YARN-1506. Changed RMNode/SchedulerNode to update resource with event notification. Contributed by Junping Du

This commit is contained in:
Jian He 2014-08-29 23:05:51 -07:00
parent 270a271f53
commit 5c14bc426b
28 changed files with 617 additions and 187 deletions

View File

@ -32,7 +32,6 @@ import org.apache.hadoop.yarn.api.records.ContainerStatus;
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.api.records.ResourceOption;
import org.apache.hadoop.yarn.server.api.protocolrecords.NodeHeartbeatResponse;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNode;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode
@ -55,7 +54,7 @@ public class NodeInfo {
private String nodeAddr;
private String httpAddress;
private int cmdPort;
private volatile ResourceOption perNode;
private volatile Resource perNode;
private String rackName;
private String healthReport;
private NodeState state;
@ -63,7 +62,7 @@ public class NodeInfo {
private List<ApplicationId> toCleanUpApplications;
public FakeRMNodeImpl(NodeId nodeId, String nodeAddr, String httpAddress,
ResourceOption perNode, String rackName, String healthReport,
Resource perNode, String rackName, String healthReport,
int cmdPort, String hostName, NodeState state) {
this.nodeId = nodeId;
this.nodeAddr = nodeAddr;
@ -111,10 +110,6 @@ public class NodeInfo {
}
public Resource getTotalCapability() {
return perNode.getResource();
}
public ResourceOption getResourceOption() {
return perNode;
}
@ -161,30 +156,24 @@ public class NodeInfo {
@Override
public String getNodeManagerVersion() {
// TODO Auto-generated method stub
return null;
}
@Override
public void setResourceOption(ResourceOption resourceOption) {
perNode = resourceOption;
}
}
public static RMNode newNodeInfo(String rackName, String hostName,
final ResourceOption resourceOption, int port) {
final Resource resource, int port) {
final NodeId nodeId = newNodeID(hostName, port);
final String nodeAddr = hostName + ":" + port;
final String httpAddress = hostName;
return new FakeRMNodeImpl(nodeId, nodeAddr, httpAddress,
resourceOption, rackName, "Me good",
resource, rackName, "Me good",
port, hostName, null);
}
public static RMNode newNodeInfo(String rackName, String hostName,
final Resource resource) {
return newNodeInfo(rackName, hostName, ResourceOption.newInstance(resource,
RMNode.OVER_COMMIT_TIMEOUT_MILLIS_DEFAULT), NODE_ID++);
return newNodeInfo(rackName, hostName, resource, NODE_ID++);
}
}

View File

@ -26,7 +26,6 @@ import org.apache.hadoop.yarn.api.records.ContainerId;
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.api.records.ResourceOption;
import org.apache.hadoop.yarn.server.api.protocolrecords.NodeHeartbeatResponse;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNode;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode
@ -148,14 +147,4 @@ public class RMNodeWrapper implements RMNode {
return node.getNodeManagerVersion();
}
@Override
public void setResourceOption(ResourceOption resourceOption) {
node.setResourceOption(resourceOption);
}
@Override
public ResourceOption getResourceOption() {
return node.getResourceOption();
}
}

View File

@ -166,6 +166,9 @@ Release 2.6.0 - UNRELEASED
YARN-2360. Fair Scheduler: Display dynamic fair share for queues on the
scheduler page. (Ashwin Shankar and Wei Yan via kasha)
YARN-1506. Changed RMNode/SchedulerNode to update resource with event
notification. (Junping Du via jianhe)
OPTIMIZATIONS
BUG FIXES

View File

@ -62,6 +62,8 @@ public abstract class ResourceOption {
@Evolving
protected abstract void setOverCommitTimeout(int overCommitTimeout);
@Private
@Evolving
protected abstract void build();
@Override

View File

@ -17,9 +17,10 @@
*/
package org.apache.hadoop.yarn.server.api.protocolrecords;
import org.apache.hadoop.classification.InterfaceAudience.Public;
import org.apache.hadoop.classification.InterfaceAudience.Private;
import org.apache.hadoop.classification.InterfaceStability.Evolving;
import org.apache.hadoop.yarn.server.api.ResourceManagerAdministrationProtocol;
import org.apache.hadoop.yarn.util.Records;
/**
* <p>The response sent by the <code>ResourceManager</code> to Admin client on
@ -30,8 +31,13 @@ import org.apache.hadoop.yarn.server.api.ResourceManagerAdministrationProtocol;
* @see ResourceManagerAdministrationProtocol#updateNodeResource(
* UpdateNodeResourceRequest)
*/
@Public
@Private
@Evolving
public interface UpdateNodeResourceResponse {
public abstract class UpdateNodeResourceResponse {
public static UpdateNodeResourceResponse newInstance(){
UpdateNodeResourceResponse response =
Records.newRecord(UpdateNodeResourceResponse.class);
return response;
}
}

View File

@ -22,14 +22,15 @@ import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.api.records.ResourceOption;
import org.apache.hadoop.yarn.proto.YarnProtos.ResourceProto;
import org.apache.hadoop.yarn.proto.YarnProtos.ResourceOptionProto;
import org.apache.hadoop.yarn.proto.YarnProtos.ResourceOptionProtoOrBuilder;
import com.google.common.base.Preconditions;
public class ResourceOptionPBImpl extends ResourceOption {
ResourceOptionProto proto = null;
ResourceOptionProto proto = ResourceOptionProto.getDefaultInstance();
ResourceOptionProto.Builder builder = null;
private Resource resource = null;
boolean viaProto = false;
public ResourceOptionPBImpl() {
builder = ResourceOptionProto.newBuilder();
@ -37,39 +38,46 @@ public class ResourceOptionPBImpl extends ResourceOption {
public ResourceOptionPBImpl(ResourceOptionProto proto) {
this.proto = proto;
this.resource = convertFromProtoFormat(proto.getResource());
viaProto = true;
}
public ResourceOptionProto getProto() {
proto = viaProto ? proto : builder.build();
viaProto = true;
return proto;
}
@Override
public Resource getResource() {
return this.resource;
ResourceOptionProtoOrBuilder p = viaProto ? proto : builder;
return convertFromProtoFormat(p.getResource());
}
@Override
protected void setResource(Resource resource) {
if (resource != null) {
Preconditions.checkNotNull(builder);
maybeInitBuilder();
builder.setResource(convertToProtoFormat(resource));
}
this.resource = resource;
}
@Override
public int getOverCommitTimeout() {
Preconditions.checkNotNull(proto);
return proto.getOverCommitTimeout();
ResourceOptionProtoOrBuilder p = viaProto ? proto : builder;
return p.getOverCommitTimeout();
}
@Override
protected void setOverCommitTimeout(int overCommitTimeout) {
Preconditions.checkNotNull(builder);
maybeInitBuilder();
builder.setOverCommitTimeout(overCommitTimeout);
}
private void maybeInitBuilder() {
if (viaProto || builder == null) {
builder = ResourceOptionProto.newBuilder(proto);
}
viaProto = false;
}
private ResourceProto convertToProtoFormat(
Resource resource) {
return ((ResourcePBImpl)resource).getProto();
@ -83,6 +91,7 @@ public class ResourceOptionPBImpl extends ResourceOption {
@Override
protected void build() {
proto = builder.build();
viaProto = true;
builder = null;
}

View File

@ -20,7 +20,7 @@ package org.apache.hadoop.yarn.server.api.protocolrecords.impl.pb;
import org.apache.hadoop.yarn.proto.YarnServerResourceManagerServiceProtos.UpdateNodeResourceResponseProto;
import org.apache.hadoop.yarn.server.api.protocolrecords.UpdateNodeResourceResponse;
public class UpdateNodeResourceResponsePBImpl implements UpdateNodeResourceResponse {
public class UpdateNodeResourceResponsePBImpl extends UpdateNodeResourceResponse {
UpdateNodeResourceResponseProto proto = UpdateNodeResourceResponseProto.getDefaultInstance();
UpdateNodeResourceResponseProto.Builder builder = null;

View File

@ -72,6 +72,7 @@ import org.apache.hadoop.yarn.server.api.protocolrecords.RefreshUserToGroupsMapp
import org.apache.hadoop.yarn.server.api.protocolrecords.UpdateNodeResourceRequest;
import org.apache.hadoop.yarn.server.api.protocolrecords.UpdateNodeResourceResponse;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNode;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeResourceUpdateEvent;
import org.apache.hadoop.yarn.server.resourcemanager.security.authorize.RMPolicyProvider;
import com.google.common.annotations.VisibleForTesting;
@ -513,9 +514,20 @@ public class AdminService extends CompositeService implements
return UserGroupInformation.createRemoteUser(user).getGroupNames();
}
@SuppressWarnings("unchecked")
@Override
public UpdateNodeResourceResponse updateNodeResource(
UpdateNodeResourceRequest request) throws YarnException, IOException {
String argName = "updateNodeResource";
UserGroupInformation user = checkAcls(argName);
if (!isRMActive()) {
RMAuditLogger.logFailure(user.getShortUserName(), argName,
adminAcl.toString(), "AdminService",
"ResourceManager is not active. Can not update node resource.");
throwStandbyException();
}
Map<NodeId, ResourceOption> nodeResourceMap = request.getNodeResourceMap();
Set<NodeId> nodeIds = nodeResourceMap.keySet();
// verify nodes are all valid first.
@ -536,20 +548,30 @@ public class AdminService extends CompositeService implements
// Notice: it is still possible to have invalid NodeIDs as nodes decommission
// may happen just at the same time. This time, only log and skip absent
// nodes without throwing any exceptions.
boolean allSuccess = true;
for (Map.Entry<NodeId, ResourceOption> entry : nodeResourceMap.entrySet()) {
ResourceOption newResourceOption = entry.getValue();
NodeId nodeId = entry.getKey();
RMNode node = this.rmContext.getRMNodes().get(nodeId);
if (node == null) {
LOG.warn("Resource update get failed on an unrecognized node: " + nodeId);
allSuccess = false;
} else {
node.setResourceOption(newResourceOption);
LOG.info("Update resource successfully on node(" + node.getNodeID()
// update resource to RMNode
this.rmContext.getDispatcher().getEventHandler()
.handle(new RMNodeResourceUpdateEvent(nodeId, newResourceOption));
LOG.info("Update resource on node(" + node.getNodeID()
+ ") with resource(" + newResourceOption.toString() + ")");
}
}
UpdateNodeResourceResponse response = recordFactory.newRecordInstance(
UpdateNodeResourceResponse.class);
if (allSuccess) {
RMAuditLogger.logSuccess(user.getShortUserName(), argName,
"AdminService");
}
UpdateNodeResourceResponse response =
UpdateNodeResourceResponse.newInstance();
return response;
}

View File

@ -36,7 +36,6 @@ import org.apache.hadoop.yarn.api.records.ContainerState;
import org.apache.hadoop.yarn.api.records.ContainerStatus;
import org.apache.hadoop.yarn.api.records.NodeId;
import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.api.records.ResourceOption;
import org.apache.hadoop.yarn.conf.YarnConfiguration;
import org.apache.hadoop.yarn.exceptions.YarnException;
import org.apache.hadoop.yarn.exceptions.YarnRuntimeException;
@ -299,8 +298,7 @@ public class ResourceTrackerService extends AbstractService implements
.getCurrentKey());
RMNode rmNode = new RMNodeImpl(nodeId, rmContext, host, cmPort, httpPort,
resolve(host), ResourceOption.newInstance(capability, RMNode.OVER_COMMIT_TIMEOUT_MILLIS_DEFAULT),
nodeManagerVersion);
resolve(host), capability, nodeManagerVersion);
RMNode oldNode = this.rmContext.getRMNodes().putIfAbsent(nodeId, rmNode);
if (oldNode == null) {

View File

@ -101,18 +101,6 @@ public interface RMNode {
*/
public Resource getTotalCapability();
/**
* Set resource option with total available resource and overCommitTimoutMillis
* @param resourceOption
*/
public void setResourceOption(ResourceOption resourceOption);
/**
* resource option with total available resource and overCommitTimoutMillis
* @return ResourceOption
*/
public ResourceOption getResourceOption();
/**
* The rack name for this node manager.
* @return the rack name.

View File

@ -25,6 +25,9 @@ public enum RMNodeEventType {
// Source: AdminService
DECOMMISSION,
// Source: AdminService, ResourceTrackerService
RESOURCE_UPDATE,
// ResourceTrackerService
STATUS_UPDATE,
REBOOTING,

View File

@ -58,6 +58,7 @@ import org.apache.hadoop.yarn.server.resourcemanager.rmapp.RMApp;
import org.apache.hadoop.yarn.server.resourcemanager.rmapp.RMAppRunningOnNodeEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeAddedSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeRemovedSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeResourceUpdateSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeUpdateSchedulerEvent;
import org.apache.hadoop.yarn.server.utils.BuilderUtils.ContainerIdComparator;
import org.apache.hadoop.yarn.state.InvalidStateTransitonException;
@ -96,7 +97,7 @@ public class RMNodeImpl implements RMNode, EventHandler<RMNodeEvent> {
private int httpPort;
private final String nodeAddress; // The containerManager address
private String httpAddress;
private volatile ResourceOption resourceOption;
private volatile Resource totalCapability;
private final Node node;
private String healthReport;
@ -129,6 +130,9 @@ public class RMNodeImpl implements RMNode, EventHandler<RMNodeEvent> {
//Transitions from NEW state
.addTransition(NodeState.NEW, NodeState.RUNNING,
RMNodeEventType.STARTED, new AddNodeTransition())
.addTransition(NodeState.NEW, NodeState.NEW,
RMNodeEventType.RESOURCE_UPDATE,
new UpdateNodeResourceWhenUnusableTransition())
//Transitions from RUNNING state
.addTransition(NodeState.RUNNING,
@ -149,6 +153,23 @@ public class RMNodeImpl implements RMNode, EventHandler<RMNodeEvent> {
RMNodeEventType.CLEANUP_CONTAINER, new CleanUpContainerTransition())
.addTransition(NodeState.RUNNING, NodeState.RUNNING,
RMNodeEventType.RECONNECTED, new ReconnectNodeTransition())
.addTransition(NodeState.RUNNING, NodeState.RUNNING,
RMNodeEventType.RESOURCE_UPDATE, new UpdateNodeResourceWhenRunningTransition())
//Transitions from REBOOTED state
.addTransition(NodeState.REBOOTED, NodeState.REBOOTED,
RMNodeEventType.RESOURCE_UPDATE,
new UpdateNodeResourceWhenUnusableTransition())
//Transitions from DECOMMISSIONED state
.addTransition(NodeState.DECOMMISSIONED, NodeState.DECOMMISSIONED,
RMNodeEventType.RESOURCE_UPDATE,
new UpdateNodeResourceWhenUnusableTransition())
//Transitions from LOST state
.addTransition(NodeState.LOST, NodeState.LOST,
RMNodeEventType.RESOURCE_UPDATE,
new UpdateNodeResourceWhenUnusableTransition())
//Transitions from UNHEALTHY state
.addTransition(NodeState.UNHEALTHY,
@ -169,6 +190,8 @@ public class RMNodeImpl implements RMNode, EventHandler<RMNodeEvent> {
RMNodeEventType.CLEANUP_APP, new CleanUpAppTransition())
.addTransition(NodeState.UNHEALTHY, NodeState.UNHEALTHY,
RMNodeEventType.CLEANUP_CONTAINER, new CleanUpContainerTransition())
.addTransition(NodeState.UNHEALTHY, NodeState.UNHEALTHY,
RMNodeEventType.RESOURCE_UPDATE, new UpdateNodeResourceWhenUnusableTransition())
// create the topology tables
.installTopology();
@ -177,13 +200,13 @@ public class RMNodeImpl implements RMNode, EventHandler<RMNodeEvent> {
RMNodeEvent> stateMachine;
public RMNodeImpl(NodeId nodeId, RMContext context, String hostName,
int cmPort, int httpPort, Node node, ResourceOption resourceOption, String nodeManagerVersion) {
int cmPort, int httpPort, Node node, Resource capability, String nodeManagerVersion) {
this.nodeId = nodeId;
this.context = context;
this.hostName = hostName;
this.commandPort = cmPort;
this.httpPort = httpPort;
this.resourceOption = resourceOption;
this.totalCapability = capability;
this.nodeAddress = hostName + ":" + cmPort;
this.httpAddress = hostName + ":" + httpPort;
this.node = node;
@ -239,17 +262,7 @@ public class RMNodeImpl implements RMNode, EventHandler<RMNodeEvent> {
@Override
public Resource getTotalCapability() {
return this.resourceOption.getResource();
}
@Override
public void setResourceOption(ResourceOption resourceOption) {
this.resourceOption = resourceOption;
}
@Override
public ResourceOption getResourceOption(){
return this.resourceOption;
return this.totalCapability;
}
@Override
@ -474,6 +487,13 @@ public class RMNodeImpl implements RMNode, EventHandler<RMNodeEvent> {
.handle(new RMAppRunningOnNodeEvent(appId, nodeId));
}
private static void updateNodeResourceFromEvent(RMNodeImpl rmNode,
RMNodeResourceUpdateEvent event){
ResourceOption resourceOption = event.getResourceOption();
// Set resource on RMNode
rmNode.totalCapability = resourceOption.getResource();
}
public static class AddNodeTransition implements
SingleArcTransition<RMNodeImpl, RMNodeEvent> {
@ -526,7 +546,7 @@ public class RMNodeImpl implements RMNode, EventHandler<RMNodeEvent> {
rmNode.nodeManagerVersion = newNode.getNodeManagerVersion();
rmNode.httpPort = newNode.getHttpPort();
rmNode.httpAddress = newNode.getHttpAddress();
rmNode.resourceOption = newNode.getResourceOption();
rmNode.totalCapability = newNode.getTotalCapability();
// Reset heartbeat ID since node just restarted.
rmNode.getLastNodeHeartBeatResponse().setResponseId(0);
@ -540,6 +560,40 @@ public class RMNodeImpl implements RMNode, EventHandler<RMNodeEvent> {
rmNode.context.getDispatcher().getEventHandler().handle(
new NodesListManagerEvent(
NodesListManagerEventType.NODE_USABLE, rmNode));
if (rmNode.getState().equals(NodeState.RUNNING)) {
// Update scheduler node's capacity for reconnect node.
rmNode.context.getDispatcher().getEventHandler().handle(
new NodeResourceUpdateSchedulerEvent(rmNode,
ResourceOption.newInstance(rmNode.totalCapability, -1)));
}
}
}
public static class UpdateNodeResourceWhenRunningTransition
implements SingleArcTransition<RMNodeImpl, RMNodeEvent> {
@Override
public void transition(RMNodeImpl rmNode, RMNodeEvent event) {
RMNodeResourceUpdateEvent updateEvent = (RMNodeResourceUpdateEvent)event;
updateNodeResourceFromEvent(rmNode, updateEvent);
// Notify new resourceOption to scheduler
rmNode.context.getDispatcher().getEventHandler().handle(
new NodeResourceUpdateSchedulerEvent(rmNode, updateEvent.getResourceOption()));
}
}
public static class UpdateNodeResourceWhenUnusableTransition
implements SingleArcTransition<RMNodeImpl, RMNodeEvent> {
@Override
public void transition(RMNodeImpl rmNode, RMNodeEvent event) {
// The node is not usable, only log a warn message
LOG.warn("Try to update resource on a "+ rmNode.getState().toString() +
" node: "+rmNode.toString());
updateNodeResourceFromEvent(rmNode, (RMNodeResourceUpdateEvent)event);
// No need to notify scheduler as schedulerNode is not function now
// and can sync later from RMnode.
}
}

View File

@ -0,0 +1,37 @@
/**
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.hadoop.yarn.server.resourcemanager.rmnode;
import org.apache.hadoop.yarn.api.records.NodeId;
import org.apache.hadoop.yarn.api.records.ResourceOption;
public class RMNodeResourceUpdateEvent extends RMNodeEvent {
private final ResourceOption resourceOption;
public RMNodeResourceUpdateEvent(NodeId nodeId, ResourceOption resourceOption) {
super(nodeId, RMNodeEventType.RESOURCE_UPDATE);
this.resourceOption = resourceOption;
}
public ResourceOption getResourceOption() {
return resourceOption;
}
}

View File

@ -41,6 +41,7 @@ import org.apache.hadoop.yarn.api.records.ContainerState;
import org.apache.hadoop.yarn.api.records.ContainerStatus;
import org.apache.hadoop.yarn.api.records.NodeId;
import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.api.records.ResourceOption;
import org.apache.hadoop.yarn.api.records.ResourceRequest;
import org.apache.hadoop.yarn.conf.YarnConfiguration;
import org.apache.hadoop.yarn.exceptions.YarnException;
@ -473,4 +474,32 @@ public abstract class AbstractYarnScheduler
.handle(new RMAppEvent(app.getApplicationId(), RMAppEventType.KILL));
}
}
/**
* Process resource update on a node.
*/
public synchronized void updateNodeResource(RMNode nm,
ResourceOption resourceOption) {
SchedulerNode node = getSchedulerNode(nm.getNodeID());
Resource newResource = resourceOption.getResource();
Resource oldResource = node.getTotalResource();
if(!oldResource.equals(newResource)) {
// Log resource change
LOG.info("Update resource on node: " + node.getNodeName()
+ " from: " + oldResource + ", to: "
+ newResource);
// update resource to node
node.setTotalResource(newResource);
// update resource to clusterResource
Resources.subtractFrom(clusterResource, oldResource);
Resources.addTo(clusterResource, newResource);
} else {
// Log resource change
LOG.warn("Update resource on node: " + node.getNodeName()
+ " with the same resource: " + newResource);
}
}
}

View File

@ -77,6 +77,16 @@ public abstract class SchedulerNode {
return this.rmNode;
}
/**
* Set total resources on the node.
* @param resource total resources on the node.
*/
public synchronized void setTotalResource(Resource resource){
this.totalResourceCapability = resource;
this.availableResource = Resources.subtract(totalResourceCapability,
this.usedResource);
}
/**
* Get the ID of the node which contains both its hostname and port.
*
@ -158,7 +168,7 @@ public abstract class SchedulerNode {
*
* @return total resources on the node.
*/
public Resource getTotalResource() {
public synchronized Resource getTotalResource() {
return this.totalResourceCapability;
}
@ -259,19 +269,6 @@ public abstract class SchedulerNode {
this.reservedContainer = reservedContainer;
}
/**
* Apply delta resource on node's available resource.
*
* @param deltaResource
* the delta of resource need to apply to node
*/
public synchronized void
applyDeltaOnAvailableResource(Resource deltaResource) {
// 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;

View File

@ -19,7 +19,6 @@ package org.apache.hadoop.yarn.server.resourcemanager.scheduler;
import java.util.List;
import org.apache.commons.logging.Log;
import org.apache.hadoop.classification.InterfaceAudience.Private;
import org.apache.hadoop.classification.InterfaceStability.Unstable;
import org.apache.hadoop.yarn.api.records.ContainerExitStatus;
@ -147,42 +146,6 @@ public class SchedulerUtils {
ask.setCapability(normalized);
}
/**
* Update resource in SchedulerNode if any resource change in RMNode.
* @param node SchedulerNode with old resource view
* @param rmNode RMNode with new resource view
* @param clusterResource the cluster's resource that need to update
* @param log Scheduler's log for resource change
* @return true if the resources have changed
*/
public static boolean updateResourceIfChanged(SchedulerNode node,
RMNode rmNode, Resource clusterResource, Log log) {
boolean result = false;
Resource oldAvailableResource = node.getAvailableResource();
Resource newAvailableResource = Resources.subtract(
rmNode.getTotalCapability(), node.getUsedResource());
if (!newAvailableResource.equals(oldAvailableResource)) {
result = true;
Resource deltaResource = Resources.subtract(newAvailableResource,
oldAvailableResource);
// Reflect resource change to scheduler node.
node.applyDeltaOnAvailableResource(deltaResource);
// Reflect resource change to clusterResource.
Resources.addTo(clusterResource, deltaResource);
// TODO process resource over-commitment case (allocated containers
// > total capacity) in different option by getting value of
// overCommitTimeoutMillis.
// Log resource change
log.info("Resource change on node: " + rmNode.getNodeAddress()
+ " with delta: CPU: " + deltaResource.getMemory() + "core, Memory: "
+ deltaResource.getMemory() +"MB");
}
return result;
}
/**
* Utility method to normalize a list of resource requests, by insuring that
* the memory for each request is a multiple of minMemory and is not zero.

View File

@ -50,6 +50,8 @@ import org.apache.hadoop.yarn.api.records.NodeId;
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.api.records.ResourceOption;
import org.apache.hadoop.yarn.api.records.ResourceRequest;
import org.apache.hadoop.yarn.conf.YarnConfiguration;
import org.apache.hadoop.yarn.exceptions.YarnException;
@ -82,6 +84,7 @@ import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.AppRemovedS
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.ContainerExpiredSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeAddedSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeRemovedSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeResourceUpdateSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeUpdateSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.SchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.security.RMContainerTokenSecretManager;
@ -866,12 +869,6 @@ public class CapacityScheduler extends
FiCaSchedulerNode node = getNode(nm.getNodeID());
// Update resource if any change
if (SchedulerUtils.updateResourceIfChanged(node, nm, clusterResource,
LOG)) {
root.updateClusterResource(clusterResource);
}
List<UpdatedContainerInfo> containerInfoList = nm.pullContainerUpdates();
List<ContainerStatus> newlyLaunchedContainers = new ArrayList<ContainerStatus>();
List<ContainerStatus> completedContainers = new ArrayList<ContainerStatus>();
@ -900,6 +897,15 @@ public class CapacityScheduler extends
}
}
/**
* Process resource update on a node.
*/
private synchronized void updateNodeAndQueueResource(RMNode nm,
ResourceOption resourceOption) {
updateNodeResource(nm, resourceOption);
root.updateClusterResource(clusterResource);
}
private synchronized void allocateContainersToNode(FiCaSchedulerNode node) {
// Assign new containers...
@ -969,6 +975,14 @@ public class CapacityScheduler extends
removeNode(nodeRemovedEvent.getRemovedRMNode());
}
break;
case NODE_RESOURCE_UPDATE:
{
NodeResourceUpdateSchedulerEvent nodeResourceUpdatedEvent =
(NodeResourceUpdateSchedulerEvent)event;
updateNodeAndQueueResource(nodeResourceUpdatedEvent.getRMNode(),
nodeResourceUpdatedEvent.getResourceOption());
}
break;
case NODE_UPDATE:
{
NodeUpdateSchedulerEvent nodeUpdatedEvent = (NodeUpdateSchedulerEvent)event;

View File

@ -0,0 +1,43 @@
/**
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.hadoop.yarn.server.resourcemanager.scheduler.event;
import org.apache.hadoop.yarn.api.records.ResourceOption;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNode;
public class NodeResourceUpdateSchedulerEvent extends SchedulerEvent {
private final RMNode rmNode;
private final ResourceOption resourceOption;
public NodeResourceUpdateSchedulerEvent(RMNode rmNode,
ResourceOption resourceOption) {
super(SchedulerEventType.NODE_RESOURCE_UPDATE);
this.rmNode = rmNode;
this.resourceOption = resourceOption;
}
public RMNode getRMNode() {
return rmNode;
}
public ResourceOption getResourceOption() {
return resourceOption;
}
}

View File

@ -24,6 +24,7 @@ public enum SchedulerEventType {
NODE_ADDED,
NODE_REMOVED,
NODE_UPDATE,
NODE_RESOURCE_UPDATE,
// Source: RMApp
APP_ADDED,

View File

@ -45,6 +45,7 @@ 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.api.records.ResourceOption;
import org.apache.hadoop.yarn.api.records.ResourceRequest;
import org.apache.hadoop.yarn.conf.YarnConfiguration;
import org.apache.hadoop.yarn.exceptions.YarnException;
@ -79,6 +80,7 @@ import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.AppRemovedS
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.ContainerExpiredSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeAddedSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeRemovedSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeResourceUpdateSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeUpdateSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.SchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.security.RMContainerTokenSecretManager;
@ -968,9 +970,6 @@ public class FairScheduler extends
eventLog.log("HEARTBEAT", nm.getHostName());
FSSchedulerNode node = getFSSchedulerNode(nm.getNodeID());
// Update resource if any change
SchedulerUtils.updateResourceIfChanged(node, nm, clusterResource, LOG);
List<UpdatedContainerInfo> containerInfoList = nm.pullContainerUpdates();
List<ContainerStatus> newlyLaunchedContainers = new ArrayList<ContainerStatus>();
List<ContainerStatus> completedContainers = new ArrayList<ContainerStatus>();
@ -1173,6 +1172,15 @@ public class FairScheduler extends
removeApplication(appRemovedEvent.getApplicationID(),
appRemovedEvent.getFinalState());
break;
case NODE_RESOURCE_UPDATE:
if (!(event instanceof NodeResourceUpdateSchedulerEvent)) {
throw new RuntimeException("Unexpected event type: " + event);
}
NodeResourceUpdateSchedulerEvent nodeResourceUpdatedEvent =
(NodeResourceUpdateSchedulerEvent)event;
updateNodeResource(nodeResourceUpdatedEvent.getRMNode(),
nodeResourceUpdatedEvent.getResourceOption());
break;
case APP_ATTEMPT_ADDED:
if (!(event instanceof AppAttemptAddedSchedulerEvent)) {
throw new RuntimeException("Unexpected event type: " + event);
@ -1534,4 +1542,16 @@ public class FairScheduler extends
}
return queue1; // names are identical
}
/**
* Process resource update on a node and update Queue.
*/
@Override
public synchronized void updateNodeResource(RMNode nm,
ResourceOption resourceOption) {
super.updateNodeResource(nm, resourceOption);
updateRootQueueMetrics();
queueMgr.getRootQueue().setSteadyFairShare(clusterResource);
queueMgr.getRootQueue().recomputeSteadyShares();
}
}

View File

@ -85,6 +85,7 @@ import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.AppRemovedS
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.ContainerExpiredSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeAddedSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeRemovedSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeResourceUpdateSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeUpdateSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.SchedulerEvent;
import org.apache.hadoop.yarn.server.utils.BuilderUtils;
@ -681,9 +682,6 @@ public class FifoScheduler extends
private synchronized void nodeUpdate(RMNode rmNode) {
FiCaSchedulerNode node = getNode(rmNode.getNodeID());
// Update resource if any change
SchedulerUtils.updateResourceIfChanged(node, rmNode, clusterResource, LOG);
List<UpdatedContainerInfo> containerInfoList = rmNode.pullContainerUpdates();
List<ContainerStatus> newlyLaunchedContainers = new ArrayList<ContainerStatus>();
List<ContainerStatus> completedContainers = new ArrayList<ContainerStatus>();
@ -750,6 +748,14 @@ public class FifoScheduler extends
removeNode(nodeRemovedEvent.getRemovedRMNode());
}
break;
case NODE_RESOURCE_UPDATE:
{
NodeResourceUpdateSchedulerEvent nodeResourceUpdatedEvent =
(NodeResourceUpdateSchedulerEvent)event;
updateNodeResource(nodeResourceUpdatedEvent.getRMNode(),
nodeResourceUpdatedEvent.getResourceOption());
}
break;
case NODE_UPDATE:
{
NodeUpdateSchedulerEvent nodeUpdatedEvent =

View File

@ -27,7 +27,6 @@ import org.apache.hadoop.yarn.api.records.ContainerId;
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.api.records.ResourceOption;
import org.apache.hadoop.yarn.factories.RecordFactory;
import org.apache.hadoop.yarn.factory.providers.RecordFactoryProvider;
import org.apache.hadoop.yarn.server.api.protocolrecords.NodeHeartbeatResponse;
@ -94,14 +93,14 @@ public class MockNodes {
private String nodeAddr;
private String httpAddress;
private int cmdPort;
private ResourceOption perNode;
private Resource perNode;
private String rackName;
private String healthReport;
private long lastHealthReportTime;
private NodeState state;
public MockRMNodeImpl(NodeId nodeId, String nodeAddr, String httpAddress,
ResourceOption perNode, String rackName, String healthReport,
Resource perNode, String rackName, String healthReport,
long lastHealthReportTime, int cmdPort, String hostName, NodeState state) {
this.nodeId = nodeId;
this.nodeAddr = nodeAddr;
@ -147,7 +146,7 @@ public class MockNodes {
@Override
public Resource getTotalCapability() {
return this.perNode.getResource();
return this.perNode;
}
@Override
@ -204,16 +203,6 @@ public class MockNodes {
return lastHealthReportTime;
}
@Override
public void setResourceOption(ResourceOption resourceOption) {
this.perNode = resourceOption;
}
@Override
public ResourceOption getResourceOption(){
return this.perNode;
}
};
private static RMNode buildRMNode(int rack, final Resource perNode, NodeState state, String httpAddr) {
@ -232,8 +221,7 @@ public class MockNodes {
final String httpAddress = httpAddr;
String healthReport = (state == NodeState.UNHEALTHY) ? null : "HealthyMe";
return new MockRMNodeImpl(nodeID, nodeAddr, httpAddress,
ResourceOption.newInstance(perNode, RMNode.OVER_COMMIT_TIMEOUT_MILLIS_DEFAULT),
return new MockRMNodeImpl(nodeID, nodeAddr, httpAddress, perNode,
rackName, healthReport, 0, nid, hostName, state);
}

View File

@ -457,7 +457,6 @@ public class MockRM extends ResourceManager {
@Override
protected ResourceTrackerService createResourceTrackerService() {
Configuration conf = new Configuration();
RMContainerTokenSecretManager containerTokenSecretManager =
getRMContext().getContainerTokenSecretManager();
@ -548,6 +547,10 @@ public class MockRM extends ResourceManager {
return this.rmAppManager;
}
public AdminService getAdminService() {
return this.adminService;
}
@Override
protected void startWepApp() {
if (getConfig().getBoolean(ENABLE_WEBAPP, false)) {

View File

@ -23,7 +23,9 @@ import static org.junit.Assert.fail;
import java.util.ArrayList;
import java.util.Collections;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
@ -35,9 +37,13 @@ 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.ContainerStatus;
import org.apache.hadoop.yarn.api.records.NodeId;
import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.api.records.ResourceOption;
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.api.protocolrecords.UpdateNodeResourceRequest;
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;
@ -509,6 +515,85 @@ public class TestFifoScheduler {
rm.stop();
}
@Test
public void testResourceOverCommit() throws Exception {
MockRM rm = new MockRM(conf);
rm.start();
MockNM nm1 = rm.registerNode("127.0.0.1:1234", 4 * GB);
RMApp app1 = rm.submitApp(2048);
// kick the scheduling, 2 GB given to AM1, remaining 2GB on nm1
nm1.nodeHeartbeat(true);
RMAppAttempt attempt1 = app1.getCurrentAppAttempt();
MockAM am1 = rm.sendAMLaunched(attempt1.getAppAttemptId());
am1.registerAppAttempt();
SchedulerNodeReport report_nm1 = rm.getResourceScheduler().getNodeReport(
nm1.getNodeId());
// check node report, 2 GB used and 2 GB available
Assert.assertEquals(2 * GB, report_nm1.getUsedResource().getMemory());
Assert.assertEquals(2 * GB, report_nm1.getAvailableResource().getMemory());
// add request for containers
am1.addRequests(new String[] { "127.0.0.1", "127.0.0.2" }, 2 * GB, 1, 1);
AllocateResponse alloc1Response = am1.schedule(); // send the request
// kick the scheduler, 2 GB given to AM1, resource remaining 0
nm1.nodeHeartbeat(true);
while (alloc1Response.getAllocatedContainers().size() < 1) {
LOG.info("Waiting for containers to be created for app 1...");
Thread.sleep(1000);
alloc1Response = am1.schedule();
}
List<Container> allocated1 = alloc1Response.getAllocatedContainers();
Assert.assertEquals(1, allocated1.size());
Assert.assertEquals(2 * GB, allocated1.get(0).getResource().getMemory());
Assert.assertEquals(nm1.getNodeId(), allocated1.get(0).getNodeId());
report_nm1 = rm.getResourceScheduler().getNodeReport(nm1.getNodeId());
// check node report, 4 GB used and 0 GB available
Assert.assertEquals(0, report_nm1.getAvailableResource().getMemory());
Assert.assertEquals(4 * GB, report_nm1.getUsedResource().getMemory());
// check container is assigned with 2 GB.
Container c1 = allocated1.get(0);
Assert.assertEquals(2 * GB, c1.getResource().getMemory());
// update node resource to 2 GB, so resource is over-consumed.
Map<NodeId, ResourceOption> nodeResourceMap =
new HashMap<NodeId, ResourceOption>();
nodeResourceMap.put(nm1.getNodeId(),
ResourceOption.newInstance(Resource.newInstance(2 * GB, 1), -1));
UpdateNodeResourceRequest request =
UpdateNodeResourceRequest.newInstance(nodeResourceMap);
AdminService as = rm.adminService;
as.updateNodeResource(request);
// Now, the used resource is still 4 GB, and available resource is minus value.
report_nm1 = rm.getResourceScheduler().getNodeReport(nm1.getNodeId());
Assert.assertEquals(4 * GB, report_nm1.getUsedResource().getMemory());
Assert.assertEquals(-2 * GB, report_nm1.getAvailableResource().getMemory());
// Check container can complete successfully in case of resource over-commitment.
ContainerStatus containerStatus = BuilderUtils.newContainerStatus(
c1.getId(), ContainerState.COMPLETE, "", 0);
nm1.containerStatus(containerStatus);
int waitCount = 0;
while (attempt1.getJustFinishedContainers().size() < 1
&& waitCount++ != 20) {
LOG.info("Waiting for containers to be finished for app 1... Tried "
+ waitCount + " times already..");
Thread.sleep(100);
}
Assert.assertEquals(1, attempt1.getJustFinishedContainers().size());
Assert.assertEquals(1, am1.schedule().getCompletedContainersStatuses().size());
report_nm1 = rm.getResourceScheduler().getNodeReport(nm1.getNodeId());
Assert.assertEquals(2 * GB, report_nm1.getUsedResource().getMemory());
// As container return 2 GB back, the available resource becomes 0 again.
Assert.assertEquals(0 * GB, report_nm1.getAvailableResource().getMemory());
rm.stop();
}
public static void main(String[] args) throws Exception {
TestFifoScheduler t = new TestFifoScheduler();

View File

@ -17,6 +17,7 @@
*/
package org.apache.hadoop.yarn.server.resourcemanager;
import static org.junit.Assert.assertEquals;
import static org.mockito.Matchers.any;
import static org.mockito.Mockito.doAnswer;
import static org.mockito.Mockito.doReturn;
@ -48,6 +49,7 @@ 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.RMNodeReconnectEvent;
import org.apache.hadoop.yarn.server.resourcemanager.rmnode.RMNodeResourceUpdateEvent;
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.rmnode.UpdatedContainerInfo;
@ -463,8 +465,7 @@ public class TestRMNodeTransitions {
NodeId nodeId = BuilderUtils.newNodeId("localhost", 0);
Resource capability = Resource.newInstance(4096, 4);
RMNodeImpl node = new RMNodeImpl(nodeId, rmContext,null, 0, 0,
null, ResourceOption.newInstance(capability,
RMNode.OVER_COMMIT_TIMEOUT_MILLIS_DEFAULT), nmVersion);
null, capability, nmVersion);
node.handle(new RMNodeStartedEvent(node.getNodeID(), null, null));
Assert.assertEquals(NodeState.RUNNING, node.getState());
return node;
@ -487,6 +488,25 @@ public class TestRMNodeTransitions {
return node;
}
private RMNodeImpl getNewNode(Resource capability) {
NodeId nodeId = BuilderUtils.newNodeId("localhost", 0);
RMNodeImpl node = new RMNodeImpl(nodeId, rmContext, null, 0, 0, null,
capability, null);
return node;
}
private RMNodeImpl getRebootedNode() {
NodeId nodeId = BuilderUtils.newNodeId("localhost", 0);
Resource capability = Resource.newInstance(4096, 4);
RMNodeImpl node = new RMNodeImpl(nodeId, rmContext,null, 0, 0,
null, capability, null);
node.handle(new RMNodeStartedEvent(node.getNodeID(), null, null));
Assert.assertEquals(NodeState.RUNNING, node.getState());
node.handle(new RMNodeEvent(node.getNodeID(), RMNodeEventType.REBOOTING));
Assert.assertEquals(NodeState.REBOOTED, node.getState());
return node;
}
@Test
public void testAdd() {
RMNodeImpl node = getNewNode();
@ -535,6 +555,57 @@ public class TestRMNodeTransitions {
nodesListManagerEvent.getType());
}
@Test
public void testResourceUpdateOnRunningNode() {
RMNodeImpl node = getRunningNode();
Resource oldCapacity = node.getTotalCapability();
assertEquals("Memory resource is not match.", oldCapacity.getMemory(), 4096);
assertEquals("CPU resource is not match.", oldCapacity.getVirtualCores(), 4);
node.handle(new RMNodeResourceUpdateEvent(node.getNodeID(),
ResourceOption.newInstance(Resource.newInstance(2048, 2),
RMNode.OVER_COMMIT_TIMEOUT_MILLIS_DEFAULT)));
Resource newCapacity = node.getTotalCapability();
assertEquals("Memory resource is not match.", newCapacity.getMemory(), 2048);
assertEquals("CPU resource is not match.", newCapacity.getVirtualCores(), 2);
Assert.assertEquals(NodeState.RUNNING, node.getState());
Assert.assertNotNull(nodesListManagerEvent);
Assert.assertEquals(NodesListManagerEventType.NODE_USABLE,
nodesListManagerEvent.getType());
}
@Test
public void testResourceUpdateOnNewNode() {
RMNodeImpl node = getNewNode(Resource.newInstance(4096, 4));
Resource oldCapacity = node.getTotalCapability();
assertEquals("Memory resource is not match.", oldCapacity.getMemory(), 4096);
assertEquals("CPU resource is not match.", oldCapacity.getVirtualCores(), 4);
node.handle(new RMNodeResourceUpdateEvent(node.getNodeID(),
ResourceOption.newInstance(Resource.newInstance(2048, 2),
RMNode.OVER_COMMIT_TIMEOUT_MILLIS_DEFAULT)));
Resource newCapacity = node.getTotalCapability();
assertEquals("Memory resource is not match.", newCapacity.getMemory(), 2048);
assertEquals("CPU resource is not match.", newCapacity.getVirtualCores(), 2);
Assert.assertEquals(NodeState.NEW, node.getState());
}
@Test
public void testResourceUpdateOnRebootedNode() {
RMNodeImpl node = getRebootedNode();
Resource oldCapacity = node.getTotalCapability();
assertEquals("Memory resource is not match.", oldCapacity.getMemory(), 4096);
assertEquals("CPU resource is not match.", oldCapacity.getVirtualCores(), 4);
node.handle(new RMNodeResourceUpdateEvent(node.getNodeID(),
ResourceOption.newInstance(Resource.newInstance(2048, 2),
RMNode.OVER_COMMIT_TIMEOUT_MILLIS_DEFAULT)));
Resource newCapacity = node.getTotalCapability();
assertEquals("Memory resource is not match.", newCapacity.getMemory(), 2048);
assertEquals("CPU resource is not match.", newCapacity.getVirtualCores(), 2);
Assert.assertEquals(NodeState.REBOOTED, node.getState());
}
@Test
public void testReconnnectUpdate() {
final String nmVersion1 = "nm version 1";

View File

@ -18,6 +18,9 @@
package org.apache.hadoop.yarn.server.resourcemanager.resourcetracker;
import java.util.ArrayList;
import java.util.List;
import org.junit.Assert;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.yarn.api.records.NodeId;
@ -47,14 +50,14 @@ public class TestNMReconnect {
private static final RecordFactory recordFactory =
RecordFactoryProvider.getRecordFactory(null);
private RMNodeEvent rmNodeEvent = null;
private List<RMNodeEvent> rmNodeEvents = new ArrayList<RMNodeEvent>();
private class TestRMNodeEventDispatcher implements
EventHandler<RMNodeEvent> {
@Override
public void handle(RMNodeEvent event) {
rmNodeEvent = event;
rmNodeEvents.add(event);
}
}
@ -109,16 +112,18 @@ public class TestNMReconnect {
request1.setResource(capability);
resourceTrackerService.registerNodeManager(request1);
Assert.assertEquals(RMNodeEventType.STARTED, rmNodeEvent.getType());
Assert.assertEquals(RMNodeEventType.STARTED, rmNodeEvents.get(0).getType());
rmNodeEvent = null;
rmNodeEvents.clear();
resourceTrackerService.registerNodeManager(request1);
Assert.assertEquals(RMNodeEventType.RECONNECTED, rmNodeEvent.getType());
Assert.assertEquals(RMNodeEventType.RECONNECTED,
rmNodeEvents.get(0).getType());
rmNodeEvent = null;
rmNodeEvents.clear();
resourceTrackerService.registerNodeManager(request1);
capability = BuilderUtils.newResource(1024, 2);
request1.setResource(capability);
Assert.assertEquals(RMNodeEventType.RECONNECTED, rmNodeEvent.getType());
Assert.assertEquals(RMNodeEventType.RECONNECTED,
rmNodeEvents.get(0).getType());
}
}

View File

@ -47,23 +47,30 @@ import org.apache.hadoop.security.token.TokenIdentifier;
import org.apache.hadoop.yarn.LocalConfigurationProvider;
import org.apache.hadoop.yarn.api.ApplicationMasterProtocol;
import org.apache.hadoop.yarn.api.protocolrecords.AllocateRequest;
import org.apache.hadoop.yarn.api.protocolrecords.AllocateResponse;
import org.apache.hadoop.yarn.api.protocolrecords.RegisterApplicationMasterRequest;
import org.apache.hadoop.yarn.api.records.ApplicationAccessType;
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.ContainerId;
import org.apache.hadoop.yarn.api.records.ContainerState;
import org.apache.hadoop.yarn.api.records.ContainerStatus;
import org.apache.hadoop.yarn.api.records.NodeId;
import org.apache.hadoop.yarn.api.records.Priority;
import org.apache.hadoop.yarn.api.records.QueueInfo;
import org.apache.hadoop.yarn.api.records.QueueState;
import org.apache.hadoop.yarn.api.records.QueueUserACLInfo;
import org.apache.hadoop.yarn.api.records.Resource;
import org.apache.hadoop.yarn.api.records.ResourceOption;
import org.apache.hadoop.yarn.api.records.ResourceRequest;
import org.apache.hadoop.yarn.conf.YarnConfiguration;
import org.apache.hadoop.yarn.event.AsyncDispatcher;
import org.apache.hadoop.yarn.exceptions.YarnException;
import org.apache.hadoop.yarn.exceptions.YarnRuntimeException;
import org.apache.hadoop.yarn.ipc.YarnRPC;
import org.apache.hadoop.yarn.server.api.protocolrecords.UpdateNodeResourceRequest;
import org.apache.hadoop.yarn.server.resourcemanager.AdminService;
import org.apache.hadoop.yarn.server.resourcemanager.Application;
import org.apache.hadoop.yarn.server.resourcemanager.MockAM;
import org.apache.hadoop.yarn.server.resourcemanager.MockNM;
@ -90,6 +97,7 @@ import org.apache.hadoop.yarn.server.resourcemanager.scheduler.ResourceScheduler
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.SchedulerNode;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.SchedulerNodeReport;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.TestSchedulerUtils;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.common.fica.FiCaSchedulerApp;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.common.fica.FiCaSchedulerNode;
@ -591,7 +599,6 @@ public class TestCapacityScheduler {
return result;
}
@SuppressWarnings("resource")
@Test
public void testBlackListNodes() throws Exception {
Configuration conf = new Configuration();
@ -628,6 +635,104 @@ public class TestCapacityScheduler {
rm.stop();
}
@Test
public void testResourceOverCommit() throws Exception {
Configuration conf = new Configuration();
conf.setClass(YarnConfiguration.RM_SCHEDULER, CapacityScheduler.class,
ResourceScheduler.class);
MockRM rm = new MockRM(conf);
rm.start();
MockNM nm1 = rm.registerNode("127.0.0.1:1234", 4 * GB);
RMApp app1 = rm.submitApp(2048);
// kick the scheduling, 2 GB given to AM1, remaining 2GB on nm1
nm1.nodeHeartbeat(true);
RMAppAttempt attempt1 = app1.getCurrentAppAttempt();
MockAM am1 = rm.sendAMLaunched(attempt1.getAppAttemptId());
am1.registerAppAttempt();
SchedulerNodeReport report_nm1 = rm.getResourceScheduler().getNodeReport(
nm1.getNodeId());
// check node report, 2 GB used and 2 GB available
Assert.assertEquals(2 * GB, report_nm1.getUsedResource().getMemory());
Assert.assertEquals(2 * GB, report_nm1.getAvailableResource().getMemory());
// add request for containers
am1.addRequests(new String[] { "127.0.0.1", "127.0.0.2" }, 2 * GB, 1, 1);
AllocateResponse alloc1Response = am1.schedule(); // send the request
// kick the scheduler, 2 GB given to AM1, resource remaining 0
nm1.nodeHeartbeat(true);
while (alloc1Response.getAllocatedContainers().size() < 1) {
LOG.info("Waiting for containers to be created for app 1...");
Thread.sleep(100);
alloc1Response = am1.schedule();
}
List<Container> allocated1 = alloc1Response.getAllocatedContainers();
Assert.assertEquals(1, allocated1.size());
Assert.assertEquals(2 * GB, allocated1.get(0).getResource().getMemory());
Assert.assertEquals(nm1.getNodeId(), allocated1.get(0).getNodeId());
report_nm1 = rm.getResourceScheduler().getNodeReport(nm1.getNodeId());
// check node report, 4 GB used and 0 GB available
Assert.assertEquals(0, report_nm1.getAvailableResource().getMemory());
Assert.assertEquals(4 * GB, report_nm1.getUsedResource().getMemory());
// check container is assigned with 2 GB.
Container c1 = allocated1.get(0);
Assert.assertEquals(2 * GB, c1.getResource().getMemory());
// update node resource to 2 GB, so resource is over-consumed.
Map<NodeId, ResourceOption> nodeResourceMap =
new HashMap<NodeId, ResourceOption>();
nodeResourceMap.put(nm1.getNodeId(),
ResourceOption.newInstance(Resource.newInstance(2 * GB, 1), -1));
UpdateNodeResourceRequest request =
UpdateNodeResourceRequest.newInstance(nodeResourceMap);
AdminService as = ((MockRM)rm).getAdminService();
as.updateNodeResource(request);
// Now, the used resource is still 4 GB, and available resource is minus value.
report_nm1 = rm.getResourceScheduler().getNodeReport(nm1.getNodeId());
Assert.assertEquals(4 * GB, report_nm1.getUsedResource().getMemory());
Assert.assertEquals(-2 * GB, report_nm1.getAvailableResource().getMemory());
// Check container can complete successfully in case of resource over-commitment.
ContainerStatus containerStatus = BuilderUtils.newContainerStatus(
c1.getId(), ContainerState.COMPLETE, "", 0);
nm1.containerStatus(containerStatus);
int waitCount = 0;
while (attempt1.getJustFinishedContainers().size() < 1
&& waitCount++ != 20) {
LOG.info("Waiting for containers to be finished for app 1... Tried "
+ waitCount + " times already..");
Thread.sleep(100);
}
Assert.assertEquals(1, attempt1.getJustFinishedContainers().size());
Assert.assertEquals(1, am1.schedule().getCompletedContainersStatuses().size());
report_nm1 = rm.getResourceScheduler().getNodeReport(nm1.getNodeId());
Assert.assertEquals(2 * GB, report_nm1.getUsedResource().getMemory());
// As container return 2 GB back, the available resource becomes 0 again.
Assert.assertEquals(0 * GB, report_nm1.getAvailableResource().getMemory());
// Verify no NPE is trigger in schedule after resource is updated.
am1.addRequests(new String[] { "127.0.0.1", "127.0.0.2" }, 3 * GB, 1, 1);
alloc1Response = am1.schedule();
Assert.assertEquals("Shouldn't have enough resource to allocate containers",
0, alloc1Response.getAllocatedContainers().size());
int times = 0;
// try 10 times as scheduling is async process.
while (alloc1Response.getAllocatedContainers().size() < 1
&& times++ < 10) {
LOG.info("Waiting for containers to be allocated for app 1... Tried "
+ times + " times already..");
Thread.sleep(100);
}
Assert.assertEquals("Shouldn't have enough resource to allocate containers",
0, alloc1Response.getAllocatedContainers().size());
rm.stop();
}
@Test (timeout = 5000)
public void testApplicationComparator()
{

View File

@ -68,6 +68,7 @@ import org.apache.hadoop.yarn.server.resourcemanager.scheduler.common.fica.FiCaS
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.AppAddedSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.AppAttemptAddedSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeAddedSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeResourceUpdateSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.NodeUpdateSchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.scheduler.event.SchedulerEvent;
import org.apache.hadoop.yarn.server.resourcemanager.security.NMTokenSecretManagerInRM;
@ -278,17 +279,16 @@ public class TestFifoScheduler {
(Map<NodeId, FiCaSchedulerNode>) method.invoke(scheduler);
assertEquals(schedulerNodes.values().size(), 1);
// set resource of RMNode to 1024 and verify it works.
node0.setResourceOption(ResourceOption.newInstance(
Resources.createResource(1024, 4), RMNode.OVER_COMMIT_TIMEOUT_MILLIS_DEFAULT));
assertEquals(node0.getTotalCapability().getMemory(), 1024);
// verify that SchedulerNode's resource hasn't been changed.
assertEquals(schedulerNodes.get(node0.getNodeID()).
getAvailableResource().getMemory(), 2048);
// now, NM heartbeat comes.
NodeUpdateSchedulerEvent node0Update = new NodeUpdateSchedulerEvent(node0);
scheduler.handle(node0Update);
// SchedulerNode's available resource is changed.
Resource newResource = Resources.createResource(1024, 4);
NodeResourceUpdateSchedulerEvent node0ResourceUpdate = new
NodeResourceUpdateSchedulerEvent(node0, ResourceOption.newInstance(
newResource, RMNode.OVER_COMMIT_TIMEOUT_MILLIS_DEFAULT));
scheduler.handle(node0ResourceUpdate);
// SchedulerNode's total resource and available resource are changed.
assertEquals(schedulerNodes.get(node0.getNodeID()).getTotalResource()
.getMemory(), 1024);
assertEquals(schedulerNodes.get(node0.getNodeID()).
getAvailableResource().getMemory(), 1024);
QueueInfo queueInfo = scheduler.getQueueInfo(null, false, false);
@ -324,6 +324,7 @@ public class TestFifoScheduler {
// Before the node update event, there are one local request
Assert.assertEquals(1, nodeLocal.getNumContainers());
NodeUpdateSchedulerEvent node0Update = new NodeUpdateSchedulerEvent(node0);
// Now schedule.
scheduler.handle(node0Update);
@ -544,7 +545,6 @@ public class TestFifoScheduler {
LOG.info("--- END: testFifoScheduler ---");
}
@SuppressWarnings("resource")
@Test
public void testBlackListNodes() throws Exception {
Configuration conf = new Configuration();