.getStatus();
}
}, 5, TimeUnit.MINUTES);
// Get the data for stage "source" in the PreviewStore.
checkPreviewStore(previewRunner, "source", 2);
// Get the data for stage "transform" in the PreviewStore, should contain one less record than source.
checkPreviewStore(previewRunner, "transform", 1);
// Get the data for stage "sink" in the PreviewStore, should contain one less record than source.
checkPreviewStore(previewRunner, "sink", 1);
// Validate the metrics for preview
validateMetric(2, previewId, "source.records.in", previewRunner);
validateMetric(2, previewId, "source.records.out", previewRunner);
validateMetric(2, previewId, "transform.records.in", previewRunner);
validateMetric(1, previewId, "transform.records.out", previewRunner);
validateMetric(1, previewId, "sink.records.out", previewRunner);
validateMetric(1, previewId, "sink.records.in", previewRunner);
// Check the sink table is not created in the real space.
DataSetManager<Table> sinkManager = getDataset(sinkTableName);
Assert.assertNull(sinkManager.get());
deleteDatasetInstance(NamespaceId.DEFAULT.dataset(sourceTableName));
}
use of co.cask.cdap.app.preview.PreviewManager in project cdap by caskdata.
the class PreviewDataPipelineTest method testDataPipelinePreviewRun.
private void testDataPipelinePreviewRun(Engine engine) throws Exception {
PreviewManager previewManager = getPreviewManager();
String sourceTableName = "singleInput";
String sinkTableName = "singleOutput";
Schema schema = Schema.recordOf("testRecord", Schema.Field.of("name", Schema.of(Schema.Type.STRING)));
/*
* source --> transform -> sink
*/
ETLBatchConfig etlConfig = ETLBatchConfig.builder("* * * * *").addStage(new ETLStage("source", MockSource.getPlugin(sourceTableName, schema))).addStage(new ETLStage("transform", IdentityTransform.getPlugin())).addStage(new ETLStage("sink", MockSink.getPlugin(sinkTableName))).addConnection("source", "transform").addConnection("transform", "sink").setEngine(engine).setNumOfRecordsPreview(100).build();
// Construct the preview config with the program name and program type
PreviewConfig previewConfig = new PreviewConfig(SmartWorkflow.NAME, ProgramType.WORKFLOW, Collections.<String, String>emptyMap(), 10);
// Create the table for the mock source
addDatasetInstance(Table.class.getName(), sourceTableName, DatasetProperties.of(ImmutableMap.of("schema", schema.toString())));
DataSetManager<Table> inputManager = getDataset(NamespaceId.DEFAULT.dataset(sourceTableName));
StructuredRecord recordSamuel = StructuredRecord.builder(schema).set("name", "samuel").build();
StructuredRecord recordBob = StructuredRecord.builder(schema).set("name", "bob").build();
MockSource.writeInput(inputManager, ImmutableList.of(recordSamuel, recordBob));
AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT_RANGE, etlConfig, previewConfig);
// Start the preview and get the corresponding PreviewRunner.
ApplicationId previewId = previewManager.start(NamespaceId.DEFAULT, appRequest);
final PreviewRunner previewRunner = previewManager.getRunner(previewId);
// Wait for the preview status go into COMPLETED.
Tasks.waitFor(PreviewStatus.Status.COMPLETED, new Callable<PreviewStatus.Status>() {
@Override
public PreviewStatus.Status call() throws Exception {
PreviewStatus status = previewRunner.getStatus();
return status == null ? null : status.getStatus();
}
}, 5, TimeUnit.MINUTES);
// Get the data for stage "source" in the PreviewStore, should contain two records.
checkPreviewStore(previewRunner, "source", 2);
// Get the data for stage "transform" in the PreviewStore, should contain two records.
checkPreviewStore(previewRunner, "transform", 2);
// Get the data for stage "sink" in the PreviewStore, should contain two records.
checkPreviewStore(previewRunner, "sink", 2);
// Validate the metrics for preview
validateMetric(2, previewId, "source.records.in", previewRunner);
validateMetric(2, previewId, "source.records.out", previewRunner);
validateMetric(2, previewId, "transform.records.in", previewRunner);
validateMetric(2, previewId, "transform.records.out", previewRunner);
validateMetric(2, previewId, "sink.records.out", previewRunner);
validateMetric(2, previewId, "sink.records.in", previewRunner);
// Check the sink table is not created in the real space.
DataSetManager<Table> sinkManager = getDataset(sinkTableName);
Assert.assertNull(sinkManager.get());
deleteDatasetInstance(NamespaceId.DEFAULT.dataset(sourceTableName));
Assert.assertNotNull(previewRunner.getRunRecord());
}
use of co.cask.cdap.app.preview.PreviewManager in project cdap by caskdata.
the class PreviewDataPipelineTest method testMultiplePhase.
private void testMultiplePhase(Engine engine) throws Exception {
/*
* source1 ----> t1 ------
* | --> innerjoin ----> t4 ------
* source2 ----> t2 ------ |
* | ---> outerjoin --> sink1
* |
* source3 -------------------- t3 ------------------------
*/
PreviewManager previewManager = getPreviewManager();
Schema inputSchema1 = Schema.recordOf("customerRecord", Schema.Field.of("customer_id", Schema.of(Schema.Type.STRING)), Schema.Field.of("customer_name", Schema.of(Schema.Type.STRING)));
Schema inputSchema2 = Schema.recordOf("itemRecord", Schema.Field.of("item_id", Schema.of(Schema.Type.STRING)), Schema.Field.of("item_price", Schema.of(Schema.Type.LONG)), Schema.Field.of("cust_id", Schema.of(Schema.Type.STRING)), Schema.Field.of("cust_name", Schema.of(Schema.Type.STRING)));
Schema inputSchema3 = Schema.recordOf("transactionRecord", Schema.Field.of("t_id", Schema.of(Schema.Type.STRING)), Schema.Field.of("c_id", Schema.of(Schema.Type.STRING)), Schema.Field.of("i_id", Schema.of(Schema.Type.STRING)));
Schema outSchema2 = Schema.recordOf("join.output", Schema.Field.of("t_id", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("c_id", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("i_id", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("customer_id", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("customer_name", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("item_id", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("item_price", Schema.nullableOf(Schema.of(Schema.Type.LONG))), Schema.Field.of("cust_id", Schema.nullableOf(Schema.of(Schema.Type.STRING))), Schema.Field.of("cust_name", Schema.nullableOf(Schema.of(Schema.Type.STRING))));
String source1MulitJoinInput = "multiJoinSource1-" + engine;
String source2MultiJoinInput = "multiJoinSource2-" + engine;
String source3MultiJoinInput = "multiJoinSource3-" + engine;
String outputName = "multiJoinOutput-" + engine;
String sinkName = "multiJoinOutputSink-" + engine;
String outerJoinName = "multiJoinOuter-" + engine;
ETLBatchConfig etlConfig = ETLBatchConfig.builder("* * * * *").addStage(new ETLStage("source1", MockSource.getPlugin(source1MulitJoinInput, inputSchema1))).addStage(new ETLStage("source2", MockSource.getPlugin(source2MultiJoinInput, inputSchema2))).addStage(new ETLStage("source3", MockSource.getPlugin(source3MultiJoinInput, inputSchema3))).addStage(new ETLStage("t1", IdentityTransform.getPlugin())).addStage(new ETLStage("t2", IdentityTransform.getPlugin())).addStage(new ETLStage("t3", IdentityTransform.getPlugin())).addStage(new ETLStage("t4", IdentityTransform.getPlugin())).addStage(new ETLStage("innerjoin", MockJoiner.getPlugin("t1.customer_id=t2.cust_id", "t1,t2", ""))).addStage(new ETLStage(outerJoinName, MockJoiner.getPlugin("t4.item_id=t3.i_id", "", ""))).addStage(new ETLStage(sinkName, MockSink.getPlugin(outputName))).addConnection("source1", "t1").addConnection("source2", "t2").addConnection("source3", "t3").addConnection("t1", "innerjoin").addConnection("t2", "innerjoin").addConnection("innerjoin", "t4").addConnection("t3", outerJoinName).addConnection("t4", outerJoinName).addConnection(outerJoinName, sinkName).setEngine(engine).setNumOfRecordsPreview(100).build();
// Construct the preview config with the program name and program type
PreviewConfig previewConfig = new PreviewConfig(SmartWorkflow.NAME, ProgramType.WORKFLOW, Collections.<String, String>emptyMap(), 10);
// Create the table for the mock source
addDatasetInstance(Table.class.getName(), source1MulitJoinInput, DatasetProperties.of(ImmutableMap.of("schema", inputSchema1.toString())));
addDatasetInstance(Table.class.getName(), source2MultiJoinInput, DatasetProperties.of(ImmutableMap.of("schema", inputSchema2.toString())));
addDatasetInstance(Table.class.getName(), source3MultiJoinInput, DatasetProperties.of(ImmutableMap.of("schema", inputSchema3.toString())));
AppRequest<ETLBatchConfig> appRequest = new AppRequest<>(APP_ARTIFACT, etlConfig, previewConfig);
// Start the preview and get the corresponding PreviewRunner.
ApplicationId previewId = previewManager.start(NamespaceId.DEFAULT, appRequest);
final PreviewRunner previewRunner = previewManager.getRunner(previewId);
ingestData(inputSchema1, inputSchema2, inputSchema3, source1MulitJoinInput, source2MultiJoinInput, source3MultiJoinInput);
// Wait for the preview status go into COMPLETED.
Tasks.waitFor(PreviewStatus.Status.COMPLETED, new Callable<PreviewStatus.Status>() {
@Override
public PreviewStatus.Status call() throws Exception {
PreviewStatus status = previewRunner.getStatus();
return status == null ? null : status.getStatus();
}
}, 5, TimeUnit.MINUTES);
checkPreviewStore(previewRunner, sinkName, 3);
validateMetric(3L, previewId, sinkName + ".records.in", previewRunner);
}
use of co.cask.cdap.app.preview.PreviewManager in project cdap by caskdata.
the class TestBase method initialize.
@BeforeClass
public static void initialize() throws Exception {
if (nestedStartCount++ > 0) {
return;
}
File localDataDir = TMP_FOLDER.newFolder();
cConf = createCConf(localDataDir);
org.apache.hadoop.conf.Configuration hConf = new org.apache.hadoop.conf.Configuration();
hConf.addResource("mapred-site-local.xml");
hConf.reloadConfiguration();
hConf.set(Constants.CFG_LOCAL_DATA_DIR, localDataDir.getAbsolutePath());
hConf.set(Constants.AppFabric.OUTPUT_DIR, cConf.get(Constants.AppFabric.OUTPUT_DIR));
hConf.set("hadoop.tmp.dir", new File(localDataDir, cConf.get(Constants.AppFabric.TEMP_DIR)).getAbsolutePath());
// Windows specific requirements
if (OSDetector.isWindows()) {
File tmpDir = TMP_FOLDER.newFolder();
File binDir = new File(tmpDir, "bin");
Assert.assertTrue(binDir.mkdirs());
copyTempFile("hadoop.dll", tmpDir);
copyTempFile("winutils.exe", binDir);
System.setProperty("hadoop.home.dir", tmpDir.getAbsolutePath());
System.load(new File(tmpDir, "hadoop.dll").getAbsolutePath());
}
Injector injector = Guice.createInjector(createDataFabricModule(), new TransactionExecutorModule(), new DataSetsModules().getStandaloneModules(), new DataSetServiceModules().getInMemoryModules(), new ConfigModule(cConf, hConf), new IOModule(), new LocationRuntimeModule().getInMemoryModules(), new DiscoveryRuntimeModule().getInMemoryModules(), new AppFabricServiceRuntimeModule().getInMemoryModules(), new ServiceStoreModules().getInMemoryModules(), new InMemoryProgramRunnerModule(LocalStreamWriter.class), new SecureStoreModules().getInMemoryModules(), new AbstractModule() {
@Override
protected void configure() {
bind(StreamHandler.class).in(Scopes.SINGLETON);
bind(StreamFetchHandler.class).in(Scopes.SINGLETON);
bind(StreamViewHttpHandler.class).in(Scopes.SINGLETON);
bind(StreamFileJanitorService.class).to(LocalStreamFileJanitorService.class).in(Scopes.SINGLETON);
bind(StreamWriterSizeCollector.class).to(BasicStreamWriterSizeCollector.class).in(Scopes.SINGLETON);
bind(StreamCoordinatorClient.class).to(InMemoryStreamCoordinatorClient.class).in(Scopes.SINGLETON);
bind(MetricsManager.class).toProvider(MetricsManagerProvider.class);
}
}, // todo: do we need handler?
new MetricsHandlerModule(), new MetricsClientRuntimeModule().getInMemoryModules(), new LoggingModules().getInMemoryModules(), new LogReaderRuntimeModules().getInMemoryModules(), new ExploreRuntimeModule().getInMemoryModules(), new ExploreClientModule(), new NotificationFeedServiceRuntimeModule().getInMemoryModules(), new NotificationServiceRuntimeModule().getInMemoryModules(), new NamespaceStoreModule().getStandaloneModules(), new AuthorizationModule(), new AuthorizationEnforcementModule().getInMemoryModules(), new MessagingServerRuntimeModule().getInMemoryModules(), new PreviewHttpModule(), new AbstractModule() {
@Override
@SuppressWarnings("deprecation")
protected void configure() {
install(new FactoryModuleBuilder().implement(ApplicationManager.class, DefaultApplicationManager.class).build(ApplicationManagerFactory.class));
install(new FactoryModuleBuilder().implement(ArtifactManager.class, DefaultArtifactManager.class).build(ArtifactManagerFactory.class));
install(new FactoryModuleBuilder().implement(StreamManager.class, DefaultStreamManager.class).build(StreamManagerFactory.class));
bind(TemporaryFolder.class).toInstance(TMP_FOLDER);
bind(AuthorizationHandler.class).in(Scopes.SINGLETON);
}
});
messagingService = injector.getInstance(MessagingService.class);
if (messagingService instanceof Service) {
((Service) messagingService).startAndWait();
}
txService = injector.getInstance(TransactionManager.class);
txService.startAndWait();
dsOpService = injector.getInstance(DatasetOpExecutor.class);
dsOpService.startAndWait();
datasetService = injector.getInstance(DatasetService.class);
datasetService.startAndWait();
metricsQueryService = injector.getInstance(MetricsQueryService.class);
metricsQueryService.startAndWait();
metricsCollectionService = injector.getInstance(MetricsCollectionService.class);
metricsCollectionService.startAndWait();
programLifecycleService = injector.getInstance(ProgramLifecycleService.class);
programLifecycleService.startAndWait();
programNotificationSubscriberService = injector.getInstance(ProgramNotificationSubscriberService.class);
programNotificationSubscriberService.startAndWait();
scheduler = injector.getInstance(Scheduler.class);
if (scheduler instanceof Service) {
((Service) scheduler).startAndWait();
}
if (scheduler instanceof CoreSchedulerService) {
((CoreSchedulerService) scheduler).waitUntilFunctional(10, TimeUnit.SECONDS);
}
if (cConf.getBoolean(Constants.Explore.EXPLORE_ENABLED)) {
exploreExecutorService = injector.getInstance(ExploreExecutorService.class);
exploreExecutorService.startAndWait();
// wait for explore service to be discoverable
DiscoveryServiceClient discoveryService = injector.getInstance(DiscoveryServiceClient.class);
EndpointStrategy endpointStrategy = new RandomEndpointStrategy(discoveryService.discover(Constants.Service.EXPLORE_HTTP_USER_SERVICE));
Preconditions.checkNotNull(endpointStrategy.pick(5, TimeUnit.SECONDS), "%s service is not up after 5 seconds", Constants.Service.EXPLORE_HTTP_USER_SERVICE);
exploreClient = injector.getInstance(ExploreClient.class);
}
streamCoordinatorClient = injector.getInstance(StreamCoordinatorClient.class);
streamCoordinatorClient.startAndWait();
programScheduler = injector.getInstance(Scheduler.class);
if (programScheduler instanceof Service) {
((Service) programScheduler).startAndWait();
}
testManager = injector.getInstance(UnitTestManager.class);
metricsManager = injector.getInstance(MetricsManager.class);
authorizerInstantiator = injector.getInstance(AuthorizerInstantiator.class);
// This is needed so the logged-in user can successfully create the default namespace
if (cConf.getBoolean(Constants.Security.Authorization.ENABLED)) {
String user = System.getProperty("user.name");
SecurityRequestContext.setUserId(user);
InstanceId instance = new InstanceId(cConf.get(Constants.INSTANCE_NAME));
Principal principal = new Principal(user, Principal.PrincipalType.USER);
authorizerInstantiator.get().grant(Authorizable.fromEntityId(instance), principal, ImmutableSet.of(Action.ADMIN));
authorizerInstantiator.get().grant(Authorizable.fromEntityId(NamespaceId.DEFAULT), principal, ImmutableSet.of(Action.ADMIN));
}
namespaceAdmin = injector.getInstance(NamespaceAdmin.class);
if (firstInit) {
// only create the default namespace on first test. if multiple tests are run in the same JVM,
// then any time after the first time, the default namespace already exists. That is because
// the namespaceAdmin.delete(Id.Namespace.DEFAULT) in finish() only clears the default namespace
// but does not remove it entirely
namespaceAdmin.create(NamespaceMeta.DEFAULT);
}
secureStore = injector.getInstance(SecureStore.class);
secureStoreManager = injector.getInstance(SecureStoreManager.class);
messagingContext = new MultiThreadMessagingContext(messagingService);
firstInit = false;
previewManager = injector.getInstance(PreviewManager.class);
}
use of co.cask.cdap.app.preview.PreviewManager in project cdap by caskdata.
the class DefaultPreviewManagerTest method testInjector.
@Test
public void testInjector() throws Exception {
PreviewManager previewManager = getInjector().getInstance(PreviewManager.class);
DefaultPreviewManager defaultPreviewManager = (DefaultPreviewManager) previewManager;
Injector previewInjector = defaultPreviewManager.createPreviewInjector(new ApplicationId("ns1", "app1"));
// Make sure same PreviewManager instance is returned for a same preview
Assert.assertEquals(previewInjector.getInstance(PreviewRunner.class), previewInjector.getInstance(PreviewRunner.class));
Injector anotherPreviewInjector = defaultPreviewManager.createPreviewInjector(new ApplicationId("ns2", "app2"));
Assert.assertNotEquals(previewInjector.getInstance(PreviewRunner.class), anotherPreviewInjector.getInstance(PreviewRunner.class));
}
Aggregations