You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-commits@hadoop.apache.org by vi...@apache.org on 2014/06/09 05:10:07 UTC
svn commit: r1601304 [1/2] - in
/hadoop/common/branches/branch-2/hadoop-yarn-project: ./
hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/
hadoop-yarn/hadoop-yarn-common/src/main/resources/
hadoop-yarn/hadoop-yarn-server/hadoop-yar...
Author: vinodkv
Date: Mon Jun 9 03:10:06 2014
New Revision: 1601304
URL: http://svn.apache.org/r1601304
Log:
YARN-1368. Added core functionality of recovering container state into schedulers after ResourceManager Restart so as to preserve running work in the cluster. Contributed by Jian He.
svn merge --ignore-ancestry -c 1601303 ../../trunk/
Added:
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmcontainer/RMContainerRecoverEvent.java
- copied unchanged from r1601303, hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmcontainer/RMContainerRecoverEvent.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmnode/RMNodeStartedEvent.java
- copied unchanged from r1601303, hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmnode/RMNodeStartedEvent.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/TestWorkPreservingRMRestart.java
- copied unchanged from r1601303, hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/TestWorkPreservingRMRestart.java
Modified:
hadoop/common/branches/branch-2/hadoop-yarn-project/CHANGES.txt
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/YarnConfiguration.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/resources/yarn-default.xml
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/RMContext.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/RMContextImpl.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ResourceManager.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ResourceTrackerService.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmapp/RMAppImpl.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmapp/attempt/RMAppAttemptImpl.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmcontainer/RMContainerEventType.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmcontainer/RMContainerImpl.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmnode/RMNodeImpl.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/AbstractYarnScheduler.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/AppSchedulingInfo.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/Queue.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/SchedulerApplicationAttempt.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/SchedulerNode.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CSQueue.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CapacityScheduler.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/ParentQueue.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/common/fica/FiCaSchedulerNode.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/event/NodeAddedSchedulerEvent.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/AppSchedulable.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/FSLeafQueue.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/FSParentQueue.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fifo/FifoScheduler.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/MockRM.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/TestFifoScheduler.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/TestMoveApplication.java
hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/TestRMNodeTransitions.java
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/CHANGES.txt
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/CHANGES.txt?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/CHANGES.txt (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/CHANGES.txt Mon Jun 9 03:10:06 2014
@@ -17,6 +17,10 @@ Release 2.5.0 - UNRELEASED
YARN-1338. Recover localized resource cache state upon nodemanager restart
(Jason Lowe via junping_du)
+ YARN-1368. Added core functionality of recovering container state into
+ schedulers after ResourceManager Restart so as to preserve running work in
+ the cluster. (Jian He via vinodkv)
+
IMPROVEMENTS
YARN-1479. Invalid NaN values in Hadoop REST API JSON response (Chen He via
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/YarnConfiguration.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/YarnConfiguration.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/YarnConfiguration.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-api/src/main/java/org/apache/hadoop/yarn/conf/YarnConfiguration.java Mon Jun 9 03:10:06 2014
@@ -318,6 +318,13 @@ public class YarnConfiguration extends C
public static final String RECOVERY_ENABLED = RM_PREFIX + "recovery.enabled";
public static final boolean DEFAULT_RM_RECOVERY_ENABLED = false;
+ @Private
+ public static final String RM_WORK_PRESERVING_RECOVERY_ENABLED = RM_PREFIX
+ + "work-preserving-recovery.enabled";
+ @Private
+ public static final boolean DEFAULT_RM_WORK_PRESERVING_RECOVERY_ENABLED =
+ false;
+
/** Zookeeper interaction configs */
public static final String RM_ZK_PREFIX = RM_PREFIX + "zk-";
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/resources/yarn-default.xml
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/resources/yarn-default.xml?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/resources/yarn-default.xml (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-common/src/main/resources/yarn-default.xml Mon Jun 9 03:10:06 2014
@@ -270,6 +270,14 @@
</property>
<property>
+ <description>Enable RM work preserving recovery. This configuration is private
+ to YARN for experimenting the feature.
+ </description>
+ <name>yarn.resourcemanager.work-preserving-recovery.enabled</name>
+ <value>false</value>
+ </property>
+
+ <property>
<description>The class to use as the persistent store.
If org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/RMContext.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/RMContext.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/RMContext.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/RMContext.java Mon Jun 9 03:10:06 2014
@@ -99,4 +99,6 @@ public interface RMContext {
RMApplicationHistoryWriter rmApplicationHistoryWriter);
ConfigurationProvider getConfigurationProvider();
+
+ boolean isWorkPreservingRecoveryEnabled();
}
\ No newline at end of file
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/RMContextImpl.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/RMContextImpl.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/RMContextImpl.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/RMContextImpl.java Mon Jun 9 03:10:06 2014
@@ -60,6 +60,7 @@ public class RMContextImpl implements RM
= new ConcurrentHashMap<String, RMNode>();
private boolean isHAEnabled;
+ private boolean isWorkPreservingRecoveryEnabled;
private HAServiceState haServiceState =
HAServiceProtocol.HAServiceState.INITIALIZING;
@@ -329,6 +330,15 @@ public class RMContextImpl implements RM
}
}
+ public void setWorkPreservingRecoveryEnabled(boolean enabled) {
+ this.isWorkPreservingRecoveryEnabled = enabled;
+ }
+
+ @Override
+ public boolean isWorkPreservingRecoveryEnabled() {
+ return this.isWorkPreservingRecoveryEnabled;
+ }
+
@Override
public RMApplicationHistoryWriter getRMApplicationHistoryWriter() {
return rmApplicationHistoryWriter;
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ResourceManager.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ResourceManager.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ResourceManager.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ResourceManager.java Mon Jun 9 03:10:06 2014
@@ -364,9 +364,15 @@ public class ResourceManager extends Com
YarnConfiguration.DEFAULT_RM_RECOVERY_ENABLED);
RMStateStore rmStore = null;
- if(isRecoveryEnabled) {
+ if (isRecoveryEnabled) {
recoveryEnabled = true;
- rmStore = RMStateStoreFactory.getStore(conf);
+ rmStore = RMStateStoreFactory.getStore(conf);
+ boolean isWorkPreservingRecoveryEnabled =
+ conf.getBoolean(
+ YarnConfiguration.RM_WORK_PRESERVING_RECOVERY_ENABLED,
+ YarnConfiguration.DEFAULT_RM_WORK_PRESERVING_RECOVERY_ENABLED);
+ rmContext
+ .setWorkPreservingRecoveryEnabled(isWorkPreservingRecoveryEnabled);
} else {
recoveryEnabled = false;
rmStore = new NullRMStateStore();
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ResourceTrackerService.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ResourceTrackerService.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ResourceTrackerService.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ResourceTrackerService.java Mon Jun 9 03:10:06 2014
@@ -60,6 +60,7 @@ import org.apache.hadoop.yarn.server.res
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeEventType;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeImpl;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeReconnectEvent;
+import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeStartedEvent;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeStatusEvent;
import org.apache.hadoop.yarn.server.resourcemanager.security.NMTokenSecretManagerInRM;
import org.apache.hadoop.yarn.server.resourcemanager.security.RMContainerTokenSecretManager;
@@ -243,11 +244,13 @@ public class ResourceTrackerService exte
Resource capability = request.getResource();
String nodeManagerVersion = request.getNMVersion();
- if (!request.getNMContainerStatuses().isEmpty()) {
- LOG.info("received container statuses on node manager register :"
- + request.getNMContainerStatuses());
- for (NMContainerStatus report : request.getNMContainerStatuses()) {
- handleNMContainerStatus(report);
+ if (!rmContext.isWorkPreservingRecoveryEnabled()) {
+ if (!request.getNMContainerStatuses().isEmpty()) {
+ LOG.info("received container statuses on node manager register :"
+ + request.getNMContainerStatuses());
+ for (NMContainerStatus status : request.getNMContainerStatuses()) {
+ handleNMContainerStatus(status);
+ }
}
}
RegisterNodeManagerResponse response = recordFactory
@@ -308,7 +311,7 @@ public class ResourceTrackerService exte
RMNode oldNode = this.rmContext.getRMNodes().putIfAbsent(nodeId, rmNode);
if (oldNode == null) {
this.rmContext.getDispatcher().getEventHandler().handle(
- new RMNodeEvent(nodeId, RMNodeEventType.STARTED));
+ new RMNodeStartedEvent(nodeId, request.getNMContainerStatuses()));
} else {
LOG.info("Reconnect from the node at: " + host);
this.nmLivelinessMonitor.unregister(nodeId);
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmapp/RMAppImpl.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmapp/RMAppImpl.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmapp/RMAppImpl.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmapp/RMAppImpl.java Mon Jun 9 03:10:06 2014
@@ -723,29 +723,36 @@ public class RMAppImpl implements RMApp,
}
}
+ // synchronously recover attempt to ensure any incoming external events
+ // to be processed after the attempt processes the recover event.
+ private void recoverAppAttempts() {
+ for (RMAppAttempt attempt : getAppAttempts().values()) {
+ attempt.handle(new RMAppAttemptEvent(attempt.getAppAttemptId(),
+ RMAppAttemptEventType.RECOVER));
+ }
+ }
+
private static final class RMAppRecoveredTransition implements
MultipleArcTransition<RMAppImpl, RMAppEvent, RMAppState> {
@Override
public RMAppState transition(RMAppImpl app, RMAppEvent event) {
- for (RMAppAttempt attempt : app.getAppAttempts().values()) {
- // synchronously recover attempt to ensure any incoming external events
- // to be processed after the attempt processes the recover event.
- attempt.handle(
- new RMAppAttemptEvent(attempt.getAppAttemptId(),
- RMAppAttemptEventType.RECOVER));
- }
-
// The app has completed.
if (app.recoveredFinalState != null) {
+ app.recoverAppAttempts();
new FinalTransition(app.recoveredFinalState).transition(app, event);
return app.recoveredFinalState;
}
- // Last attempt is in final state, do not add to scheduler and just return
- // ACCEPTED waiting for last RMAppAttempt to send finished or failed event
- // back.
+ // Notify scheduler about the app on recovery
+ new AddApplicationToSchedulerTransition().transition(app, event);
+
+ // recover attempts
+ app.recoverAppAttempts();
+
+ // Last attempt is in final state, return ACCEPTED waiting for last
+ // RMAppAttempt to send finished or failed event back.
if (app.currentAttempt != null
&& (app.currentAttempt.getState() == RMAppAttemptState.KILLED
|| app.currentAttempt.getState() == RMAppAttemptState.FINISHED
@@ -754,9 +761,6 @@ public class RMAppImpl implements RMApp,
return RMAppState.ACCEPTED;
}
- // Notify scheduler about the app on recovery
- new AddApplicationToSchedulerTransition().transition(app, event);
-
// No existent attempts means the attempt associated with this app was not
// started or started but not yet saved.
if (app.attempts.isEmpty()) {
@@ -1055,8 +1059,12 @@ public class RMAppImpl implements RMApp,
if (app.finishTime == 0 ) {
app.finishTime = System.currentTimeMillis();
}
- app.handler.handle(new AppRemovedSchedulerEvent(app.applicationId,
- finalState));
+ // Recovered apps that are completed were not added to scheduler, so no
+ // need to remove them from scheduler.
+ if (app.recoveredFinalState == null) {
+ app.handler.handle(new AppRemovedSchedulerEvent(app.applicationId,
+ finalState));
+ }
app.handler.handle(
new RMAppManagerEvent(app.applicationId,
RMAppManagerEventType.APP_COMPLETED));
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmapp/attempt/RMAppAttemptImpl.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmapp/attempt/RMAppAttemptImpl.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmapp/attempt/RMAppAttemptImpl.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmapp/attempt/RMAppAttemptImpl.java Mon Jun 9 03:10:06 2014
@@ -267,15 +267,17 @@ public class RMAppAttemptImpl implements
.addTransition(RMAppAttemptState.ALLOCATED, RMAppAttemptState.FINAL_SAVING,
RMAppAttemptEventType.CONTAINER_FINISHED,
new FinalSavingTransition(
- new AMContainerCrashedTransition(), RMAppAttemptState.FAILED))
+ new AMContainerCrashedBeforeRunningTransition(), RMAppAttemptState.FAILED))
// Transitions from LAUNCHED State
.addTransition(RMAppAttemptState.LAUNCHED, RMAppAttemptState.RUNNING,
RMAppAttemptEventType.REGISTERED, new AMRegisteredTransition())
- .addTransition(RMAppAttemptState.LAUNCHED, RMAppAttemptState.FINAL_SAVING,
+ .addTransition(RMAppAttemptState.LAUNCHED,
+ EnumSet.of(RMAppAttemptState.LAUNCHED, RMAppAttemptState.FINAL_SAVING),
RMAppAttemptEventType.CONTAINER_FINISHED,
- new FinalSavingTransition(
- new AMContainerCrashedTransition(), RMAppAttemptState.FAILED))
+ new ContainerFinishedTransition(
+ new AMContainerCrashedBeforeRunningTransition(),
+ RMAppAttemptState.LAUNCHED))
.addTransition(
RMAppAttemptState.LAUNCHED, RMAppAttemptState.FINAL_SAVING,
RMAppAttemptEventType.EXPIRE,
@@ -302,7 +304,9 @@ public class RMAppAttemptImpl implements
RMAppAttemptState.RUNNING,
EnumSet.of(RMAppAttemptState.RUNNING, RMAppAttemptState.FINAL_SAVING),
RMAppAttemptEventType.CONTAINER_FINISHED,
- new ContainerFinishedTransition())
+ new ContainerFinishedTransition(
+ new AMContainerCrashedAtRunningTransition(),
+ RMAppAttemptState.RUNNING))
.addTransition(
RMAppAttemptState.RUNNING, RMAppAttemptState.FINAL_SAVING,
RMAppAttemptEventType.EXPIRE,
@@ -904,6 +908,12 @@ public class RMAppAttemptImpl implements
}
return appAttempt.recoveredFinalState;
} else {
+ // Add the current attempt to the scheduler.
+ if (appAttempt.rmContext.isWorkPreservingRecoveryEnabled()) {
+ appAttempt.eventHandler.handle(new AppAttemptAddedSchedulerEvent(
+ appAttempt.getAppAttemptId(), false));
+ }
+
/*
* Since the application attempt's final state is not saved that means
* for AM container (previous attempt) state must be one of these.
@@ -1207,17 +1217,16 @@ public class RMAppAttemptImpl implements
}
}
- private static final class AMContainerCrashedTransition extends
+ private static final class AMContainerCrashedBeforeRunningTransition extends
BaseFinalTransition {
- public AMContainerCrashedTransition() {
+ public AMContainerCrashedBeforeRunningTransition() {
super(RMAppAttemptState.FAILED);
}
@Override
public void transition(RMAppAttemptImpl appAttempt,
RMAppAttemptEvent event) {
-
RMAppAttemptContainerFinishedEvent finishEvent =
((RMAppAttemptContainerFinishedEvent)event);
@@ -1410,6 +1419,16 @@ public class RMAppAttemptImpl implements
implements
MultipleArcTransition<RMAppAttemptImpl, RMAppAttemptEvent, RMAppAttemptState> {
+ // The transition To Do after attempt final state is saved.
+ private BaseTransition transitionToDo;
+ private RMAppAttemptState currentState;
+
+ public ContainerFinishedTransition(BaseTransition transitionToDo,
+ RMAppAttemptState currentState) {
+ this.transitionToDo = transitionToDo;
+ this.currentState = currentState;
+ }
+
@Override
public RMAppAttemptState transition(RMAppAttemptImpl appAttempt,
RMAppAttemptEvent event) {
@@ -1426,14 +1445,13 @@ public class RMAppAttemptImpl implements
containerStatus.getContainerId())) {
// Remember the follow up transition and save the final attempt state.
appAttempt.rememberTargetTransitionsAndStoreState(event,
- new ContainerFinishedFinalStateSavedTransition(),
- RMAppAttemptState.FAILED, RMAppAttemptState.FAILED);
+ transitionToDo, RMAppAttemptState.FAILED, RMAppAttemptState.FAILED);
return RMAppAttemptState.FINAL_SAVING;
}
// Normal container.Put it in completedcontainers list
appAttempt.justFinishedContainers.add(containerStatus);
- return RMAppAttemptState.RUNNING;
+ return this.currentState;
}
}
@@ -1451,7 +1469,7 @@ public class RMAppAttemptImpl implements
}
}
- private static class ContainerFinishedFinalStateSavedTransition extends
+ private static class AMContainerCrashedAtRunningTransition extends
BaseTransition {
@Override
public void
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmcontainer/RMContainerEventType.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmcontainer/RMContainerEventType.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmcontainer/RMContainerEventType.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmcontainer/RMContainerEventType.java Mon Jun 9 03:10:06 2014
@@ -33,5 +33,7 @@ public enum RMContainerEventType {
RELEASED,
// Source: ContainerAllocationExpirer
- EXPIRE
+ EXPIRE,
+
+ RECOVER
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmcontainer/RMContainerImpl.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmcontainer/RMContainerImpl.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmcontainer/RMContainerImpl.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmcontainer/RMContainerImpl.java Mon Jun 9 03:10:06 2014
@@ -35,12 +35,14 @@ import org.apache.hadoop.yarn.api.record
import org.apache.hadoop.yarn.api.records.Priority;
import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.event.EventHandler;
+import org.apache.hadoop.yarn.server.api.protocolrecords.NMContainerStatus;
import org.apache.hadoop.yarn.server.resourcemanager.RMContext;
import org.apache.hadoop.yarn.server.resourcemanager.rmapp.attempt.event.RMAppAttemptContainerAcquiredEvent;
import org.apache.hadoop.yarn.server.resourcemanager.rmapp.attempt.event.RMAppAttemptContainerAllocatedEvent;
import org.apache.hadoop.yarn.server.resourcemanager.rmapp.attempt.event.RMAppAttemptContainerFinishedEvent;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeCleanContainerEvent;
import org.apache.hadoop.yarn.state.InvalidStateTransitonException;
+import org.apache.hadoop.yarn.state.MultipleArcTransition;
import org.apache.hadoop.yarn.state.SingleArcTransition;
import org.apache.hadoop.yarn.state.StateMachine;
import org.apache.hadoop.yarn.state.StateMachineFactory;
@@ -65,6 +67,9 @@ public class RMContainerImpl implements
RMContainerEventType.KILL)
.addTransition(RMContainerState.NEW, RMContainerState.RESERVED,
RMContainerEventType.RESERVED, new ContainerReservedTransition())
+ .addTransition(RMContainerState.NEW,
+ EnumSet.of(RMContainerState.RUNNING, RMContainerState.COMPLETED),
+ RMContainerEventType.RECOVER, new ContainerRecoveredTransition())
// Transitions from RESERVED state
.addTransition(RMContainerState.RESERVED, RMContainerState.RESERVED,
@@ -341,6 +346,38 @@ public class RMContainerImpl implements
}
}
+ private static final class ContainerRecoveredTransition
+ implements
+ MultipleArcTransition<RMContainerImpl, RMContainerEvent, RMContainerState> {
+ @Override
+ public RMContainerState transition(RMContainerImpl container,
+ RMContainerEvent event) {
+ NMContainerStatus report =
+ ((RMContainerRecoverEvent) event).getContainerReport();
+ if (report.getContainerState().equals(ContainerState.COMPLETE)) {
+ ContainerStatus status =
+ ContainerStatus.newInstance(report.getContainerId(),
+ report.getContainerState(), report.getDiagnostics(),
+ report.getContainerExitStatus());
+
+ new FinishedTransition().transition(container,
+ new RMContainerFinishedEvent(container.containerId, status,
+ RMContainerEventType.FINISHED));
+ return RMContainerState.COMPLETED;
+ } else if (report.getContainerState().equals(ContainerState.RUNNING)) {
+ // Tell the appAttempt
+ container.eventHandler.handle(new RMAppAttemptContainerAcquiredEvent(
+ container.getApplicationAttemptId(), container.getContainer()));
+ return RMContainerState.RUNNING;
+ } else {
+ // This can never happen.
+ LOG.warn("RMContainer received unexpected recover event with container"
+ + " state " + report.getContainerState() + " while recovering.");
+ return RMContainerState.RUNNING;
+ }
+ }
+ }
+
private static final class ContainerReservedTransition extends
BaseTransition {
@@ -398,7 +435,6 @@ public class RMContainerImpl implements
// Inform AppAttempt
container.eventHandler.handle(new RMAppAttemptContainerFinishedEvent(
container.appAttemptId, finishedEvent.getRemoteContainerStatus()));
-
container.rmContext.getRMApplicationHistoryWriter()
.containerFinished(container);
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmnode/RMNodeImpl.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmnode/RMNodeImpl.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmnode/RMNodeImpl.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/rmnode/RMNodeImpl.java Mon Jun 9 03:10:06 2014
@@ -48,6 +48,7 @@ import org.apache.hadoop.yarn.api.record
import org.apache.hadoop.yarn.event.EventHandler;
import org.apache.hadoop.yarn.factories.RecordFactory;
import org.apache.hadoop.yarn.factory.providers.RecordFactoryProvider;
+import org.apache.hadoop.yarn.server.api.protocolrecords.NMContainerStatus;
import org.apache.hadoop.yarn.server.api.protocolrecords.NodeHeartbeatResponse;
import org.apache.hadoop.yarn.server.api.records.NodeHealthStatus;
import org.apache.hadoop.yarn.server.resourcemanager.ClusterMetrics;
@@ -460,13 +461,9 @@ public class RMNodeImpl implements RMNod
@Override
public void transition(RMNodeImpl rmNode, RMNodeEvent event) {
// Inform the scheduler
+ RMNodeStartedEvent startEvent = (RMNodeStartedEvent) event;
+ List<NMContainerStatus> containers = null;
- rmNode.context.getDispatcher().getEventHandler().handle(
- new NodeAddedSchedulerEvent(rmNode));
- rmNode.context.getDispatcher().getEventHandler().handle(
- new NodesListManagerEvent(
- NodesListManagerEventType.NODE_USABLE, rmNode));
-
String host = rmNode.nodeId.getHost();
if (rmNode.context.getInactiveRMNodes().containsKey(host)) {
// Old node rejoining
@@ -476,10 +473,17 @@ public class RMNodeImpl implements RMNod
} else {
// Increment activeNodes explicitly because this is a new node.
ClusterMetrics.getMetrics().incrNumActiveNodes();
+ containers = startEvent.getContainerRecoveryReports();
}
+
+ rmNode.context.getDispatcher().getEventHandler()
+ .handle(new NodeAddedSchedulerEvent(rmNode, containers));
+ rmNode.context.getDispatcher().getEventHandler().handle(
+ new NodesListManagerEvent(
+ NodesListManagerEventType.NODE_USABLE, rmNode));
}
}
-
+
public static class ReconnectNodeTransition implements
SingleArcTransition<RMNodeImpl, RMNodeEvent> {
@@ -513,7 +517,7 @@ public class RMNodeImpl implements RMNod
}
rmNode.context.getRMNodes().put(newNode.getNodeID(), newNode);
rmNode.context.getDispatcher().getEventHandler().handle(
- new RMNodeEvent(newNode.getNodeID(), RMNodeEventType.STARTED));
+ new RMNodeStartedEvent(newNode.getNodeID(), null));
}
rmNode.context.getDispatcher().getEventHandler().handle(
new NodesListManagerEvent(
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/AbstractYarnScheduler.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/AbstractYarnScheduler.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/AbstractYarnScheduler.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/AbstractYarnScheduler.java Mon Jun 9 03:10:06 2014
@@ -32,14 +32,21 @@ import org.apache.hadoop.yarn.api.record
import org.apache.hadoop.yarn.api.records.ApplicationResourceUsageReport;
import org.apache.hadoop.yarn.api.records.Container;
import org.apache.hadoop.yarn.api.records.ContainerId;
+import org.apache.hadoop.yarn.api.records.ContainerState;
import org.apache.hadoop.yarn.api.records.NodeId;
import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.exceptions.YarnException;
+import org.apache.hadoop.yarn.server.api.protocolrecords.NMContainerStatus;
import org.apache.hadoop.yarn.server.resourcemanager.RMContext;
import org.apache.hadoop.yarn.server.resourcemanager.rmapp.RMApp;
import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainer;
+import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerImpl;
+import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerRecoverEvent;
+import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNode;
+import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeCleanContainerEvent;
import org.apache.hadoop.yarn.util.resource.Resources;
+@SuppressWarnings("unchecked")
public abstract class AbstractYarnScheduler
<T extends SchedulerApplicationAttempt, N extends SchedulerNode>
extends AbstractService implements ResourceScheduler {
@@ -47,8 +54,7 @@ public abstract class AbstractYarnSchedu
private static final Log LOG = LogFactory.getLog(AbstractYarnScheduler.class);
// Nodes in the cluster, indexed by NodeId
- protected Map<NodeId, N> nodes =
- new ConcurrentHashMap<NodeId, N>();
+ protected Map<NodeId, N> nodes = new ConcurrentHashMap<NodeId, N>();
// Whole capacity of the cluster
protected Resource clusterResource = Resource.newInstance(0, 0);
@@ -58,6 +64,7 @@ public abstract class AbstractYarnSchedu
protected RMContext rmContext;
protected Map<ApplicationId, SchedulerApplication<T>> applications;
+
protected final static List<Container> EMPTY_CONTAINER_LIST =
new ArrayList<Container>();
protected static final Allocation EMPTY_ALLOCATION = new Allocation(
@@ -169,4 +176,90 @@ public abstract class AbstractYarnSchedu
throw new YarnException(getClass().getSimpleName()
+ " does not support moving apps between queues");
}
+
+ private void killOrphanContainerOnNode(RMNode node,
+ NMContainerStatus container) {
+ if (!container.getContainerState().equals(ContainerState.COMPLETE)) {
+ this.rmContext.getDispatcher().getEventHandler().handle(
+ new RMNodeCleanContainerEvent(node.getNodeID(),
+ container.getContainerId()));
+ }
+ }
+
+ public synchronized void recoverContainersOnNode(
+ List<NMContainerStatus> containerReports, RMNode nm) {
+ if (!rmContext.isWorkPreservingRecoveryEnabled()
+ || containerReports == null
+ || (containerReports != null && containerReports.isEmpty())) {
+ return;
+ }
+
+ for (NMContainerStatus container : containerReports) {
+ ApplicationId appId =
+ container.getContainerId().getApplicationAttemptId().getApplicationId();
+ RMApp rmApp = rmContext.getRMApps().get(appId);
+ if (rmApp == null) {
+ LOG.error("Skip recovering container " + container
+ + " for unknown application.");
+ killOrphanContainerOnNode(nm, container);
+ continue;
+ }
+
+ // Unmanaged AM recovery is addressed in YARN-1815
+ if (rmApp.getApplicationSubmissionContext().getUnmanagedAM()) {
+ LOG.info("Skip recovering container " + container + " for unmanaged AM."
+ + rmApp.getApplicationId());
+ killOrphanContainerOnNode(nm, container);
+ continue;
+ }
+
+ SchedulerApplication<T> schedulerApp = applications.get(appId);
+ if (schedulerApp == null) {
+ LOG.info("Skip recovering container " + container
+ + " for unknown SchedulerApplication. Application current state is "
+ + rmApp.getState());
+ killOrphanContainerOnNode(nm, container);
+ continue;
+ }
+
+ LOG.info("Recovering container " + container);
+ SchedulerApplicationAttempt schedulerAttempt =
+ schedulerApp.getCurrentAppAttempt();
+
+ // create container
+ RMContainer rmContainer = recoverAndCreateContainer(container, nm);
+
+ // recover RMContainer
+ rmContainer.handle(new RMContainerRecoverEvent(container.getContainerId(),
+ container));
+
+ // recover scheduler node
+ nodes.get(nm.getNodeID()).recoverContainer(rmContainer);
+
+ // recover queue: update headroom etc.
+ Queue queue = schedulerAttempt.getQueue();
+ queue.recoverContainer(clusterResource, schedulerAttempt, rmContainer);
+
+ // recover scheduler attempt
+ schedulerAttempt.recoverContainer(rmContainer);
+ }
+ }
+
+ private RMContainer recoverAndCreateContainer(NMContainerStatus report,
+ RMNode node) {
+ Container container =
+ Container.newInstance(report.getContainerId(), node.getNodeID(),
+ node.getHttpAddress(), report.getAllocatedResource(),
+ report.getPriority(), null);
+ ApplicationAttemptId attemptId =
+ container.getId().getApplicationAttemptId();
+ RMContainer rmContainer =
+ new RMContainerImpl(container, attemptId, node.getNodeID(),
+ applications.get(attemptId.getApplicationId()).getUser(), rmContext);
+ return rmContainer;
+ }
+
+ public SchedulerNode getSchedulerNode(NodeId nodeId) {
+ return nodes.get(nodeId);
+ }
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/AppSchedulingInfo.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/AppSchedulingInfo.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/AppSchedulingInfo.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/AppSchedulingInfo.java Mon Jun 9 03:10:06 2014
@@ -39,6 +39,8 @@ import org.apache.hadoop.yarn.api.record
import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.api.records.ResourceRequest;
import org.apache.hadoop.yarn.server.resourcemanager.rmapp.attempt.RMAppAttemptState;
+import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainer;
+import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerState;
import org.apache.hadoop.yarn.util.resource.Resources;
/**
@@ -409,4 +411,25 @@ public class AppSchedulingInfo {
// this.requests = appInfo.getRequests();
this.blacklist = appInfo.getBlackList();
}
+
+ public synchronized void recoverContainer(RMContainer rmContainer) {
+ // ContainerIdCounter on recovery will be addressed in YARN-2052
+ this.containerIdCounter.incrementAndGet();
+
+ QueueMetrics metrics = queue.getMetrics();
+ if (pending) {
+ // If there was any container to recover, the application was
+ // running from scheduler's POV.
+ pending = false;
+ metrics.runAppAttempt(applicationId, user);
+ }
+
+ // Container is completed. Skip recovering resources.
+ if (rmContainer.getState().equals(RMContainerState.COMPLETED)) {
+ return;
+ }
+
+ metrics.allocateResources(user, 1, rmContainer.getAllocatedResource(),
+ false);
+ }
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/Queue.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/Queue.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/Queue.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/Queue.java Mon Jun 9 03:10:06 2014
@@ -26,6 +26,8 @@ import org.apache.hadoop.security.UserGr
import org.apache.hadoop.yarn.api.records.QueueACL;
import org.apache.hadoop.yarn.api.records.QueueInfo;
import org.apache.hadoop.yarn.api.records.QueueUserACLInfo;
+import org.apache.hadoop.yarn.api.records.Resource;
+import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainer;
@Evolving
@LimitedPrivate("yarn")
@@ -60,4 +62,13 @@ public interface Queue {
boolean hasAccess(QueueACL acl, UserGroupInformation user);
public ActiveUsersManager getActiveUsersManager();
+
+ /**
+ * Recover the state of the queue for a given container.
+ * @param clusterResource the resource of the cluster
+ * @param schedulerAttempt the application for which the container was allocated
+ * @param rmContainer the container that was recovered.
+ */
+ public void recoverContainer(Resource clusterResource,
+ SchedulerApplicationAttempt schedulerAttempt, RMContainer rmContainer);
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/SchedulerApplicationAttempt.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/SchedulerApplicationAttempt.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/SchedulerApplicationAttempt.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/SchedulerApplicationAttempt.java Mon Jun 9 03:10:06 2014
@@ -47,6 +47,7 @@ import org.apache.hadoop.yarn.server.res
import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerEventType;
import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerImpl;
import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerReservedEvent;
+import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerState;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeCleanContainerEvent;
import org.apache.hadoop.yarn.util.resource.Resources;
@@ -534,5 +535,24 @@ public class SchedulerApplicationAttempt
appSchedulingInfo.move(newQueue);
this.queue = newQueue;
- }
+ }
+
+ public synchronized void recoverContainer(RMContainer rmContainer) {
+ // recover app scheduling info
+ appSchedulingInfo.recoverContainer(rmContainer);
+
+ if (rmContainer.getState().equals(RMContainerState.COMPLETED)) {
+ return;
+ }
+ LOG.info("SchedulerAttempt " + getApplicationAttemptId()
+ + " is recovering container " + rmContainer.getContainerId());
+ liveContainers.put(rmContainer.getContainerId(), rmContainer);
+ Resources.addTo(currentConsumption, rmContainer.getContainer()
+ .getResource());
+ // resourceLimit: updated when LeafQueue#recoverContainer#allocateResource
+ // is called.
+ // newlyAllocatedContainers.add(rmContainer);
+ // schedulingOpportunities
+ // lastScheduledContainer
+ }
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/SchedulerNode.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/SchedulerNode.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/SchedulerNode.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/SchedulerNode.java Mon Jun 9 03:10:06 2014
@@ -27,7 +27,6 @@ import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.classification.InterfaceAudience.Private;
import org.apache.hadoop.classification.InterfaceStability.Unstable;
-import org.apache.hadoop.yarn.api.records.ApplicationId;
import org.apache.hadoop.yarn.api.records.Container;
import org.apache.hadoop.yarn.api.records.ContainerId;
import org.apache.hadoop.yarn.api.records.NodeId;
@@ -35,10 +34,10 @@ import org.apache.hadoop.yarn.api.record
import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.conf.YarnConfiguration;
import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainer;
+import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerState;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNode;
import org.apache.hadoop.yarn.util.resource.Resources;
-import com.google.common.base.Preconditions;
/**
* Represents a YARN Cluster Node from the viewpoint of the scheduler.
@@ -119,13 +118,10 @@ public abstract class SchedulerNode {
* The Scheduler has allocated containers on this node to the given
* application.
*
- * @param applicationId
- * application
* @param rmContainer
* allocated container
*/
- public synchronized void allocateContainer(ApplicationId applicationId,
- RMContainer rmContainer) {
+ public synchronized void allocateContainer(RMContainer rmContainer) {
Container container = rmContainer.getContainer();
deductAvailableResource(container.getResource());
++numContainers;
@@ -166,8 +162,8 @@ public abstract class SchedulerNode {
return this.totalResourceCapability;
}
- private synchronized boolean isValidContainer(Container c) {
- if (launchedContainers.containsKey(c.getId())) {
+ public synchronized boolean isValidContainer(ContainerId containerId) {
+ if (launchedContainers.containsKey(containerId)) {
return true;
}
return false;
@@ -185,7 +181,7 @@ public abstract class SchedulerNode {
* container to be released
*/
public synchronized void releaseContainer(Container container) {
- if (!isValidContainer(container)) {
+ if (!isValidContainer(container.getId())) {
LOG.error("Invalid container released " + container);
return;
}
@@ -274,4 +270,12 @@ public abstract class SchedulerNode {
// we can only adjust available resource if total resource is changed.
Resources.addTo(this.availableResource, deltaResource);
}
+
+
+ public synchronized void recoverContainer(RMContainer rmContainer) {
+ if (rmContainer.getState().equals(RMContainerState.COMPLETED)) {
+ return;
+ }
+ allocateContainer(rmContainer);
+ }
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CSQueue.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CSQueue.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CSQueue.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CSQueue.java Mon Jun 9 03:10:06 2014
@@ -28,7 +28,6 @@ import org.apache.hadoop.security.Access
import org.apache.hadoop.security.UserGroupInformation;
import org.apache.hadoop.yarn.api.records.ApplicationAttemptId;
import org.apache.hadoop.yarn.api.records.ApplicationId;
-import org.apache.hadoop.yarn.api.records.Container;
import org.apache.hadoop.yarn.api.records.ContainerStatus;
import org.apache.hadoop.yarn.api.records.QueueACL;
import org.apache.hadoop.yarn.api.records.QueueState;
@@ -235,15 +234,6 @@ extends org.apache.hadoop.yarn.server.re
public ActiveUsersManager getActiveUsersManager();
/**
- * Recover the state of the queue
- * @param clusterResource the resource of the cluster
- * @param application the application for which the container was allocated
- * @param container the container that was recovered.
- */
- public void recoverContainer(Resource clusterResource, FiCaSchedulerApp application,
- Container container);
-
- /**
* Adds all applications in the queue and its subqueues to the given collection.
* @param apps the collection to add the applications to
*/
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CapacityScheduler.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CapacityScheduler.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CapacityScheduler.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CapacityScheduler.java Mon Jun 9 03:10:06 2014
@@ -872,6 +872,8 @@ public class CapacityScheduler extends
{
NodeAddedSchedulerEvent nodeAddedEvent = (NodeAddedSchedulerEvent)event;
addNode(nodeAddedEvent.getAddedRMNode());
+ recoverContainersOnNode(nodeAddedEvent.getContainerReports(),
+ nodeAddedEvent.getAddedRMNode());
}
break;
case NODE_REMOVED:
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java Mon Jun 9 03:10:06 2014
@@ -59,15 +59,17 @@ import org.apache.hadoop.yarn.server.res
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.NodeType;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.QueueMetrics;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerAppUtils;
+import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerApplicationAttempt;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.common.fica.FiCaSchedulerApp;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.common.fica.FiCaSchedulerNode;
-import org.apache.hadoop.yarn.server.resourcemanager.security.RMContainerTokenSecretManager;
import org.apache.hadoop.yarn.server.utils.BuilderUtils;
import org.apache.hadoop.yarn.server.utils.Lock;
import org.apache.hadoop.yarn.server.utils.Lock.NoLock;
import org.apache.hadoop.yarn.util.resource.ResourceCalculator;
import org.apache.hadoop.yarn.util.resource.Resources;
+import com.google.common.annotations.VisibleForTesting;
+
@Private
@Unstable
public class LeafQueue implements CSQueue {
@@ -564,7 +566,8 @@ public class LeafQueue implements CSQueu
"numContainers=" + getNumContainers();
}
- private synchronized User getUser(String userName) {
+ @VisibleForTesting
+ public synchronized User getUser(String userName) {
User user = users.get(userName);
if (user == null) {
user = new User();
@@ -1346,8 +1349,7 @@ public class LeafQueue implements CSQueu
}
// Inform the node
- node.allocateContainer(application.getApplicationId(),
- allocatedContainer);
+ node.allocateContainer(allocatedContainer);
LOG.info("assignedContainer" +
" application attempt=" + application.getApplicationAttemptId() +
@@ -1446,7 +1448,7 @@ public class LeafQueue implements CSQueu
}
synchronized void allocateResource(Resource clusterResource,
- FiCaSchedulerApp application, Resource resource) {
+ SchedulerApplicationAttempt application, Resource resource) {
// Update queue metrics
Resources.addTo(usedResources, resource);
CSQueueUtils.updateQueueStatistics(
@@ -1530,7 +1532,8 @@ public class LeafQueue implements CSQueu
return metrics;
}
- static class User {
+ @VisibleForTesting
+ public static class User {
Resource consumed = Resources.createResource(0, 0);
int pendingApplications = 0;
int activeApplications = 0;
@@ -1580,13 +1583,16 @@ public class LeafQueue implements CSQueu
@Override
public void recoverContainer(Resource clusterResource,
- FiCaSchedulerApp application, Container container) {
+ SchedulerApplicationAttempt attempt, RMContainer rmContainer) {
+ if (rmContainer.getState().equals(RMContainerState.COMPLETED)) {
+ return;
+ }
// Careful! Locking order is important!
synchronized (this) {
- allocateResource(clusterResource, application, container.getResource());
+ allocateResource(clusterResource, attempt, rmContainer.getContainer()
+ .getResource());
}
- getParent().recoverContainer(clusterResource, application, container);
-
+ getParent().recoverContainer(clusterResource, attempt, rmContainer);
}
/**
@@ -1613,5 +1619,4 @@ public class LeafQueue implements CSQueu
apps.add(app.getApplicationAttemptId());
}
}
-
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/ParentQueue.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/ParentQueue.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/ParentQueue.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/ParentQueue.java Mon Jun 9 03:10:06 2014
@@ -38,7 +38,6 @@ import org.apache.hadoop.security.UserGr
import org.apache.hadoop.security.authorize.AccessControlList;
import org.apache.hadoop.yarn.api.records.ApplicationAttemptId;
import org.apache.hadoop.yarn.api.records.ApplicationId;
-import org.apache.hadoop.yarn.api.records.Container;
import org.apache.hadoop.yarn.api.records.ContainerStatus;
import org.apache.hadoop.yarn.api.records.QueueACL;
import org.apache.hadoop.yarn.api.records.QueueInfo;
@@ -49,9 +48,11 @@ import org.apache.hadoop.yarn.factories.
import org.apache.hadoop.yarn.factory.providers.RecordFactoryProvider;
import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainer;
import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerEventType;
+import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerState;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.ActiveUsersManager;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.NodeType;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.QueueMetrics;
+import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerApplicationAttempt;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.common.fica.FiCaSchedulerApp;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.common.fica.FiCaSchedulerNode;
import org.apache.hadoop.yarn.util.resource.ResourceCalculator;
@@ -770,13 +771,16 @@ public class ParentQueue implements CSQu
@Override
public void recoverContainer(Resource clusterResource,
- FiCaSchedulerApp application, Container container) {
+ SchedulerApplicationAttempt attempt, RMContainer rmContainer) {
+ if (rmContainer.getState().equals(RMContainerState.COMPLETED)) {
+ return;
+ }
// Careful! Locking order is important!
synchronized (this) {
- allocateResource(clusterResource, container.getResource());
+ allocateResource(clusterResource,rmContainer.getContainer().getResource());
}
if (parent != null) {
- parent.recoverContainer(clusterResource, application, container);
+ parent.recoverContainer(clusterResource, attempt, rmContainer);
}
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/common/fica/FiCaSchedulerNode.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/common/fica/FiCaSchedulerNode.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/common/fica/FiCaSchedulerNode.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/common/fica/FiCaSchedulerNode.java Mon Jun 9 03:10:06 2014
@@ -24,6 +24,7 @@ import org.apache.commons.logging.LogFac
import org.apache.hadoop.yarn.api.records.ApplicationAttemptId;
import org.apache.hadoop.yarn.api.records.Priority;
import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainer;
+import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerState;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNode;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerApplicationAttempt;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerNode;
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/event/NodeAddedSchedulerEvent.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/event/NodeAddedSchedulerEvent.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/event/NodeAddedSchedulerEvent.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/event/NodeAddedSchedulerEvent.java Mon Jun 9 03:10:06 2014
@@ -18,19 +18,34 @@
package org.apache.hadoop.yarn.server.resourcemanager.scheduler.event;
+import java.util.List;
+
+import org.apache.hadoop.yarn.server.api.protocolrecords.NMContainerStatus;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNode;
public class NodeAddedSchedulerEvent extends SchedulerEvent {
private final RMNode rmNode;
+ private final List<NMContainerStatus> containerReports;
public NodeAddedSchedulerEvent(RMNode rmNode) {
super(SchedulerEventType.NODE_ADDED);
this.rmNode = rmNode;
+ this.containerReports = null;
+ }
+
+ public NodeAddedSchedulerEvent(RMNode rmNode,
+ List<NMContainerStatus> containerReports) {
+ super(SchedulerEventType.NODE_ADDED);
+ this.rmNode = rmNode;
+ this.containerReports = containerReports;
}
public RMNode getAddedRMNode() {
return rmNode;
}
+ public List<NMContainerStatus> getContainerReports() {
+ return containerReports;
+ }
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/AppSchedulable.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/AppSchedulable.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/AppSchedulable.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/AppSchedulable.java Mon Jun 9 03:10:06 2014
@@ -264,8 +264,7 @@ public class AppSchedulable extends Sche
}
// Inform the node
- node.allocateContainer(app.getApplicationId(),
- allocatedContainer);
+ node.allocateContainer(allocatedContainer);
// If this container is used to run AM, update the leaf queue's AM usage
if (app.getLiveContainers().size() == 1 &&
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/FSLeafQueue.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/FSLeafQueue.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/FSLeafQueue.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/FSLeafQueue.java Mon Jun 9 03:10:06 2014
@@ -36,6 +36,7 @@ import org.apache.hadoop.yarn.api.record
import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainer;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.ActiveUsersManager;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerAppUtils;
+import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerApplicationAttempt;
import org.apache.hadoop.yarn.util.resource.Resources;
@Private
@@ -318,4 +319,10 @@ public class FSLeafQueue extends FSQueue
Resources.addTo(amResourceUsage, amResource);
}
}
+
+ @Override
+ public void recoverContainer(Resource clusterResource,
+ SchedulerApplicationAttempt schedulerAttempt, RMContainer rmContainer) {
+ // TODO Auto-generated method stub
+ }
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/FSParentQueue.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/FSParentQueue.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/FSParentQueue.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fair/FSParentQueue.java Mon Jun 9 03:10:06 2014
@@ -35,7 +35,9 @@ import org.apache.hadoop.yarn.api.record
import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainer;
import org.apache.hadoop.yarn.util.resource.Resources;
+import org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainer;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.ActiveUsersManager;
+import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerApplicationAttempt;
@Private
@Unstable
@@ -228,4 +230,11 @@ public class FSParentQueue extends FSQue
// Should never be called since all applications are submitted to LeafQueues
return null;
}
+
+ @Override
+ public void recoverContainer(Resource clusterResource,
+ SchedulerApplicationAttempt schedulerAttempt, RMContainer rmContainer) {
+ // TODO Auto-generated method stub
+
+ }
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fifo/FifoScheduler.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fifo/FifoScheduler.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fifo/FifoScheduler.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/fifo/FifoScheduler.java Mon Jun 9 03:10:06 2014
@@ -76,6 +76,7 @@ import org.apache.hadoop.yarn.server.res
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.QueueMetrics;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerAppUtils;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerApplication;
+import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerApplicationAttempt;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerApplicationAttempt.ContainersAndNMTokensAllocation;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerUtils;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.common.fica.FiCaSchedulerApp;
@@ -178,6 +179,17 @@ public class FifoScheduler extends
public ActiveUsersManager getActiveUsersManager() {
return activeUsersManager;
}
+
+ @Override
+ public void recoverContainer(Resource clusterResource,
+ SchedulerApplicationAttempt schedulerAttempt, RMContainer rmContainer) {
+ if (rmContainer.getState().equals(RMContainerState.COMPLETED)) {
+ return;
+ }
+ increaseUsedResources(rmContainer);
+ updateAppHeadRoom(schedulerAttempt);
+ updateAvailableResourcesMetrics();
+ }
};
public FifoScheduler() {
@@ -488,7 +500,7 @@ public class FifoScheduler extends
if (attempt == null) {
continue;
}
- attempt.setHeadroom(Resources.subtract(clusterResource, usedResource));
+ updateAppHeadRoom(attempt);
}
}
@@ -659,11 +671,10 @@ public class FifoScheduler extends
application.allocate(type, node, priority, request, container);
// Inform the node
- node.allocateContainer(application.getApplicationId(),
- rmContainer);
+ node.allocateContainer(rmContainer);
// Update usage for this container
- Resources.addTo(usedResource, capability);
+ increaseUsedResources(rmContainer);
}
}
@@ -707,9 +718,22 @@ public class FifoScheduler extends
LOG.debug("Node after allocation " + rmNode.getNodeID() + " resource = "
+ node.getAvailableResource());
}
-
- metrics.setAvailableResourcesToQueue(
- Resources.subtract(clusterResource, usedResource));
+
+ updateAvailableResourcesMetrics();
+ }
+
+ private void increaseUsedResources(RMContainer rmContainer) {
+ Resources.addTo(usedResource, rmContainer.getAllocatedResource());
+ }
+
+ private void updateAppHeadRoom(SchedulerApplicationAttempt schedulerAttempt) {
+ schedulerAttempt.setHeadroom(Resources.subtract(clusterResource,
+ usedResource));
+ }
+
+ private void updateAvailableResourcesMetrics() {
+ metrics.setAvailableResourcesToQueue(Resources.subtract(clusterResource,
+ usedResource));
}
@Override
@@ -719,6 +743,9 @@ public class FifoScheduler extends
{
NodeAddedSchedulerEvent nodeAddedEvent = (NodeAddedSchedulerEvent)event;
addNode(nodeAddedEvent.getAddedRMNode());
+ recoverContainersOnNode(nodeAddedEvent.getContainerReports(),
+ nodeAddedEvent.getAddedRMNode());
+
}
break;
case NODE_REMOVED:
@@ -923,4 +950,8 @@ public class FifoScheduler extends
return null;
}
}
+
+ public Resource getUsedResource() {
+ return usedResource;
+ }
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/MockRM.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/MockRM.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/MockRM.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/MockRM.java Mon Jun 9 03:10:06 2014
@@ -21,10 +21,9 @@ package org.apache.hadoop.yarn.server.re
import java.io.IOException;
import java.nio.ByteBuffer;
import java.security.PrivilegedAction;
+import java.util.List;
import java.util.Map;
-import org.junit.Assert;
-
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.io.DataOutputBuffer;
import org.apache.hadoop.security.Credentials;
@@ -47,12 +46,14 @@ import org.apache.hadoop.yarn.api.record
import org.apache.hadoop.yarn.api.records.ContainerId;
import org.apache.hadoop.yarn.api.records.ContainerLaunchContext;
import org.apache.hadoop.yarn.api.records.ContainerState;
+import org.apache.hadoop.yarn.api.records.ContainerStatus;
import org.apache.hadoop.yarn.api.records.FinalApplicationStatus;
import org.apache.hadoop.yarn.api.records.NodeId;
import org.apache.hadoop.yarn.api.records.NodeState;
import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.conf.YarnConfiguration;
import org.apache.hadoop.yarn.exceptions.YarnException;
+import org.apache.hadoop.yarn.server.api.protocolrecords.NMContainerStatus;
import org.apache.hadoop.yarn.server.resourcemanager.amlauncher.AMLauncherEvent;
import org.apache.hadoop.yarn.server.resourcemanager.amlauncher.ApplicationMasterLauncher;
import org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore;
@@ -69,6 +70,10 @@ import org.apache.hadoop.yarn.server.res
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeEvent;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeEventType;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeImpl;
+import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeStartedEvent;
+import org.apache.hadoop.yarn.server.resourcemanager.scheduler.AbstractYarnScheduler;
+import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerApplicationAttempt;
+import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerNode;
import org.apache.hadoop.yarn.server.resourcemanager.security.ClientToAMTokenSecretManagerInRM;
import org.apache.hadoop.yarn.server.resourcemanager.security.NMTokenSecretManagerInRM;
import org.apache.hadoop.yarn.server.resourcemanager.security.RMContainerTokenSecretManager;
@@ -76,6 +81,7 @@ import org.apache.hadoop.yarn.util.Recor
import org.apache.log4j.Level;
import org.apache.log4j.LogManager;
import org.apache.log4j.Logger;
+import org.junit.Assert;
@SuppressWarnings("unchecked")
public class MockRM extends ResourceManager {
@@ -144,11 +150,26 @@ public class MockRM extends ResourceMana
}
}
+ public void waitForContainerToComplete(RMAppAttempt attempt,
+ NMContainerStatus completedContainer) throws InterruptedException {
+ while (true) {
+ List<ContainerStatus> containers = attempt.getJustFinishedContainers();
+ System.out.println("Received completed containers " + containers);
+ for (ContainerStatus container : containers) {
+ if (container.getContainerId().equals(
+ completedContainer.getContainerId())) {
+ return;
+ }
+ }
+ Thread.sleep(200);
+ }
+ }
+
public void waitForState(MockNM nm, ContainerId containerId,
RMContainerState containerState) throws Exception {
RMContainer container = getResourceScheduler().getRMContainer(containerId);
int timeoutSecs = 0;
- while(container == null && timeoutSecs++ < 20) {
+ while(container == null && timeoutSecs++ < 100) {
nm.nodeHeartbeat(true);
container = getResourceScheduler().getRMContainer(containerId);
System.out.println("Waiting for container " + containerId + " to be allocated.");
@@ -333,7 +354,7 @@ public class MockRM extends ResourceMana
public void sendNodeStarted(MockNM nm) throws Exception {
RMNodeImpl node = (RMNodeImpl) getRMContext().getRMNodes().get(
nm.getNodeId());
- node.handle(new RMNodeEvent(nm.getNodeId(), RMNodeEventType.STARTED));
+ node.handle(new RMNodeStartedEvent(nm.getNodeId(), null));
}
public void sendNodeLost(MockNM nm) throws Exception {
@@ -542,4 +563,12 @@ public class MockRM extends ResourceMana
.newInstance(appId));
return response.getApplicationReport();
}
+
+ // Explicitly reset queue metrics for testing.
+ @SuppressWarnings("static-access")
+ public void clearQueueMetrics(RMApp app) {
+ ((AbstractYarnScheduler<SchedulerApplicationAttempt, SchedulerNode>) getResourceScheduler())
+ .getSchedulerApplications().get(app.getApplicationId()).getQueue()
+ .getMetrics().clearQueueMetrics();
+ }
}
Modified: hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/TestFifoScheduler.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/TestFifoScheduler.java?rev=1601304&r1=1601303&r2=1601304&view=diff
==============================================================================
--- hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/TestFifoScheduler.java (original)
+++ hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/TestFifoScheduler.java Mon Jun 9 03:10:06 2014
@@ -18,16 +18,16 @@
package org.apache.hadoop.yarn.server.resourcemanager;
+import static org.junit.Assert.assertTrue;
+import static org.junit.Assert.fail;
+
import java.util.ArrayList;
import java.util.Collections;
import java.util.List;
-import org.junit.Assert;
-
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.conf.Configuration;
-import org.apache.hadoop.yarn.exceptions.YarnRuntimeException;
import org.apache.hadoop.yarn.api.protocolrecords.AllocateResponse;
import org.apache.hadoop.yarn.api.records.ApplicationAttemptId;
import org.apache.hadoop.yarn.api.records.ApplicationId;
@@ -37,6 +37,7 @@ import org.apache.hadoop.yarn.api.record
import org.apache.hadoop.yarn.api.records.ContainerStatus;
import org.apache.hadoop.yarn.api.records.ResourceRequest;
import org.apache.hadoop.yarn.conf.YarnConfiguration;
+import org.apache.hadoop.yarn.exceptions.YarnRuntimeException;
import org.apache.hadoop.yarn.server.resourcemanager.rmapp.RMApp;
import org.apache.hadoop.yarn.server.resourcemanager.rmapp.attempt.RMAppAttempt;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNode;
@@ -56,11 +57,10 @@ import org.apache.hadoop.yarn.util.resou
import org.apache.log4j.Level;
import org.apache.log4j.LogManager;
import org.apache.log4j.Logger;
+import org.junit.Assert;
import org.junit.BeforeClass;
import org.junit.Test;
-
-import static org.junit.Assert.assertTrue;
-import static org.junit.Assert.fail;
+import static org.mockito.Mockito.mock;
public class TestFifoScheduler {
private static final Log LOG = LogFactory.getLog(TestFifoScheduler.class);
@@ -298,7 +298,10 @@ public class TestFifoScheduler {
FifoScheduler fs = new FifoScheduler();
fs.init(conf);
fs.start();
+ // mock rmContext to avoid NPE.
+ RMContext context = mock(RMContext.class);
fs.reinitialize(conf, null);
+ fs.setRMContext(context);
RMNode n1 =
MockNodes.newNodeInfo(0, MockNodes.newResource(4 * GB), 1, "127.0.0.2");