Search in sources :

Example 96 with ContainerId

use of org.apache.hadoop.yarn.api.records.ContainerId in project hadoop by apache.

the class TestProtocolRecords method testRegisterNodeManagerRequest.

@Test
public void testRegisterNodeManagerRequest() {
    ApplicationId appId = ApplicationId.newInstance(123456789, 1);
    ApplicationAttemptId attemptId = ApplicationAttemptId.newInstance(appId, 1);
    ContainerId containerId = ContainerId.newContainerId(attemptId, 1);
    NMContainerStatus containerReport = NMContainerStatus.newInstance(containerId, 0, ContainerState.RUNNING, Resource.newInstance(1024, 1), "diagnostics", 0, Priority.newInstance(10), 1234);
    List<NMContainerStatus> reports = Arrays.asList(containerReport);
    RegisterNodeManagerRequest request = RegisterNodeManagerRequest.newInstance(NodeId.newInstance("1.1.1.1", 1000), 8080, Resource.newInstance(1024, 1), "NM-version-id", reports, Arrays.asList(appId));
    RegisterNodeManagerRequest requestProto = new RegisterNodeManagerRequestPBImpl(((RegisterNodeManagerRequestPBImpl) request).getProto());
    Assert.assertEquals(containerReport, requestProto.getNMContainerStatuses().get(0));
    Assert.assertEquals(8080, requestProto.getHttpPort());
    Assert.assertEquals("NM-version-id", requestProto.getNMVersion());
    Assert.assertEquals(NodeId.newInstance("1.1.1.1", 1000), requestProto.getNodeId());
    Assert.assertEquals(Resource.newInstance(1024, 1), requestProto.getResource());
    Assert.assertEquals(1, requestProto.getRunningApplications().size());
    Assert.assertEquals(appId, requestProto.getRunningApplications().get(0));
}
Also used : ContainerId(org.apache.hadoop.yarn.api.records.ContainerId) RegisterNodeManagerRequestPBImpl(org.apache.hadoop.yarn.server.api.protocolrecords.impl.pb.RegisterNodeManagerRequestPBImpl) ApplicationAttemptId(org.apache.hadoop.yarn.api.records.ApplicationAttemptId) ApplicationId(org.apache.hadoop.yarn.api.records.ApplicationId) Test(org.junit.Test)

Example 97 with ContainerId

use of org.apache.hadoop.yarn.api.records.ContainerId in project hadoop by apache.

the class LinuxContainerExecutor method launchContainer.

@Override
public int launchContainer(ContainerStartContext ctx) throws IOException {
    Container container = ctx.getContainer();
    Path nmPrivateContainerScriptPath = ctx.getNmPrivateContainerScriptPath();
    Path nmPrivateTokensPath = ctx.getNmPrivateTokensPath();
    String user = ctx.getUser();
    String appId = ctx.getAppId();
    Path containerWorkDir = ctx.getContainerWorkDir();
    List<String> localDirs = ctx.getLocalDirs();
    List<String> logDirs = ctx.getLogDirs();
    List<String> filecacheDirs = ctx.getFilecacheDirs();
    List<String> userLocalDirs = ctx.getUserLocalDirs();
    List<String> containerLocalDirs = ctx.getContainerLocalDirs();
    List<String> containerLogDirs = ctx.getContainerLogDirs();
    Map<Path, List<String>> localizedResources = ctx.getLocalizedResources();
    verifyUsernamePattern(user);
    String runAsUser = getRunAsUser(user);
    ContainerId containerId = container.getContainerId();
    String containerIdStr = containerId.toString();
    resourcesHandler.preExecute(containerId, container.getResource());
    String resourcesOptions = resourcesHandler.getResourcesOption(containerId);
    String tcCommandFile = null;
    try {
        if (resourceHandlerChain != null) {
            List<PrivilegedOperation> ops = resourceHandlerChain.preStart(container);
            if (ops != null) {
                List<PrivilegedOperation> resourceOps = new ArrayList<>();
                resourceOps.add(new PrivilegedOperation(PrivilegedOperation.OperationType.ADD_PID_TO_CGROUP, resourcesOptions));
                for (PrivilegedOperation op : ops) {
                    switch(op.getOperationType()) {
                        case ADD_PID_TO_CGROUP:
                            resourceOps.add(op);
                            break;
                        case TC_MODIFY_STATE:
                            tcCommandFile = op.getArguments().get(0);
                            break;
                        default:
                            LOG.warn("PrivilegedOperation type unsupported in launch: " + op.getOperationType());
                    }
                }
                if (resourceOps.size() > 1) {
                    //squash resource operations
                    try {
                        PrivilegedOperation operation = PrivilegedOperationExecutor.squashCGroupOperations(resourceOps);
                        resourcesOptions = operation.getArguments().get(0);
                    } catch (PrivilegedOperationException e) {
                        LOG.error("Failed to squash cgroup operations!", e);
                        throw new ResourceHandlerException("Failed to squash cgroup operations!");
                    }
                }
            }
        }
    } catch (ResourceHandlerException e) {
        LOG.error("ResourceHandlerChain.preStart() failed!", e);
        throw new IOException("ResourceHandlerChain.preStart() failed!", e);
    }
    try {
        Path pidFilePath = getPidFilePath(containerId);
        if (pidFilePath != null) {
            List<String> prefixCommands = new ArrayList<>();
            ContainerRuntimeContext.Builder builder = new ContainerRuntimeContext.Builder(container);
            addSchedPriorityCommand(prefixCommands);
            if (prefixCommands.size() > 0) {
                builder.setExecutionAttribute(CONTAINER_LAUNCH_PREFIX_COMMANDS, prefixCommands);
            }
            builder.setExecutionAttribute(LOCALIZED_RESOURCES, localizedResources).setExecutionAttribute(RUN_AS_USER, runAsUser).setExecutionAttribute(USER, user).setExecutionAttribute(APPID, appId).setExecutionAttribute(CONTAINER_ID_STR, containerIdStr).setExecutionAttribute(CONTAINER_WORK_DIR, containerWorkDir).setExecutionAttribute(NM_PRIVATE_CONTAINER_SCRIPT_PATH, nmPrivateContainerScriptPath).setExecutionAttribute(NM_PRIVATE_TOKENS_PATH, nmPrivateTokensPath).setExecutionAttribute(PID_FILE_PATH, pidFilePath).setExecutionAttribute(LOCAL_DIRS, localDirs).setExecutionAttribute(LOG_DIRS, logDirs).setExecutionAttribute(FILECACHE_DIRS, filecacheDirs).setExecutionAttribute(USER_LOCAL_DIRS, userLocalDirs).setExecutionAttribute(CONTAINER_LOCAL_DIRS, containerLocalDirs).setExecutionAttribute(CONTAINER_LOG_DIRS, containerLogDirs).setExecutionAttribute(RESOURCES_OPTIONS, resourcesOptions);
            if (tcCommandFile != null) {
                builder.setExecutionAttribute(TC_COMMAND_FILE, tcCommandFile);
            }
            linuxContainerRuntime.launchContainer(builder.build());
        } else {
            LOG.info("Container was marked as inactive. Returning terminated error");
            return ExitCode.TERMINATED.getExitCode();
        }
    } catch (ContainerExecutionException e) {
        int exitCode = e.getExitCode();
        LOG.warn("Exit code from container " + containerId + " is : " + exitCode);
        // output
        if (exitCode != ExitCode.FORCE_KILLED.getExitCode() && exitCode != ExitCode.TERMINATED.getExitCode()) {
            LOG.warn("Exception from container-launch with container ID: " + containerId + " and exit code: " + exitCode, e);
            StringBuilder builder = new StringBuilder();
            builder.append("Exception from container-launch.\n");
            builder.append("Container id: " + containerId + "\n");
            builder.append("Exit code: " + exitCode + "\n");
            if (!Optional.fromNullable(e.getErrorOutput()).or("").isEmpty()) {
                builder.append("Exception message: " + e.getErrorOutput() + "\n");
            }
            builder.append("Stack trace: " + StringUtils.stringifyException(e) + "\n");
            if (!e.getOutput().isEmpty()) {
                builder.append("Shell output: " + e.getOutput() + "\n");
            }
            String diagnostics = builder.toString();
            logOutput(diagnostics);
            container.handle(new ContainerDiagnosticsUpdateEvent(containerId, diagnostics));
        } else {
            container.handle(new ContainerDiagnosticsUpdateEvent(containerId, "Container killed on request. Exit code is " + exitCode));
        }
        return exitCode;
    } finally {
        resourcesHandler.postExecute(containerId);
        try {
            if (resourceHandlerChain != null) {
                resourceHandlerChain.postComplete(containerId);
            }
        } catch (ResourceHandlerException e) {
            LOG.warn("ResourceHandlerChain.postComplete failed for " + "containerId: " + containerId + ". Exception: " + e);
        }
    }
    return 0;
}
Also used : Path(org.apache.hadoop.fs.Path) ArrayList(java.util.ArrayList) ContainerDiagnosticsUpdateEvent(org.apache.hadoop.yarn.server.nodemanager.containermanager.container.ContainerDiagnosticsUpdateEvent) IOException(java.io.IOException) ContainerRuntimeContext(org.apache.hadoop.yarn.server.nodemanager.containermanager.runtime.ContainerRuntimeContext) ResourceHandlerException(org.apache.hadoop.yarn.server.nodemanager.containermanager.linux.resources.ResourceHandlerException) Container(org.apache.hadoop.yarn.server.nodemanager.containermanager.container.Container) ContainerExecutionException(org.apache.hadoop.yarn.server.nodemanager.containermanager.runtime.ContainerExecutionException) ContainerId(org.apache.hadoop.yarn.api.records.ContainerId) PrivilegedOperationException(org.apache.hadoop.yarn.server.nodemanager.containermanager.linux.privileged.PrivilegedOperationException) ArrayList(java.util.ArrayList) List(java.util.List) PrivilegedOperation(org.apache.hadoop.yarn.server.nodemanager.containermanager.linux.privileged.PrivilegedOperation)

Example 98 with ContainerId

use of org.apache.hadoop.yarn.api.records.ContainerId in project hadoop by apache.

the class TestYarnServerApiClasses method getIncreasedContainer.

private Container getIncreasedContainer(int containerID, int appAttemptId, int memory, int vCores) {
    ContainerId containerId = getContainerId(containerID, appAttemptId);
    Resource capability = Resource.newInstance(memory, vCores);
    return Container.newInstance(containerId, null, null, capability, null, null);
}
Also used : ContainerId(org.apache.hadoop.yarn.api.records.ContainerId) Resource(org.apache.hadoop.yarn.api.records.Resource)

Example 99 with ContainerId

use of org.apache.hadoop.yarn.api.records.ContainerId in project hadoop by apache.

the class TestNMLeveldbStateStoreService method testContainerStorage.

@Test
public void testContainerStorage() throws IOException {
    // test empty when no state
    List<RecoveredContainerState> recoveredContainers = stateStore.loadContainersState();
    assertTrue(recoveredContainers.isEmpty());
    // create a container request
    ApplicationId appId = ApplicationId.newInstance(1234, 3);
    ApplicationAttemptId appAttemptId = ApplicationAttemptId.newInstance(appId, 4);
    ContainerId containerId = ContainerId.newContainerId(appAttemptId, 5);
    StartContainerRequest containerReq = createContainerRequest(containerId);
    // store a container and verify recovered
    stateStore.storeContainer(containerId, 0, containerReq);
    // verify the container version key is not stored for new containers
    DB db = stateStore.getDB();
    assertNull("version key present for new container", db.get(bytes(stateStore.getContainerVersionKey(containerId.toString()))));
    restartStateStore();
    recoveredContainers = stateStore.loadContainersState();
    assertEquals(1, recoveredContainers.size());
    RecoveredContainerState rcs = recoveredContainers.get(0);
    assertEquals(0, rcs.getVersion());
    assertEquals(RecoveredContainerStatus.REQUESTED, rcs.getStatus());
    assertEquals(ContainerExitStatus.INVALID, rcs.getExitCode());
    assertEquals(false, rcs.getKilled());
    assertEquals(containerReq, rcs.getStartRequest());
    assertTrue(rcs.getDiagnostics().isEmpty());
    // store a new container record without StartContainerRequest
    ContainerId containerId1 = ContainerId.newContainerId(appAttemptId, 6);
    stateStore.storeContainerLaunched(containerId1);
    recoveredContainers = stateStore.loadContainersState();
    // check whether the new container record is discarded
    assertEquals(1, recoveredContainers.size());
    // queue the container, and verify recovered
    stateStore.storeContainerQueued(containerId);
    restartStateStore();
    recoveredContainers = stateStore.loadContainersState();
    assertEquals(1, recoveredContainers.size());
    rcs = recoveredContainers.get(0);
    assertEquals(RecoveredContainerStatus.QUEUED, rcs.getStatus());
    assertEquals(ContainerExitStatus.INVALID, rcs.getExitCode());
    assertEquals(false, rcs.getKilled());
    assertEquals(containerReq, rcs.getStartRequest());
    assertTrue(rcs.getDiagnostics().isEmpty());
    // launch the container, add some diagnostics, and verify recovered
    StringBuilder diags = new StringBuilder();
    stateStore.storeContainerLaunched(containerId);
    diags.append("some diags for container");
    stateStore.storeContainerDiagnostics(containerId, diags);
    restartStateStore();
    recoveredContainers = stateStore.loadContainersState();
    assertEquals(1, recoveredContainers.size());
    rcs = recoveredContainers.get(0);
    assertEquals(RecoveredContainerStatus.LAUNCHED, rcs.getStatus());
    assertEquals(ContainerExitStatus.INVALID, rcs.getExitCode());
    assertEquals(false, rcs.getKilled());
    assertEquals(containerReq, rcs.getStartRequest());
    assertEquals(diags.toString(), rcs.getDiagnostics());
    // increase the container size, and verify recovered
    stateStore.storeContainerResourceChanged(containerId, 2, Resource.newInstance(2468, 4));
    restartStateStore();
    recoveredContainers = stateStore.loadContainersState();
    assertEquals(1, recoveredContainers.size());
    rcs = recoveredContainers.get(0);
    assertEquals(2, rcs.getVersion());
    assertEquals(RecoveredContainerStatus.LAUNCHED, rcs.getStatus());
    assertEquals(ContainerExitStatus.INVALID, rcs.getExitCode());
    assertEquals(false, rcs.getKilled());
    assertEquals(Resource.newInstance(2468, 4), rcs.getCapability());
    // mark the container killed, add some more diags, and verify recovered
    diags.append("some more diags for container");
    stateStore.storeContainerDiagnostics(containerId, diags);
    stateStore.storeContainerKilled(containerId);
    restartStateStore();
    recoveredContainers = stateStore.loadContainersState();
    assertEquals(1, recoveredContainers.size());
    rcs = recoveredContainers.get(0);
    assertEquals(RecoveredContainerStatus.LAUNCHED, rcs.getStatus());
    assertEquals(ContainerExitStatus.INVALID, rcs.getExitCode());
    assertTrue(rcs.getKilled());
    assertEquals(containerReq, rcs.getStartRequest());
    assertEquals(diags.toString(), rcs.getDiagnostics());
    // add yet more diags, mark container completed, and verify recovered
    diags.append("some final diags");
    stateStore.storeContainerDiagnostics(containerId, diags);
    stateStore.storeContainerCompleted(containerId, 21);
    restartStateStore();
    recoveredContainers = stateStore.loadContainersState();
    assertEquals(1, recoveredContainers.size());
    rcs = recoveredContainers.get(0);
    assertEquals(RecoveredContainerStatus.COMPLETED, rcs.getStatus());
    assertEquals(21, rcs.getExitCode());
    assertTrue(rcs.getKilled());
    assertEquals(containerReq, rcs.getStartRequest());
    assertEquals(diags.toString(), rcs.getDiagnostics());
    // store remainingRetryAttempts, workDir and logDir
    stateStore.storeContainerRemainingRetryAttempts(containerId, 6);
    stateStore.storeContainerWorkDir(containerId, "/test/workdir");
    stateStore.storeContainerLogDir(containerId, "/test/logdir");
    restartStateStore();
    recoveredContainers = stateStore.loadContainersState();
    assertEquals(1, recoveredContainers.size());
    rcs = recoveredContainers.get(0);
    assertEquals(6, rcs.getRemainingRetryAttempts());
    assertEquals("/test/workdir", rcs.getWorkDir());
    assertEquals("/test/logdir", rcs.getLogDir());
    // remove the container and verify not recovered
    stateStore.removeContainer(containerId);
    restartStateStore();
    recoveredContainers = stateStore.loadContainersState();
    assertTrue(recoveredContainers.isEmpty());
}
Also used : ContainerId(org.apache.hadoop.yarn.api.records.ContainerId) RecoveredContainerState(org.apache.hadoop.yarn.server.nodemanager.recovery.NMStateStoreService.RecoveredContainerState) ApplicationAttemptId(org.apache.hadoop.yarn.api.records.ApplicationAttemptId) ApplicationId(org.apache.hadoop.yarn.api.records.ApplicationId) DB(org.iq80.leveldb.DB) StartContainerRequest(org.apache.hadoop.yarn.api.protocolrecords.StartContainerRequest) Test(org.junit.Test)

Example 100 with ContainerId

use of org.apache.hadoop.yarn.api.records.ContainerId in project hadoop by apache.

the class TestContainerLogsPage method testContainerLogPageAccess.

@Test(timeout = 10000)
public void testContainerLogPageAccess() throws IOException {
    // SecureIOUtils require Native IO to be enabled. This test will run
    // only if it is enabled.
    assumeTrue(NativeIO.isAvailable());
    String user = "randomUser" + System.currentTimeMillis();
    File absLogDir = null, appDir = null, containerDir = null, syslog = null;
    try {
        // target log directory
        absLogDir = new File("target", TestContainerLogsPage.class.getSimpleName() + "LogDir").getAbsoluteFile();
        absLogDir.mkdir();
        Configuration conf = new Configuration();
        conf.set(YarnConfiguration.NM_LOG_DIRS, absLogDir.toURI().toString());
        conf.set(CommonConfigurationKeysPublic.HADOOP_SECURITY_AUTHENTICATION, "kerberos");
        UserGroupInformation.setConfiguration(conf);
        NodeHealthCheckerService healthChecker = createNodeHealthCheckerService(conf);
        healthChecker.init(conf);
        LocalDirsHandlerService dirsHandler = healthChecker.getDiskHandler();
        // Add an application and the corresponding containers
        RecordFactory recordFactory = RecordFactoryProvider.getRecordFactory(conf);
        long clusterTimeStamp = 1234;
        ApplicationId appId = BuilderUtils.newApplicationId(recordFactory, clusterTimeStamp, 1);
        Application app = mock(Application.class);
        when(app.getAppId()).thenReturn(appId);
        // Making sure that application returns a random user. This is required
        // for SecureIOUtils' file owner check.
        when(app.getUser()).thenReturn(user);
        ApplicationAttemptId appAttemptId = BuilderUtils.newApplicationAttemptId(appId, 1);
        ContainerId container1 = BuilderUtils.newContainerId(recordFactory, appId, appAttemptId, 0);
        // Testing secure read access for log files
        // Creating application and container directory and syslog file.
        appDir = new File(absLogDir, appId.toString());
        appDir.mkdir();
        containerDir = new File(appDir, container1.toString());
        containerDir.mkdir();
        syslog = new File(containerDir, "syslog");
        syslog.createNewFile();
        BufferedOutputStream out = new BufferedOutputStream(new FileOutputStream(syslog));
        out.write("Log file Content".getBytes());
        out.close();
        Context context = mock(Context.class);
        ConcurrentMap<ApplicationId, Application> appMap = new ConcurrentHashMap<ApplicationId, Application>();
        appMap.put(appId, app);
        when(context.getApplications()).thenReturn(appMap);
        ConcurrentHashMap<ContainerId, Container> containers = new ConcurrentHashMap<ContainerId, Container>();
        when(context.getContainers()).thenReturn(containers);
        when(context.getLocalDirsHandler()).thenReturn(dirsHandler);
        MockContainer container = new MockContainer(appAttemptId, new AsyncDispatcher(), conf, user, appId, 1);
        container.setState(ContainerState.RUNNING);
        context.getContainers().put(container1, container);
        ContainersLogsBlock cLogsBlock = new ContainersLogsBlock(context);
        Map<String, String> params = new HashMap<String, String>();
        params.put(YarnWebParams.CONTAINER_ID, container1.toString());
        params.put(YarnWebParams.CONTAINER_LOG_TYPE, "syslog");
        Injector injector = WebAppTests.testPage(ContainerLogsPage.class, ContainersLogsBlock.class, cLogsBlock, params, (Module[]) null);
        PrintWriter spyPw = WebAppTests.getPrintWriter(injector);
        verify(spyPw).write("Exception reading log file. Application submitted by '" + user + "' doesn't own requested log file : syslog");
    } finally {
        if (syslog != null) {
            syslog.delete();
        }
        if (containerDir != null) {
            containerDir.delete();
        }
        if (appDir != null) {
            appDir.delete();
        }
        if (absLogDir != null) {
            absLogDir.delete();
        }
    }
}
Also used : Configuration(org.apache.hadoop.conf.Configuration) YarnConfiguration(org.apache.hadoop.yarn.conf.YarnConfiguration) NodeHealthCheckerService(org.apache.hadoop.yarn.server.nodemanager.NodeHealthCheckerService) ConcurrentHashMap(java.util.concurrent.ConcurrentHashMap) HashMap(java.util.HashMap) Container(org.apache.hadoop.yarn.server.nodemanager.containermanager.container.Container) ContainerId(org.apache.hadoop.yarn.api.records.ContainerId) Injector(com.google.inject.Injector) ConcurrentHashMap(java.util.concurrent.ConcurrentHashMap) BufferedOutputStream(java.io.BufferedOutputStream) PrintWriter(java.io.PrintWriter) NMContext(org.apache.hadoop.yarn.server.nodemanager.NodeManager.NMContext) Context(org.apache.hadoop.yarn.server.nodemanager.Context) ContainersLogsBlock(org.apache.hadoop.yarn.server.nodemanager.webapp.ContainerLogsPage.ContainersLogsBlock) ApplicationAttemptId(org.apache.hadoop.yarn.api.records.ApplicationAttemptId) LocalDirsHandlerService(org.apache.hadoop.yarn.server.nodemanager.LocalDirsHandlerService) RecordFactory(org.apache.hadoop.yarn.factories.RecordFactory) AsyncDispatcher(org.apache.hadoop.yarn.event.AsyncDispatcher) FileOutputStream(java.io.FileOutputStream) ApplicationId(org.apache.hadoop.yarn.api.records.ApplicationId) Module(com.google.inject.Module) File(java.io.File) Application(org.apache.hadoop.yarn.server.nodemanager.containermanager.application.Application) Test(org.junit.Test)

Aggregations

ContainerId (org.apache.hadoop.yarn.api.records.ContainerId)588 Test (org.junit.Test)339 ApplicationId (org.apache.hadoop.yarn.api.records.ApplicationId)173 ApplicationAttemptId (org.apache.hadoop.yarn.api.records.ApplicationAttemptId)169 ArrayList (java.util.ArrayList)161 RMApp (org.apache.hadoop.yarn.server.resourcemanager.rmapp.RMApp)119 Container (org.apache.hadoop.yarn.api.records.Container)104 YarnConfiguration (org.apache.hadoop.yarn.conf.YarnConfiguration)94 MockNM (org.apache.hadoop.yarn.server.resourcemanager.MockNM)79 MockRM (org.apache.hadoop.yarn.server.resourcemanager.MockRM)78 MockAM (org.apache.hadoop.yarn.server.resourcemanager.MockAM)77 Path (org.apache.hadoop.fs.Path)76 HashMap (java.util.HashMap)74 Configuration (org.apache.hadoop.conf.Configuration)74 IOException (java.io.IOException)72 RMContainer (org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainer)68 Resource (org.apache.hadoop.yarn.api.records.Resource)67 ContainerStatus (org.apache.hadoop.yarn.api.records.ContainerStatus)66 ResourceRequest (org.apache.hadoop.yarn.api.records.ResourceRequest)61 NodeId (org.apache.hadoop.yarn.api.records.NodeId)59