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

YARN-5117. QueuingContainerManager does not start GUARANTEED Container even if Resources are available. (Konstantinos Karanasos via asuresh)

This commit is contained in:
Arun Suresh 2016-05-27 17:49:15 -07:00
parent 21890c4239
commit 4fc09a897b
3 changed files with 43 additions and 50 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/containermanager/monitor/ContainersMonitorImpl.java
View File

@ -769,12 +769,14 @@ public class ContainersMonitorImpl extends AbstractService implements
(int) (getVmemAllocatedForContainers() >> 20), 1.0f);
}
/**
* Calculates the vCores CPU usage that is assigned to the given
* {@link ProcessTreeInfo}. In particular, it takes into account the number of
* vCores that are allowed to be used by the NM and returns the CPU usage
* as a normalized value between {@literal >=} 0 and {@literal <=} 1.
*/
private float allocatedCpuUsage(ProcessTreeInfo pti) {
float cpuUsagePercentPerCore = pti.getCpuVcores() * 100.0f;
float cpuUsageTotalCoresPercentage = cpuUsagePercentPerCore
/ resourceCalculatorPlugin.getNumProcessors();
return (cpuUsageTotalCoresPercentage * 1000 *
maxVCoresAllottedForContainers / nodeCpuPercentageForYARN) / 1000.0f;
return (float) pti.getCpuVcores() / getVCoresAllocatedForContainers();
}
@Override

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

@ -127,17 +127,19 @@ public class QueuingContainerManagerImpl extends ContainerManagerImpl {
hasResourcesAvailable(allocatedContInfo.getPti())) {
startAllocatedContainer(allocatedContInfo);
} else {
this.context.getNMStateStore().storeContainer(containerTokenIdentifier
.getContainerID(), request);
this.context.getNMStateStore().storeContainerQueued(
containerTokenIdentifier.getContainerID());
ContainerId cIdToStart = containerTokenIdentifier.getContainerID();
this.context.getNMStateStore().storeContainer(cIdToStart, request);
this.context.getNMStateStore().storeContainerQueued(cIdToStart);
LOG.info("No available resources for container {} to start its execution "
+ "immediately.", cIdToStart);
if (allocatedContInfo.getExecutionType() == ExecutionType.GUARANTEED) {
queuedGuaranteedContainers.add(allocatedContInfo);
// Kill running opportunistic containers to make space for
// guaranteed container.
killOpportunisticContainers(allocatedContInfo);
} else {
LOG.info("Opportunistic container {} will be queued at the NM.",
cIdToStart);
queuedOpportunisticContainers.add(allocatedContInfo);
}
}
@ -401,7 +403,6 @@ public class QueuingContainerManagerImpl extends ContainerManagerImpl {
// Subtract the overall node resources.
getContainersMonitor().subtractNodeResourcesFromResourceUtilization(
resourceAllocationToFreeUp);
return resourceAllocationToFreeUp;
}

68
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/test/java/org/apache/hadoop/yarn/server/nodemanager/containermanager/queuing/TestQueuingContainerManager.java
View File

@ -30,7 +30,6 @@ import org.apache.hadoop.yarn.api.records.ApplicationAttemptId;
import org.apache.hadoop.yarn.api.records.ApplicationId;
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.ExecutionType;
import org.apache.hadoop.yarn.api.records.LocalResource;
@ -43,7 +42,6 @@ import org.apache.hadoop.yarn.security.NMTokenIdentifier;
import org.apache.hadoop.yarn.server.nodemanager.ContainerExecutor;
import org.apache.hadoop.yarn.server.nodemanager.Context;
import org.apache.hadoop.yarn.server.nodemanager.DeletionService;
import org.apache.hadoop.yarn.server.nodemanager.containermanager.BaseContainerManagerTest;
import org.apache.hadoop.yarn.server.nodemanager.containermanager.ContainerManagerImpl;
import org.apache.hadoop.yarn.server.nodemanager.containermanager.TestContainerManager;
import org.apache.hadoop.yarn.server.nodemanager.containermanager.container.Container;
@ -83,12 +81,11 @@ public class TestQueuingContainerManager extends TestContainerManager {
LOG = LogFactory.getLog(TestQueuingContainerManager.class);
}
HasResources hasResources = null;
boolean shouldDeleteWait = false;
@Override
protected ContainerManagerImpl
createContainerManager(DeletionService delSrvc) {
protected ContainerManagerImpl createContainerManager(
DeletionService delSrvc) {
return new QueuingContainerManagerImpl(context, exec, delSrvc,
nodeStatusUpdater, metrics, dirsHandler) {
@ -123,21 +120,35 @@ public class TestQueuingContainerManager extends TestContainerManager {
}
@Override
protected void authorizeGetAndStopContainerRequest(ContainerId containerId,
Container container, boolean stopRequest, NMTokenIdentifier identifier) throws YarnException {
if(container == null || container.getUser().equals("Fail")){
protected void authorizeGetAndStopContainerRequest(
ContainerId containerId, Container container, boolean stopRequest,
NMTokenIdentifier identifier) throws YarnException {
if (container == null || container.getUser().equals("Fail")) {
throw new YarnException("Reject this container");
}
}
@Override
protected ContainersMonitor createContainersMonitor(ContainerExecutor
exec) {
protected ContainersMonitor createContainersMonitor(
ContainerExecutor exec) {
return new ContainersMonitorImpl(exec, dispatcher, this.context) {
// Define resources available for containers to be executed.
@Override
public boolean hasResourcesAvailable(
ContainersMonitorImpl.ProcessTreeInfo pti) {
return hasResources.decide(this.context, pti.getContainerId());
public long getPmemAllocatedForContainers() {
return 2048 * 1024 * 1024L;
}
@Override
public long getVmemAllocatedForContainers() {
float pmemRatio = getConfig().getFloat(
YarnConfiguration.NM_VMEM_PMEM_RATIO,
YarnConfiguration.DEFAULT_NM_VMEM_PMEM_RATIO);
return (long) (pmemRatio * getPmemAllocatedForContainers());
}
@Override
public long getVCoresAllocatedForContainers() {
return 2;
}
};
}
@ -172,12 +183,6 @@ public class TestQueuingContainerManager extends TestContainerManager {
public void setup() throws IOException {
super.setup();
shouldDeleteWait = false;
hasResources = new HasResources() {
@Override
public boolean decide(Context context, ContainerId cId) {
return true;
}
};
}
/**
@ -256,31 +261,16 @@ public class TestQueuingContainerManager extends TestContainerManager {
containerLaunchContext,
createContainerToken(createContainerId(3), DUMMY_RM_IDENTIFIER,
context.getNodeId(),
user, context.getContainerTokenSecretManager())));
user, BuilderUtils.newResource(1024, 1),
context.getContainerTokenSecretManager(), null,
ExecutionType.GUARANTEED)));
StartContainersRequest allRequests =
StartContainersRequest.newInstance(list);
// Plugin to simulate that the Node is full
// It only allows 1 container to run at a time.
hasResources = new HasResources() {
@Override
public boolean decide(Context context, ContainerId cId) {
int nOpp = ((QueuingContainerManagerImpl) containerManager)
.getNumAllocatedOpportunisticContainers();
int nGuar = ((QueuingContainerManagerImpl) containerManager)
.getNumAllocatedGuaranteedContainers();
boolean val = (nOpp + nGuar < 1);
System.out.println("\nHasResources : [" + cId + "]," +
"Opp[" + nOpp + "], Guar[" + nGuar + "], [" + val + "]\n");
return val;
}
};
containerManager.startContainers(allRequests);
BaseContainerManagerTest.waitForContainerState(containerManager,
createContainerId(3),
ContainerState.COMPLETE, 40);
Thread.sleep(10000);
List<ContainerId> statList = new ArrayList<ContainerId>();
for (int i = 0; i < 4; i++) {
statList.add(createContainerId(i));