Apache
/
hadoop
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

YARN-1256. NM silently ignores non-existent service in StartContainerRequest (Xuan Gong via bikas) 

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/trunk@1529039 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Bikas Saha 2013-10-04 00:23:35 +00:00
parent ac2cdb5f65
commit 8ebf37f369
6 changed files with 134 additions and 34 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
hadoop-yarn-project/CHANGES.txt
View File

@ -144,6 +144,9 @@ Release 2.1.2 - UNRELEASED
YARN-1236. FairScheduler setting queue name in RMApp is not working.
(Sandy Ryza)
YARN-1256. NM silently ignores non-existent service in
StartContainerRequest (Xuan Gong via bikas)
Release 2.1.1-beta - 2013-09-23
INCOMPATIBLE CHANGES

7
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/java/org/apache/hadoop/yarn/util/AuxiliaryServiceHelper.java
View File

@ -30,8 +30,11 @@ public class AuxiliaryServiceHelper {
public static ByteBuffer getServiceDataFromEnv(String serviceName,
Map<String, String> env) {
byte[] metaData =
Base64.decodeBase64(env.get(getPrefixServiceName(serviceName)));
String meta = env.get(getPrefixServiceName(serviceName));
if (null == meta) {
return null;
}
byte[] metaData = Base64.decodeBase64(meta);
return ByteBuffer.wrap(metaData);
}

38
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/containermanager/AuxServices.java
View File

@ -177,33 +177,50 @@ public class AuxServices extends AbstractService
switch (event.getType()) {
case APPLICATION_INIT:
LOG.info("Got APPLICATION_INIT for service " + event.getServiceID());
AuxiliaryService service = serviceMap.get(event.getServiceID());
if (null == service) {
LOG.info("service is null");
// TODO kill all containers waiting on Application
return;
}
service.initializeApplication(new ApplicationInitializationContext(event
AuxiliaryService service = null;
try {
service = serviceMap.get(event.getServiceID());
service
.initializeApplication(new ApplicationInitializationContext(event
.getUser(), event.getApplicationID(), event.getServiceData()));
} catch (Throwable th) {
logWarningWhenAuxServiceThrowExceptions(service,
AuxServicesEventType.APPLICATION_INIT, th);
}
break;
case APPLICATION_STOP:
for (AuxiliaryService serv : serviceMap.values()) {
try {
serv.stopApplication(new ApplicationTerminationContext(event
.getApplicationID()));
} catch (Throwable th) {
logWarningWhenAuxServiceThrowExceptions(serv,
AuxServicesEventType.APPLICATION_STOP, th);
}
}
break;
case CONTAINER_INIT:
for (AuxiliaryService serv : serviceMap.values()) {
try {
serv.initializeContainer(new ContainerInitializationContext(
event.getUser(), event.getContainer().getContainerId(),
event.getContainer().getResource()));
} catch (Throwable th) {
logWarningWhenAuxServiceThrowExceptions(serv,
AuxServicesEventType.CONTAINER_INIT, th);
}
}
break;
case CONTAINER_STOP:
for (AuxiliaryService serv : serviceMap.values()) {
try {
serv.stopContainer(new ContainerTerminationContext(
event.getUser(), event.getContainer().getContainerId(),
event.getContainer().getResource()));
} catch (Throwable th) {
logWarningWhenAuxServiceThrowExceptions(serv,
AuxServicesEventType.CONTAINER_STOP, th);
}
}
break;
default:
@ -217,4 +234,11 @@ public class AuxServices extends AbstractService
}
return p.matcher(name).matches();
}
private void logWarningWhenAuxServiceThrowExceptions(AuxiliaryService service,
AuxServicesEventType eventType, Throwable th) {
LOG.warn((null == service ? "The auxService is null"
: "The auxService name is " + service.getName())
+ " and it got an error at event: " + eventType, th);
}
}

13
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/containermanager/ContainerManagerImpl.java
View File

@ -67,6 +67,7 @@ import org.apache.hadoop.yarn.api.records.SerializedException;
import org.apache.hadoop.yarn.conf.YarnConfiguration;
import org.apache.hadoop.yarn.event.AsyncDispatcher;
import org.apache.hadoop.yarn.event.EventHandler;
import org.apache.hadoop.yarn.exceptions.InvalidAuxServiceException;
import org.apache.hadoop.yarn.exceptions.InvalidContainerException;
import org.apache.hadoop.yarn.exceptions.NMNotYetReadyException;
import org.apache.hadoop.yarn.exceptions.YarnException;
@ -451,6 +452,18 @@ public class ContainerManagerImpl extends CompositeService implements
ContainerLaunchContext launchContext = request.getContainerLaunchContext();
Map<String, ByteBuffer> serviceData = getAuxServiceMetaData();
if (launchContext.getServiceData()!=null &&
!launchContext.getServiceData().isEmpty()) {
for (Map.Entry<String, ByteBuffer> meta : launchContext.getServiceData()
.entrySet()) {
if (null == serviceData.get(meta.getKey())) {
throw new InvalidAuxServiceException("The auxService:" + meta.getKey()
+ " does not exist");
}
}
}
Credentials credentials = parseCredentials(launchContext);
Container container =

11
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/test/java/org/apache/hadoop/yarn/server/nodemanager/TestContainerManagerWithLCE.java
View File

@ -178,6 +178,17 @@ public class TestContainerManagerWithLCE extends TestContainerManager {
super.testMultipleContainersStopAndGetStatus();
}
@Override
public void testStartContainerFailureWithUnknownAuxService() throws Exception {
// Don't run the test if the binary is not available.
if (!shouldRunTest()) {
LOG.info("LCE binary path is not passed. Not running the test");
return;
}
LOG.info("Running testContainerLaunchFromPreviousRM");
super.testStartContainerFailureWithUnknownAuxService();
}
private boolean shouldRunTest() {
return System
.getProperty(YarnConfiguration.NM_LINUX_CONTAINER_EXECUTOR_PATH) != null;

46
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/test/java/org/apache/hadoop/yarn/server/nodemanager/containermanager/TestContainerManager.java
View File

@ -24,6 +24,7 @@ import java.io.FileReader;
import java.io.IOException;
import java.io.PrintWriter;
import java.net.InetAddress;
import java.nio.ByteBuffer;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.HashMap;
@ -37,6 +38,7 @@ import org.apache.hadoop.fs.FileContext;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.fs.UnsupportedFileSystemException;
import org.apache.hadoop.security.UserGroupInformation;
import org.apache.hadoop.service.Service;
import org.apache.hadoop.util.Shell;
import org.apache.hadoop.yarn.api.protocolrecords.GetContainerStatusesRequest;
import org.apache.hadoop.yarn.api.protocolrecords.GetContainerStatusesResponse;
@ -59,6 +61,7 @@ import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.api.records.SerializedException;
import org.apache.hadoop.yarn.api.records.Token;
import org.apache.hadoop.yarn.api.records.URL;
import org.apache.hadoop.yarn.conf.YarnConfiguration;
import org.apache.hadoop.yarn.exceptions.InvalidContainerException;
import org.apache.hadoop.yarn.exceptions.YarnException;
import org.apache.hadoop.yarn.security.ContainerTokenIdentifier;
@ -68,6 +71,7 @@ import org.apache.hadoop.yarn.server.nodemanager.CMgrCompletedAppsEvent;
import org.apache.hadoop.yarn.server.nodemanager.ContainerExecutor.ExitCode;
import org.apache.hadoop.yarn.server.nodemanager.DefaultContainerExecutor;
import org.apache.hadoop.yarn.server.nodemanager.DeletionService;
import org.apache.hadoop.yarn.server.nodemanager.containermanager.TestAuxServices.ServiceA;
import org.apache.hadoop.yarn.server.nodemanager.containermanager.application.ApplicationState;
import org.apache.hadoop.yarn.server.nodemanager.containermanager.container.Container;
import org.apache.hadoop.yarn.server.nodemanager.containermanager.localizer.ContainerLocalizer;
@ -746,6 +750,48 @@ public class TestContainerManager extends BaseContainerManagerTest {
}
}
@Test
public void testStartContainerFailureWithUnknownAuxService() throws Exception {
conf.setStrings(YarnConfiguration.NM_AUX_SERVICES,
new String[] { "existService" });
conf.setClass(
String.format(YarnConfiguration.NM_AUX_SERVICE_FMT, "existService"),
ServiceA.class, Service.class);
containerManager.start();
List<StartContainerRequest> startRequest =
new ArrayList<StartContainerRequest>();
ContainerLaunchContext containerLaunchContext =
recordFactory.newRecordInstance(ContainerLaunchContext.class);
Map<String, ByteBuffer> serviceData = new HashMap<String, ByteBuffer>();
String serviceName = "non_exist_auxService";
serviceData.put(serviceName, ByteBuffer.wrap(serviceName.getBytes()));
containerLaunchContext.setServiceData(serviceData);
ContainerId cId = createContainerId(0);
String user = "start_container_fail";
Token containerToken =
createContainerToken(cId, DUMMY_RM_IDENTIFIER, context.getNodeId(),
user, context.getContainerTokenSecretManager());
StartContainerRequest request =
StartContainerRequest.newInstance(containerLaunchContext,
containerToken);
// start containers
startRequest.add(request);
StartContainersRequest requestList =
StartContainersRequest.newInstance(startRequest);
StartContainersResponse response =
containerManager.startContainers(requestList);
Assert.assertTrue(response.getFailedRequests().size() == 1);
Assert.assertTrue(response.getSuccessfullyStartedContainers().size() == 0);
Assert.assertTrue(response.getFailedRequests().containsKey(cId));
Assert.assertTrue(response.getFailedRequests().get(cId).getMessage()
.contains("The auxService:" + serviceName + " does not exist"));
}
public static Token createContainerToken(ContainerId cId, long rmIdentifier,
NodeId nodeId, String user,
NMContainerTokenSecretManager containerTokenSecretManager)