Search in sources :

Example 1 with WorkflowManager

use of co.cask.cdap.test.WorkflowManager in project cdap by caskdata.

the class AuthorizationTest method testScheduleAuth.

@Test
public void testScheduleAuth() throws Exception {
    createAuthNamespace();
    ApplicationId appId = AUTH_NAMESPACE.app(AppWithSchedule.class.getSimpleName());
    Map<EntityId, Set<Action>> neededPrivileges = ImmutableMap.<EntityId, Set<Action>>builder().put(appId, EnumSet.of(Action.ADMIN)).put(AUTH_NAMESPACE.artifact(AppWithSchedule.class.getSimpleName(), "1.0-SNAPSHOT"), EnumSet.of(Action.ADMIN)).put(AUTH_NAMESPACE.dataset(AppWithSchedule.INPUT_NAME), EnumSet.of(Action.ADMIN)).put(AUTH_NAMESPACE.dataset(AppWithSchedule.OUTPUT_NAME), EnumSet.of(Action.ADMIN)).put(AUTH_NAMESPACE.datasetType(ObjectStore.class.getName()), EnumSet.of(Action.ADMIN)).build();
    setUpPrivilegeAndRegisterForDeletion(ALICE, neededPrivileges);
    ApplicationManager appManager = deployApplication(AUTH_NAMESPACE, AppWithSchedule.class);
    String workflowName = AppWithSchedule.SampleWorkflow.class.getSimpleName();
    ProgramId workflowID = new ProgramId(AUTH_NAMESPACE.getNamespace(), AppWithSchedule.class.getSimpleName(), ProgramType.WORKFLOW, workflowName);
    cleanUpEntities.add(workflowID);
    final WorkflowManager workflowManager = appManager.getWorkflowManager(workflowName);
    ScheduleManager scheduleManager = workflowManager.getSchedule(AppWithSchedule.EVERY_HOUR_SCHEDULE);
    // switch to BOB
    SecurityRequestContext.setUserId(BOB.getName());
    // try to resume schedule as BOB. It should fail since BOB does not have execute privileges on the programs
    try {
        scheduleManager.resume();
        Assert.fail("Resuming schedule should have failed since BOB does not have EXECUTE on the program");
    } catch (UnauthorizedException e) {
    // Expected
    }
    // bob should also not be able see the status of the schedule
    try {
        scheduleManager.status(HttpURLConnection.HTTP_FORBIDDEN);
        Assert.fail("Getting schedule status should have failed since BOB does not have any privilege on the program");
    } catch (UnauthorizedException e) {
    // Expected
    }
    // give BOB READ permission in the workflow
    grantAndAssertSuccess(workflowID, BOB, EnumSet.of(Action.READ));
    // switch to BOB
    SecurityRequestContext.setUserId(BOB.getName());
    // try to resume schedule as BOB. It should fail since BOB has READ but not EXECUTE on the workflow
    try {
        scheduleManager.resume();
        Assert.fail("Resuming schedule should have failed since BOB does not have EXECUTE on the program");
    } catch (UnauthorizedException e) {
    // Expected
    }
    // but BOB should be able to get schedule status now
    Assert.assertEquals(ProgramScheduleStatus.SUSPENDED.name(), scheduleManager.status(HttpURLConnection.HTTP_OK));
    // give BOB EXECUTE permission in the workflow
    grantAndAssertSuccess(workflowID, BOB, EnumSet.of(Action.EXECUTE));
    // switch to BOB
    SecurityRequestContext.setUserId(BOB.getName());
    // try to resume the schedule. This should pass and workflow should run
    scheduleManager.resume();
    Assert.assertEquals(ProgramScheduleStatus.SCHEDULED.name(), scheduleManager.status(HttpURLConnection.HTTP_OK));
    // suspend the schedule so that it does not start running again
    scheduleManager.suspend();
    Assert.assertEquals(ProgramScheduleStatus.SUSPENDED.name(), scheduleManager.status(HttpURLConnection.HTTP_OK));
    ScheduleId scheduleId = new ScheduleId(appId.getNamespace(), appId.getApplication(), appId.getVersion(), "testSchedule");
    ScheduleDetail scheduleDetail = new ScheduleDetail(AUTH_NAMESPACE.getNamespace(), AppWithSchedule.class.getSimpleName(), "1.0-SNAPSHOT", "testSchedule", "Something 2", new ScheduleProgramInfo(SchedulableProgramType.WORKFLOW, workflowName), Collections.<String, String>emptyMap(), new TimeTrigger("*/1 * * * *"), Collections.<Constraint>emptyList(), TimeUnit.HOURS.toMillis(6), null);
    try {
        addSchedule(scheduleId, scheduleDetail);
        Assert.fail("Adding schedule should fail since BOB does not have AMDIN on the app");
    } catch (UnauthorizedException e) {
    // expected
    }
    // grant BOB ADMIN on the app
    grantAndAssertSuccess(appId, BOB, EnumSet.of(Action.ADMIN));
    // add schedule should succeed
    addSchedule(scheduleId, scheduleDetail);
    Assert.assertEquals(ProgramScheduleStatus.SUSPENDED.name(), workflowManager.getSchedule(scheduleId.getSchedule()).status(HttpURLConnection.HTTP_OK));
    // update schedule should succeed
    updateSchedule(scheduleId, scheduleDetail);
    Assert.assertEquals(ProgramScheduleStatus.SUSPENDED.name(), workflowManager.getSchedule(scheduleId.getSchedule()).status(HttpURLConnection.HTTP_OK));
    // revoke ADMIN from BOB
    getAuthorizer().revoke(Authorizable.fromEntityId(appId), BOB, EnumSet.of(Action.ADMIN));
    try {
        // delete schedule should fail since we revoke the ADMIN privilege from BOB
        deleteSchedule(scheduleId);
        Assert.fail("Deleting schedule should fail since BOB does not have AMDIN on the app");
    } catch (UnauthorizedException e) {
    // expected
    }
    try {
        updateSchedule(scheduleId, scheduleDetail);
        Assert.fail("Updating schedule should fail since BOB does not have AMDIN on the app");
    } catch (UnauthorizedException e) {
    // expected
    }
    // grant BOB ADMIN on the app again
    grantAndAssertSuccess(appId, BOB, EnumSet.of(Action.ADMIN));
    deleteSchedule(scheduleId);
    workflowManager.getSchedule(scheduleId.getSchedule()).status(HttpURLConnection.HTTP_NOT_FOUND);
    // switch to Alice
    SecurityRequestContext.setUserId(ALICE.getName());
}
Also used : ScheduleManager(co.cask.cdap.test.ScheduleManager) ObjectStore(co.cask.cdap.api.dataset.lib.ObjectStore) ApplicationManager(co.cask.cdap.test.ApplicationManager) EnumSet(java.util.EnumSet) Set(java.util.Set) ImmutableSet(com.google.common.collect.ImmutableSet) HashSet(java.util.HashSet) PartitionedFileSet(co.cask.cdap.api.dataset.lib.PartitionedFileSet) TimeTrigger(co.cask.cdap.internal.app.runtime.schedule.trigger.TimeTrigger) WorkflowManager(co.cask.cdap.test.WorkflowManager) AppWithSchedule(co.cask.cdap.test.app.AppWithSchedule) ProgramId(co.cask.cdap.proto.id.ProgramId) ScheduleId(co.cask.cdap.proto.id.ScheduleId) EntityId(co.cask.cdap.proto.id.EntityId) UnauthorizedException(co.cask.cdap.security.spi.authorization.UnauthorizedException) ScheduleDetail(co.cask.cdap.proto.ScheduleDetail) ApplicationId(co.cask.cdap.proto.id.ApplicationId) ScheduleProgramInfo(co.cask.cdap.api.workflow.ScheduleProgramInfo) Test(org.junit.Test)

Example 2 with WorkflowManager

use of co.cask.cdap.test.WorkflowManager in project cdap by caskdata.

the class SparkTest method testSparkProgramStatusSchedule.

@Test
public void testSparkProgramStatusSchedule() throws Exception {
    ApplicationManager appManager = deploy(TestSparkApp.class);
    ScheduleId scheduleId = new ScheduleId(NamespaceId.DEFAULT.getNamespace(), TestSparkApp.class.getSimpleName(), "schedule");
    appManager.enableSchedule(scheduleId);
    // Start the upstream program
    appManager.getSparkManager(TestSparkApp.ScalaClassicSpark.class.getSimpleName()).start();
    // Wait for the downstream to complete
    WorkflowManager workflowManager = appManager.getWorkflowManager(TestSparkApp.TriggeredWorkflow.class.getSimpleName());
    workflowManager.waitForRun(ProgramRunStatus.COMPLETED, 5, TimeUnit.MINUTES);
    // Run again with the kryo serializer
    appManager.getSparkManager(TestSparkApp.ScalaClassicSpark.class.getSimpleName()).start(Collections.singletonMap("spark.serializer", "org.apache.spark.serializer.KryoSerializer"));
    // Wait for the downstream to complete again
    workflowManager.waitForRuns(ProgramRunStatus.COMPLETED, 2, 5, TimeUnit.MINUTES);
}
Also used : ApplicationManager(co.cask.cdap.test.ApplicationManager) WorkflowManager(co.cask.cdap.test.WorkflowManager) TestSparkApp(co.cask.cdap.spark.app.TestSparkApp) ScheduleId(co.cask.cdap.proto.id.ScheduleId) Test(org.junit.Test)

Example 3 with WorkflowManager

use of co.cask.cdap.test.WorkflowManager in project cdap by caskdata.

the class WikipediaPipelineAppTest method test.

@Test
@Category(XSlowTests.class)
public void test() throws Exception {
    WikipediaPipelineApp.WikipediaAppConfig appConfig = new WikipediaPipelineApp.WikipediaAppConfig();
    AppRequest<WikipediaPipelineApp.WikipediaAppConfig> appRequest = new AppRequest<>(ARTIFACT_SUMMARY, appConfig);
    ApplicationManager appManager = deployApplication(APP_ID, appRequest);
    // Setup input streams with test data
    createTestData();
    WorkflowManager workflowManager = appManager.getWorkflowManager(WikipediaPipelineWorkflow.NAME);
    // Test with default threshold. Workflow should not proceed beyond first condition.
    testWorkflow(workflowManager, appConfig, 1);
    // Test with a reduced threshold, so the workflow proceeds beyond the first predicate
    testWorkflow(workflowManager, appConfig, 2, 1);
    // Test K-Means
    appConfig = new WikipediaPipelineApp.WikipediaAppConfig("kmeans");
    appRequest = new AppRequest<>(ARTIFACT_SUMMARY, appConfig);
    appManager = deployApplication(APP_ID, appRequest);
    workflowManager = appManager.getWorkflowManager(WikipediaPipelineWorkflow.NAME);
    testWorkflow(workflowManager, appConfig, 3, 1);
}
Also used : ApplicationManager(co.cask.cdap.test.ApplicationManager) WorkflowManager(co.cask.cdap.test.WorkflowManager) AppRequest(co.cask.cdap.proto.artifact.AppRequest) Category(org.junit.experimental.categories.Category) Test(org.junit.Test)

Example 4 with WorkflowManager

use of co.cask.cdap.test.WorkflowManager in project cdap by caskdata.

the class TestFrameworkTestRun method executeWorkflow.

private String executeWorkflow(ApplicationManager applicationManager, Map<String, String> additionalParams, int expectedComplete) throws Exception {
    WorkflowManager wfManager = applicationManager.getWorkflowManager(WorkflowAppWithLocalDatasets.WORKFLOW_NAME);
    Map<String, String> runtimeArgs = new HashMap<>();
    File waitFile = new File(TMP_FOLDER.newFolder(), "/wait.file");
    File doneFile = new File(TMP_FOLDER.newFolder(), "/done.file");
    runtimeArgs.put("input.path", "input");
    runtimeArgs.put("output.path", "output");
    runtimeArgs.put("wait.file", waitFile.getAbsolutePath());
    runtimeArgs.put("done.file", doneFile.getAbsolutePath());
    runtimeArgs.putAll(additionalParams);
    wfManager.start(runtimeArgs);
    // Wait until custom action in the Workflow is triggered.
    while (!waitFile.exists()) {
        TimeUnit.MILLISECONDS.sleep(50);
    }
    // Now the Workflow should have RUNNING status. Get its runid.
    List<RunRecord> history = wfManager.getHistory(ProgramRunStatus.RUNNING);
    Assert.assertEquals(1, history.size());
    String runId = history.get(0).getPid();
    // Get the local datasets for this Workflow run
    DataSetManager<KeyValueTable> localDataset = getDataset(testSpace.dataset(WorkflowAppWithLocalDatasets.WORDCOUNT_DATASET + "." + runId));
    Assert.assertEquals("2", Bytes.toString(localDataset.get().read("text")));
    DataSetManager<FileSet> fileSetDataset = getDataset(testSpace.dataset(WorkflowAppWithLocalDatasets.CSV_FILESET_DATASET + "." + runId));
    Assert.assertNotNull(fileSetDataset.get());
    // Local datasets should not exist at the namespace level
    localDataset = getDataset(testSpace.dataset(WorkflowAppWithLocalDatasets.WORDCOUNT_DATASET));
    Assert.assertNull(localDataset.get());
    fileSetDataset = getDataset(testSpace.dataset(WorkflowAppWithLocalDatasets.CSV_FILESET_DATASET));
    Assert.assertNull(fileSetDataset.get());
    // Verify that the workflow hasn't completed on its own before we signal it to
    history = wfManager.getHistory(ProgramRunStatus.RUNNING);
    Assert.assertEquals(1, history.size());
    // Signal the Workflow to continue
    doneFile.createNewFile();
    // Wait for workflow to finish
    wfManager.waitForRuns(ProgramRunStatus.COMPLETED, expectedComplete, 1, TimeUnit.MINUTES);
    Map<String, WorkflowNodeStateDetail> nodeStateDetailMap = wfManager.getWorkflowNodeStates(runId);
    Map<String, String> workflowMetricsContext = new HashMap<>();
    workflowMetricsContext.put(Constants.Metrics.Tag.NAMESPACE, testSpace.getNamespace());
    workflowMetricsContext.put(Constants.Metrics.Tag.APP, applicationManager.getInfo().getName());
    workflowMetricsContext.put(Constants.Metrics.Tag.WORKFLOW, WorkflowAppWithLocalDatasets.WORKFLOW_NAME);
    workflowMetricsContext.put(Constants.Metrics.Tag.RUN_ID, runId);
    Map<String, String> writerContext = new HashMap<>(workflowMetricsContext);
    writerContext.put(Constants.Metrics.Tag.NODE, WorkflowAppWithLocalDatasets.LocalDatasetWriter.class.getSimpleName());
    Assert.assertEquals(2, getMetricsManager().getTotalMetric(writerContext, "user.num.lines"));
    Map<String, String> wfSparkMetricsContext = new HashMap<>(workflowMetricsContext);
    wfSparkMetricsContext.put(Constants.Metrics.Tag.NODE, "JavaSparkCSVToSpaceConverter");
    Assert.assertEquals(2, getMetricsManager().getTotalMetric(wfSparkMetricsContext, "user.num.lines"));
    // check in spark context
    Map<String, String> sparkMetricsContext = new HashMap<>();
    sparkMetricsContext.put(Constants.Metrics.Tag.NAMESPACE, testSpace.getNamespace());
    sparkMetricsContext.put(Constants.Metrics.Tag.APP, applicationManager.getInfo().getName());
    sparkMetricsContext.put(Constants.Metrics.Tag.SPARK, "JavaSparkCSVToSpaceConverter");
    sparkMetricsContext.put(Constants.Metrics.Tag.RUN_ID, nodeStateDetailMap.get("JavaSparkCSVToSpaceConverter").getRunId());
    Assert.assertEquals(2, getMetricsManager().getTotalMetric(sparkMetricsContext, "user.num.lines"));
    Map<String, String> appMetricsContext = new HashMap<>();
    appMetricsContext.put(Constants.Metrics.Tag.NAMESPACE, testSpace.getNamespace());
    appMetricsContext.put(Constants.Metrics.Tag.APP, applicationManager.getInfo().getName());
    // app metrics context should have sum from custom action and spark metrics.
    Assert.assertEquals(4, getMetricsManager().getTotalMetric(appMetricsContext, "user.num.lines"));
    Map<String, String> wfMRMetricsContext = new HashMap<>(workflowMetricsContext);
    wfMRMetricsContext.put(Constants.Metrics.Tag.NODE, "WordCount");
    Assert.assertEquals(7, getMetricsManager().getTotalMetric(wfMRMetricsContext, "user.num.words"));
    // mr metrics context
    Map<String, String> mrMetricsContext = new HashMap<>();
    mrMetricsContext.put(Constants.Metrics.Tag.NAMESPACE, testSpace.getNamespace());
    mrMetricsContext.put(Constants.Metrics.Tag.APP, applicationManager.getInfo().getName());
    mrMetricsContext.put(Constants.Metrics.Tag.MAPREDUCE, "WordCount");
    mrMetricsContext.put(Constants.Metrics.Tag.RUN_ID, nodeStateDetailMap.get("WordCount").getRunId());
    Assert.assertEquals(7, getMetricsManager().getTotalMetric(mrMetricsContext, "user.num.words"));
    final Map<String, String> readerContext = new HashMap<>(workflowMetricsContext);
    readerContext.put(Constants.Metrics.Tag.NODE, "readerAction");
    Tasks.waitFor(6L, new Callable<Long>() {

        @Override
        public Long call() throws Exception {
            return getMetricsManager().getTotalMetric(readerContext, "user.unique.words");
        }
    }, 60, TimeUnit.SECONDS);
    return runId;
}
Also used : FileSet(co.cask.cdap.api.dataset.lib.FileSet) HashMap(java.util.HashMap) WorkflowManager(co.cask.cdap.test.WorkflowManager) ConflictException(co.cask.cdap.common.ConflictException) IOException(java.io.IOException) TimeoutException(java.util.concurrent.TimeoutException) WorkflowNodeStateDetail(co.cask.cdap.proto.WorkflowNodeStateDetail) RunRecord(co.cask.cdap.proto.RunRecord) KeyValueTable(co.cask.cdap.api.dataset.lib.KeyValueTable) File(java.io.File)

Example 5 with WorkflowManager

use of co.cask.cdap.test.WorkflowManager in project cdap by caskdata.

the class TestFrameworkTestRun method testWorkflowStatus.

@Test
public void testWorkflowStatus() throws Exception {
    ApplicationManager appManager = deployApplication(WorkflowStatusTestApp.class);
    File workflowSuccess = new File(TMP_FOLDER.newFolder() + "/workflow.success");
    File actionSuccess = new File(TMP_FOLDER.newFolder() + "/action.success");
    File workflowKilled = new File(TMP_FOLDER.newFolder() + "/workflow.killed");
    File firstFile = new File(TMP_FOLDER.newFolder() + "/first");
    File firstFileDone = new File(TMP_FOLDER.newFolder() + "/first.done");
    WorkflowManager workflowManager = appManager.getWorkflowManager(WorkflowStatusTestApp.WORKFLOW_NAME);
    workflowManager.start(ImmutableMap.of("workflow.success.file", workflowSuccess.getAbsolutePath(), "action.success.file", actionSuccess.getAbsolutePath(), "throw.exception", "true"));
    workflowManager.waitForRun(ProgramRunStatus.FAILED, 1, TimeUnit.MINUTES);
    // Since action and workflow failed the files should not exist
    Assert.assertFalse(workflowSuccess.exists());
    Assert.assertFalse(actionSuccess.exists());
    workflowManager.start(ImmutableMap.of("workflow.success.file", workflowSuccess.getAbsolutePath(), "action.success.file", actionSuccess.getAbsolutePath()));
    workflowManager.waitForRun(ProgramRunStatus.COMPLETED, 1, TimeUnit.MINUTES);
    Assert.assertTrue(workflowSuccess.exists());
    Assert.assertTrue(actionSuccess.exists());
    // Test the killed status
    workflowManager.start(ImmutableMap.of("workflow.killed.file", workflowKilled.getAbsolutePath(), "first.file", firstFile.getAbsolutePath(), "first.done.file", firstFileDone.getAbsolutePath(), "test.killed", "true"));
    verifyFileExists(Lists.newArrayList(firstFile));
    workflowManager.stop();
    workflowManager.waitForRun(ProgramRunStatus.KILLED, 1, TimeUnit.MINUTES);
    Assert.assertTrue(workflowKilled.exists());
}
Also used : ApplicationManager(co.cask.cdap.test.ApplicationManager) WorkflowManager(co.cask.cdap.test.WorkflowManager) File(java.io.File) Test(org.junit.Test)

Aggregations

WorkflowManager (co.cask.cdap.test.WorkflowManager)59 ApplicationManager (co.cask.cdap.test.ApplicationManager)57 ApplicationId (co.cask.cdap.proto.id.ApplicationId)46 ETLBatchConfig (co.cask.cdap.etl.proto.v2.ETLBatchConfig)44 ETLStage (co.cask.cdap.etl.proto.v2.ETLStage)44 KeyValueTable (co.cask.cdap.api.dataset.lib.KeyValueTable)42 AppRequest (co.cask.cdap.proto.artifact.AppRequest)41 Table (co.cask.cdap.api.dataset.table.Table)39 Test (org.junit.Test)35 StructuredRecord (co.cask.cdap.api.data.format.StructuredRecord)34 Schema (co.cask.cdap.api.data.schema.Schema)31 HashMap (java.util.HashMap)12 RunRecord (co.cask.cdap.proto.RunRecord)9 ArrayList (java.util.ArrayList)9 ETLPlugin (co.cask.cdap.etl.proto.v2.ETLPlugin)7 HashSet (java.util.HashSet)7 ConflictException (co.cask.cdap.common.ConflictException)6 File (java.io.File)6 IOException (java.io.IOException)6 TimeoutException (java.util.concurrent.TimeoutException)6