MAPREDUCE-3747. Initialize queue metrics upfront and added start/finish time to RM Web-UI.
git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/trunk@1240886 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
parent
f88574acde
commit
c7007a8a35
@ -705,6 +705,9 @@ Release 0.23.1 - Unreleased
|
||||
MAPREDUCE-3765. FifoScheduler does not respect yarn.scheduler.fifo.minimum-
|
||||
allocation-mb setting (Hitesh Shah via mahadev)
|
||||
|
||||
MAPREDUCE-3747. Initialize queue metrics upfront and added start/finish
|
||||
time to RM Web-UI. (acmurthy)
|
||||
|
||||
Release 0.23.0 - 2011-11-01
|
||||
|
||||
INCOMPATIBLE CHANGES
|
||||
|
@ -98,6 +98,12 @@ public interface CSQueue
|
||||
*/
|
||||
public float getUsedCapacity();
|
||||
|
||||
/**
|
||||
* Set used capacity of the queue.
|
||||
* @param usedCapacity used capacity of the queue
|
||||
*/
|
||||
public void setUsedCapacity(float usedCapacity);
|
||||
|
||||
/**
|
||||
* Get the currently utilized resources in the cluster
|
||||
* by the queue and children (if any).
|
||||
@ -114,6 +120,12 @@ public interface CSQueue
|
||||
*/
|
||||
public float getUtilization();
|
||||
|
||||
/**
|
||||
* Get the current <em>utilization</em> of the queue.
|
||||
* @param utilization queue utilization
|
||||
*/
|
||||
public void setUtilization(float utilization);
|
||||
|
||||
/**
|
||||
* Get the current run-state of the queue
|
||||
* @return current run-state
|
||||
|
@ -17,7 +17,9 @@
|
||||
*/
|
||||
package org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity;
|
||||
|
||||
import org.apache.hadoop.yarn.Lock;
|
||||
import org.apache.hadoop.yarn.api.records.Resource;
|
||||
import org.apache.hadoop.yarn.server.resourcemanager.resource.Resources;
|
||||
|
||||
class CSQueueUtils {
|
||||
|
||||
@ -65,4 +67,40 @@ public static int computeMaxActiveApplicationsPerUser(
|
||||
1);
|
||||
}
|
||||
|
||||
@Lock(CSQueue.class)
|
||||
public static void updateQueueStatistics(
|
||||
final CSQueue childQueue, final CSQueue parentQueue,
|
||||
final Resource clusterResource, final Resource minimumAllocation) {
|
||||
final int clusterMemory = clusterResource.getMemory();
|
||||
final int usedMemory = childQueue.getUsedResources().getMemory();
|
||||
|
||||
float queueLimit = 0.0f;
|
||||
float utilization = 0.0f;
|
||||
float usedCapacity = 0.0f;
|
||||
if (clusterMemory > 0) {
|
||||
queueLimit = clusterMemory * childQueue.getAbsoluteCapacity();
|
||||
final float parentAbsoluteCapacity =
|
||||
(parentQueue == null) ? 1.0f : parentQueue.getAbsoluteCapacity();
|
||||
utilization = (usedMemory / queueLimit);
|
||||
usedCapacity = (usedMemory / (clusterMemory * parentAbsoluteCapacity));
|
||||
}
|
||||
|
||||
childQueue.setUtilization(utilization);
|
||||
childQueue.setUsedCapacity(usedCapacity);
|
||||
|
||||
int available =
|
||||
Math.max((roundUp(minimumAllocation, (int)queueLimit) - usedMemory), 0);
|
||||
childQueue.getMetrics().setAvailableResourcesToQueue(
|
||||
Resources.createResource(available));
|
||||
}
|
||||
|
||||
public static int roundUp(Resource minimumAllocation, int memory) {
|
||||
int minMemory = minimumAllocation.getMemory();
|
||||
return LeafQueue.divideAndCeil(memory, minMemory) * minMemory;
|
||||
}
|
||||
|
||||
public static int roundDown(Resource minimumAllocation, int memory) {
|
||||
int minMemory = minimumAllocation.getMemory();
|
||||
return (memory / minMemory) * minMemory;
|
||||
}
|
||||
}
|
||||
|
@ -180,7 +180,9 @@ public LeafQueue(CapacitySchedulerContext cs,
|
||||
Map<QueueACL, AccessControlList> acls =
|
||||
cs.getConfiguration().getAcls(getQueuePath());
|
||||
|
||||
setupQueueConfigs(capacity, absoluteCapacity,
|
||||
setupQueueConfigs(
|
||||
cs.getClusterResources(),
|
||||
capacity, absoluteCapacity,
|
||||
maximumCapacity, absoluteMaxCapacity,
|
||||
userLimit, userLimitFactor,
|
||||
maxApplications, maxApplicationsPerUser,
|
||||
@ -198,6 +200,7 @@ public LeafQueue(CapacitySchedulerContext cs,
|
||||
}
|
||||
|
||||
private synchronized void setupQueueConfigs(
|
||||
Resource clusterResource,
|
||||
float capacity, float absoluteCapacity,
|
||||
float maximumCapacity, float absoluteMaxCapacity,
|
||||
int userLimit, float userLimitFactor,
|
||||
@ -235,6 +238,10 @@ private synchronized void setupQueueConfigs(
|
||||
for (Map.Entry<QueueACL, AccessControlList> e : acls.entrySet()) {
|
||||
aclsString.append(e.getKey() + ":" + e.getValue().getAclString());
|
||||
}
|
||||
|
||||
// Update metrics
|
||||
CSQueueUtils.updateQueueStatistics(
|
||||
this, parent, clusterResource, minimumAllocation);
|
||||
|
||||
LOG.info("Initializing " + queueName + "\n" +
|
||||
"capacity = " + capacity +
|
||||
@ -386,11 +393,11 @@ public List<CSQueue> getChildQueues() {
|
||||
return null;
|
||||
}
|
||||
|
||||
synchronized void setUtilization(float utilization) {
|
||||
public synchronized void setUtilization(float utilization) {
|
||||
this.utilization = utilization;
|
||||
}
|
||||
|
||||
synchronized void setUsedCapacity(float usedCapacity) {
|
||||
public synchronized void setUsedCapacity(float usedCapacity) {
|
||||
this.usedCapacity = usedCapacity;
|
||||
}
|
||||
|
||||
@ -534,7 +541,9 @@ public synchronized void reinitialize(CSQueue queue, Resource clusterResource)
|
||||
}
|
||||
|
||||
LeafQueue leafQueue = (LeafQueue)queue;
|
||||
setupQueueConfigs(leafQueue.capacity, leafQueue.absoluteCapacity,
|
||||
setupQueueConfigs(
|
||||
clusterResource,
|
||||
leafQueue.capacity, leafQueue.absoluteCapacity,
|
||||
leafQueue.maximumCapacity, leafQueue.absoluteMaxCapacity,
|
||||
leafQueue.userLimit, leafQueue.userLimitFactor,
|
||||
leafQueue.maxApplications,
|
||||
@ -542,8 +551,6 @@ public synchronized void reinitialize(CSQueue queue, Resource clusterResource)
|
||||
leafQueue.getMaximumActiveApplications(),
|
||||
leafQueue.getMaximumActiveApplicationsPerUser(),
|
||||
leafQueue.state, leafQueue.acls);
|
||||
|
||||
updateResource(clusterResource);
|
||||
}
|
||||
|
||||
@Override
|
||||
@ -883,7 +890,8 @@ private Resource computeUserLimitAndSetHeadroom(
|
||||
|
||||
Resource queueMaxCap = // Queue Max-Capacity
|
||||
Resources.createResource(
|
||||
roundDown((int)(absoluteMaxCapacity * clusterResource.getMemory()))
|
||||
CSQueueUtils.roundDown(minimumAllocation,
|
||||
(int)(absoluteMaxCapacity * clusterResource.getMemory()))
|
||||
);
|
||||
|
||||
Resource userConsumed = getUser(user).getConsumedResources();
|
||||
@ -904,16 +912,6 @@ private Resource computeUserLimitAndSetHeadroom(
|
||||
return userLimit;
|
||||
}
|
||||
|
||||
private int roundUp(int memory) {
|
||||
int minMemory = minimumAllocation.getMemory();
|
||||
return divideAndCeil(memory, minMemory) * minMemory;
|
||||
}
|
||||
|
||||
private int roundDown(int memory) {
|
||||
int minMemory = minimumAllocation.getMemory();
|
||||
return (memory / minMemory) * minMemory;
|
||||
}
|
||||
|
||||
@Lock(NoLock.class)
|
||||
private Resource computeUserLimit(SchedulerApp application,
|
||||
Resource clusterResource, Resource required) {
|
||||
@ -927,8 +925,11 @@ private Resource computeUserLimit(SchedulerApp application,
|
||||
// Allow progress for queues with miniscule capacity
|
||||
final int queueCapacity =
|
||||
Math.max(
|
||||
roundUp((int)(absoluteCapacity * clusterResource.getMemory())),
|
||||
required.getMemory());
|
||||
CSQueueUtils.roundUp(
|
||||
minimumAllocation,
|
||||
(int)(absoluteCapacity * clusterResource.getMemory())),
|
||||
required.getMemory()
|
||||
);
|
||||
|
||||
final int consumed = usedResources.getMemory();
|
||||
final int currentCapacity =
|
||||
@ -943,7 +944,8 @@ private Resource computeUserLimit(SchedulerApp application,
|
||||
final int activeUsers = activeUsersManager.getNumActiveUsers();
|
||||
|
||||
int limit =
|
||||
roundUp(
|
||||
CSQueueUtils.roundUp(
|
||||
minimumAllocation,
|
||||
Math.min(
|
||||
Math.max(divideAndCeil(currentCapacity, activeUsers),
|
||||
divideAndCeil((int)userLimit*currentCapacity, 100)),
|
||||
@ -991,7 +993,7 @@ private synchronized boolean assignToUser(String userName, Resource limit) {
|
||||
return true;
|
||||
}
|
||||
|
||||
private static int divideAndCeil(int a, int b) {
|
||||
static int divideAndCeil(int a, int b) {
|
||||
if (b == 0) {
|
||||
LOG.info("divideAndCeil called with a=" + a + " b=" + b);
|
||||
return 0;
|
||||
@ -1325,7 +1327,8 @@ synchronized void allocateResource(Resource clusterResource,
|
||||
SchedulerApp application, Resource resource) {
|
||||
// Update queue metrics
|
||||
Resources.addTo(usedResources, resource);
|
||||
updateResource(clusterResource);
|
||||
CSQueueUtils.updateQueueStatistics(
|
||||
this, parent, clusterResource, minimumAllocation);
|
||||
++numContainers;
|
||||
|
||||
// Update user metrics
|
||||
@ -1349,7 +1352,8 @@ synchronized void releaseResource(Resource clusterResource,
|
||||
SchedulerApp application, Resource resource) {
|
||||
// Update queue metrics
|
||||
Resources.subtractFrom(usedResources, resource);
|
||||
updateResource(clusterResource);
|
||||
CSQueueUtils.updateQueueStatistics(
|
||||
this, parent, clusterResource, minimumAllocation);
|
||||
--numContainers;
|
||||
|
||||
// Update user metrics
|
||||
@ -1374,6 +1378,10 @@ public synchronized void updateClusterResource(Resource clusterResource) {
|
||||
CSQueueUtils.computeMaxActiveApplicationsPerUser(
|
||||
maxActiveApplications, userLimit, userLimitFactor);
|
||||
|
||||
// Update metrics
|
||||
CSQueueUtils.updateQueueStatistics(
|
||||
this, parent, clusterResource, minimumAllocation);
|
||||
|
||||
// Update application properties
|
||||
for (SchedulerApp application : activeApplications) {
|
||||
synchronized (application) {
|
||||
@ -1383,18 +1391,6 @@ public synchronized void updateClusterResource(Resource clusterResource) {
|
||||
}
|
||||
}
|
||||
|
||||
private synchronized void updateResource(Resource clusterResource) {
|
||||
float queueLimit = clusterResource.getMemory() * absoluteCapacity;
|
||||
setUtilization(usedResources.getMemory() / queueLimit);
|
||||
setUsedCapacity(usedResources.getMemory()
|
||||
/ (clusterResource.getMemory() * parent.getAbsoluteCapacity()));
|
||||
|
||||
Resource resourceLimit =
|
||||
Resources.createResource(roundUp((int)queueLimit));
|
||||
metrics.setAvailableResourcesToQueue(
|
||||
Resources.subtractFrom(resourceLimit, usedResources));
|
||||
}
|
||||
|
||||
@Override
|
||||
public QueueMetrics getMetrics() {
|
||||
return metrics;
|
||||
|
@ -97,7 +97,8 @@ public class ParentQueue implements CSQueue {
|
||||
RecordFactoryProvider.getRecordFactory(null);
|
||||
|
||||
public ParentQueue(CapacitySchedulerContext cs,
|
||||
String queueName, Comparator<CSQueue> comparator, CSQueue parent, CSQueue old) {
|
||||
String queueName, Comparator<CSQueue> comparator,
|
||||
CSQueue parent, CSQueue old) {
|
||||
minimumAllocation = cs.getMinimumResourceCapability();
|
||||
|
||||
this.parent = parent;
|
||||
@ -137,7 +138,8 @@ public ParentQueue(CapacitySchedulerContext cs,
|
||||
this.queueInfo.setQueueName(queueName);
|
||||
this.queueInfo.setChildQueues(new ArrayList<QueueInfo>());
|
||||
|
||||
setupQueueConfigs(capacity, absoluteCapacity,
|
||||
setupQueueConfigs(cs.getClusterResources(),
|
||||
capacity, absoluteCapacity,
|
||||
maximumCapacity, absoluteMaxCapacity, state, acls);
|
||||
|
||||
this.queueComparator = comparator;
|
||||
@ -149,9 +151,10 @@ public ParentQueue(CapacitySchedulerContext cs,
|
||||
}
|
||||
|
||||
private synchronized void setupQueueConfigs(
|
||||
float capacity, float absoluteCapacity,
|
||||
float maximumCapacity, float absoluteMaxCapacity,
|
||||
QueueState state, Map<QueueACL, AccessControlList> acls
|
||||
Resource clusterResource,
|
||||
float capacity, float absoluteCapacity,
|
||||
float maximumCapacity, float absoluteMaxCapacity,
|
||||
QueueState state, Map<QueueACL, AccessControlList> acls
|
||||
) {
|
||||
// Sanity check
|
||||
CSQueueUtils.checkMaxCapacity(getQueueName(), capacity, maximumCapacity);
|
||||
@ -174,6 +177,10 @@ private synchronized void setupQueueConfigs(
|
||||
aclsString.append(e.getKey() + ":" + e.getValue().getAclString());
|
||||
}
|
||||
|
||||
// Update metrics
|
||||
CSQueueUtils.updateQueueStatistics(
|
||||
this, parent, clusterResource, minimumAllocation);
|
||||
|
||||
LOG.info(queueName +
|
||||
", capacity=" + capacity +
|
||||
", asboluteCapacity=" + absoluteCapacity +
|
||||
@ -384,12 +391,10 @@ public synchronized void reinitialize(CSQueue queue, Resource clusterResource)
|
||||
childQueues.addAll(currentChildQueues.values());
|
||||
|
||||
// Set new configs
|
||||
setupQueueConfigs(parentQueue.capacity, parentQueue.absoluteCapacity,
|
||||
setupQueueConfigs(clusterResource,
|
||||
parentQueue.capacity, parentQueue.absoluteCapacity,
|
||||
parentQueue.maximumCapacity, parentQueue.absoluteMaxCapacity,
|
||||
parentQueue.state, parentQueue.acls);
|
||||
|
||||
// Update
|
||||
updateResource(clusterResource);
|
||||
}
|
||||
|
||||
Map<String, CSQueue> getQueues(Set<CSQueue> queues) {
|
||||
@ -485,11 +490,11 @@ public synchronized void removeApplication(SchedulerApp application,
|
||||
" #applications: " + getNumApplications());
|
||||
}
|
||||
|
||||
synchronized void setUsedCapacity(float usedCapacity) {
|
||||
public synchronized void setUsedCapacity(float usedCapacity) {
|
||||
this.usedCapacity = usedCapacity;
|
||||
}
|
||||
|
||||
synchronized void setUtilization(float utilization) {
|
||||
public synchronized void setUtilization(float utilization) {
|
||||
this.utilization = utilization;
|
||||
}
|
||||
|
||||
@ -674,14 +679,16 @@ public void completedContainer(Resource clusterResource,
|
||||
synchronized void allocateResource(Resource clusterResource,
|
||||
Resource resource) {
|
||||
Resources.addTo(usedResources, resource);
|
||||
updateResource(clusterResource);
|
||||
CSQueueUtils.updateQueueStatistics(
|
||||
this, parent, clusterResource, minimumAllocation);
|
||||
++numContainers;
|
||||
}
|
||||
|
||||
synchronized void releaseResource(Resource clusterResource,
|
||||
Resource resource) {
|
||||
Resources.subtractFrom(usedResources, resource);
|
||||
updateResource(clusterResource);
|
||||
CSQueueUtils.updateQueueStatistics(
|
||||
this, parent, clusterResource, minimumAllocation);
|
||||
--numContainers;
|
||||
}
|
||||
|
||||
@ -691,22 +698,12 @@ public synchronized void updateClusterResource(Resource clusterResource) {
|
||||
for (CSQueue childQueue : childQueues) {
|
||||
childQueue.updateClusterResource(clusterResource);
|
||||
}
|
||||
|
||||
// Update metrics
|
||||
CSQueueUtils.updateQueueStatistics(
|
||||
this, parent, clusterResource, minimumAllocation);
|
||||
}
|
||||
|
||||
private synchronized void updateResource(Resource clusterResource) {
|
||||
float queueLimit = clusterResource.getMemory() * absoluteCapacity;
|
||||
float parentAbsoluteCapacity =
|
||||
(rootQueue) ? 1.0f : parent.getAbsoluteCapacity();
|
||||
setUtilization(usedResources.getMemory() / queueLimit);
|
||||
setUsedCapacity(usedResources.getMemory()
|
||||
/ (clusterResource.getMemory() * parentAbsoluteCapacity));
|
||||
|
||||
Resource resourceLimit =
|
||||
Resources.createResource((int)queueLimit);
|
||||
metrics.setAvailableResourcesToQueue(
|
||||
Resources.subtractFrom(resourceLimit, usedResources));
|
||||
}
|
||||
|
||||
@Override
|
||||
public QueueMetrics getMetrics() {
|
||||
return metrics;
|
||||
|
@ -19,7 +19,6 @@
|
||||
package org.apache.hadoop.yarn.server.resourcemanager.webapp;
|
||||
|
||||
import static org.apache.hadoop.yarn.util.StringHelper.join;
|
||||
import static org.apache.hadoop.yarn.util.StringHelper.sjoin;
|
||||
import static org.apache.hadoop.yarn.webapp.YarnWebParams.APP_STATE;
|
||||
import static org.apache.hadoop.yarn.webapp.view.JQueryUI._PROGRESSBAR;
|
||||
import static org.apache.hadoop.yarn.webapp.view.JQueryUI._PROGRESSBAR_VALUE;
|
||||
@ -27,6 +26,7 @@
|
||||
import org.apache.hadoop.yarn.server.resourcemanager.rmapp.RMApp;
|
||||
import org.apache.hadoop.yarn.server.resourcemanager.rmapp.RMAppState;
|
||||
import org.apache.hadoop.yarn.server.resourcemanager.webapp.dao.AppInfo;
|
||||
import org.apache.hadoop.yarn.util.Times;
|
||||
import org.apache.hadoop.yarn.webapp.hamlet.Hamlet;
|
||||
import org.apache.hadoop.yarn.webapp.hamlet.Hamlet.TABLE;
|
||||
import org.apache.hadoop.yarn.webapp.hamlet.Hamlet.TBODY;
|
||||
@ -52,11 +52,12 @@ class AppsBlock extends HtmlBlock {
|
||||
th(".user", "User").
|
||||
th(".name", "Name").
|
||||
th(".queue", "Queue").
|
||||
th(".starttime", "StartTime").
|
||||
th(".finishtime", "FinishTime").
|
||||
th(".state", "State").
|
||||
th(".finalstatus", "FinalStatus").
|
||||
th(".progress", "Progress").
|
||||
th(".ui", "Tracking UI").
|
||||
th(".note", "Note")._()._().
|
||||
th(".ui", "Tracking UI")._()._().
|
||||
tbody();
|
||||
int i = 0;
|
||||
String reqState = $(APP_STATE);
|
||||
@ -67,6 +68,8 @@ class AppsBlock extends HtmlBlock {
|
||||
}
|
||||
AppInfo appInfo = new AppInfo(app, true);
|
||||
String percent = String.format("%.1f", appInfo.getProgress());
|
||||
String startTime = Times.format(appInfo.getStartTime());
|
||||
String finishTime = Times.format(appInfo.getFinishTime());
|
||||
tbody.
|
||||
tr().
|
||||
td().
|
||||
@ -75,6 +78,10 @@ class AppsBlock extends HtmlBlock {
|
||||
td(appInfo.getUser()).
|
||||
td(appInfo.getName()).
|
||||
td(appInfo.getQueue()).
|
||||
td().
|
||||
br().$title(startTime)._()._(startTime)._().
|
||||
td().
|
||||
br().$title(startTime)._()._(finishTime)._().
|
||||
td(appInfo.getState()).
|
||||
td(appInfo.getFinalStatus()).
|
||||
td().
|
||||
@ -85,8 +92,7 @@ class AppsBlock extends HtmlBlock {
|
||||
$style(join("width:", percent, '%'))._()._()._().
|
||||
td().
|
||||
a(!appInfo.isTrackingUrlReady()?
|
||||
"#" : appInfo.getTrackingUrlPretty(), appInfo.getTrackingUI())._().
|
||||
td(appInfo.getNote())._();
|
||||
"#" : appInfo.getTrackingUrlPretty(), appInfo.getTrackingUI())._()._();
|
||||
if (list.rendering != Render.HTML && ++i >= 20) break;
|
||||
}
|
||||
tbody._()._();
|
||||
|
@ -55,15 +55,19 @@ protected void render(Block html) {
|
||||
//CSS in the correct spot
|
||||
html.style(".metrics {margin-bottom:5px}");
|
||||
|
||||
ClusterMetricsInfo clusterMetrics = new ClusterMetricsInfo(this.rm, this.rmContext);
|
||||
|
||||
ClusterMetricsInfo clusterMetrics =
|
||||
new ClusterMetricsInfo(this.rm, this.rmContext);
|
||||
|
||||
DIV<Hamlet> div = html.div().$class("metrics");
|
||||
|
||||
div.table("#metricsoverview").
|
||||
div.h3("Cluster Metrics").
|
||||
table("#metricsoverview").
|
||||
thead().$class("ui-widget-header").
|
||||
tr().
|
||||
th().$class("ui-state-default")._("Apps Submitted")._().
|
||||
th().$class("ui-state-default")._("Apps Pending")._().
|
||||
th().$class("ui-state-default")._("Apps Running")._().
|
||||
th().$class("ui-state-default")._("Apps Completed")._().
|
||||
th().$class("ui-state-default")._("Containers Running")._().
|
||||
th().$class("ui-state-default")._("Memory Used")._().
|
||||
th().$class("ui-state-default")._("Memory Total")._().
|
||||
@ -78,6 +82,14 @@ protected void render(Block html) {
|
||||
tbody().$class("ui-widget-content").
|
||||
tr().
|
||||
td(String.valueOf(clusterMetrics.getAppsSubmitted())).
|
||||
td(String.valueOf(clusterMetrics.getAppsPending())).
|
||||
td(String.valueOf(clusterMetrics.getAppsRunning())).
|
||||
td(
|
||||
String.valueOf(
|
||||
clusterMetrics.getAppsCompleted() +
|
||||
clusterMetrics.getAppsFailed() + clusterMetrics.getAppsKilled()
|
||||
)
|
||||
).
|
||||
td(String.valueOf(clusterMetrics.getContainersAllocated())).
|
||||
td(StringUtils.byteDesc(clusterMetrics.getAllocatedMB() * BYTES_IN_MB)).
|
||||
td(StringUtils.byteDesc(clusterMetrics.getTotalMB() * BYTES_IN_MB)).
|
||||
@ -89,26 +101,38 @@ protected void render(Block html) {
|
||||
td().a(url("nodes/rebooted"),String.valueOf(clusterMetrics.getRebootedNodes()))._().
|
||||
_().
|
||||
_()._();
|
||||
|
||||
|
||||
String user = request().getRemoteUser();
|
||||
if (user != null) {
|
||||
UserMetricsInfo userMetrics = new UserMetricsInfo(this.rm, this.rmContext, user);
|
||||
if (userMetrics.metricsAvailable()) {
|
||||
div.table("#usermetricsoverview").
|
||||
div.h3("User Metrics for " + user).
|
||||
table("#usermetricsoverview").
|
||||
thead().$class("ui-widget-header").
|
||||
tr().
|
||||
th().$class("ui-state-default")._("Apps Submitted ("+user+")")._().
|
||||
th().$class("ui-state-default")._("Containers Running ("+user+")")._().
|
||||
th().$class("ui-state-default")._("Containers Pending ("+user+")")._().
|
||||
th().$class("ui-state-default")._("Containers Reserved ("+user+")")._().
|
||||
th().$class("ui-state-default")._("Memory Used ("+user+")")._().
|
||||
th().$class("ui-state-default")._("Memory Pending ("+user+")")._().
|
||||
th().$class("ui-state-default")._("Memory Reserved ("+user+")")._().
|
||||
th().$class("ui-state-default")._("Apps Submitted")._().
|
||||
th().$class("ui-state-default")._("Apps Pending")._().
|
||||
th().$class("ui-state-default")._("Apps Running")._().
|
||||
th().$class("ui-state-default")._("Apps Completed")._().
|
||||
th().$class("ui-state-default")._("Containers Running")._().
|
||||
th().$class("ui-state-default")._("Containers Pending")._().
|
||||
th().$class("ui-state-default")._("Containers Reserved")._().
|
||||
th().$class("ui-state-default")._("Memory Used")._().
|
||||
th().$class("ui-state-default")._("Memory Pending")._().
|
||||
th().$class("ui-state-default")._("Memory Reserved")._().
|
||||
_().
|
||||
_().
|
||||
tbody().$class("ui-widget-content").
|
||||
tr().
|
||||
td(String.valueOf(userMetrics.getAppsSubmitted())).
|
||||
td(String.valueOf(userMetrics.getAppsPending())).
|
||||
td(String.valueOf(userMetrics.getAppsRunning())).
|
||||
td(
|
||||
String.valueOf(
|
||||
(userMetrics.getAppsCompleted() +
|
||||
userMetrics.getAppsFailed() + userMetrics.getAppsKilled())
|
||||
)
|
||||
).
|
||||
td(String.valueOf(userMetrics.getRunningContainers())).
|
||||
td(String.valueOf(userMetrics.getPendingContainers())).
|
||||
td(String.valueOf(userMetrics.getReservedContainers())).
|
||||
@ -117,6 +141,7 @@ protected void render(Block html) {
|
||||
td(StringUtils.byteDesc(userMetrics.getReservedMB() * BYTES_IN_MB)).
|
||||
_().
|
||||
_()._();
|
||||
|
||||
}
|
||||
}
|
||||
|
||||
|
@ -63,10 +63,11 @@ protected Class<? extends SubView> content() {
|
||||
|
||||
private String appsTableInit() {
|
||||
AppsList list = getInstance(AppsList.class);
|
||||
// id, user, name, queue, state, progress, ui, note
|
||||
// id, user, name, queue, starttime, finishtime, state, progress, ui
|
||||
StringBuilder init = tableInit().
|
||||
append(", aoColumns:[{sType:'title-numeric'}, null, null, null, null,").
|
||||
append("null,{sType:'title-numeric', bSearchable:false}, null, null]");
|
||||
append(", aoColumns:[{sType:'title-numeric'}, null, null, null, ").
|
||||
append("null, null , null, ").
|
||||
append("null,{sType:'title-numeric', bSearchable:false}, null]");
|
||||
|
||||
// Sort by id upon page load
|
||||
init.append(", aaSorting: [[0, 'asc']]");
|
||||
|
@ -32,10 +32,20 @@
|
||||
public class ClusterMetricsInfo {
|
||||
|
||||
protected int appsSubmitted;
|
||||
protected int appsCompleted;
|
||||
protected int appsPending;
|
||||
protected int appsRunning;
|
||||
protected int appsFailed;
|
||||
protected int appsKilled;
|
||||
|
||||
protected long reservedMB;
|
||||
protected long availableMB;
|
||||
protected long allocatedMB;
|
||||
|
||||
protected int containersAllocated;
|
||||
protected int containersReserved;
|
||||
protected int containersPending;
|
||||
|
||||
protected long totalMB;
|
||||
protected int totalNodes;
|
||||
protected int lostNodes;
|
||||
@ -53,10 +63,20 @@ public ClusterMetricsInfo(final ResourceManager rm, final RMContext rmContext) {
|
||||
ClusterMetrics clusterMetrics = ClusterMetrics.getMetrics();
|
||||
|
||||
this.appsSubmitted = metrics.getAppsSubmitted();
|
||||
this.appsCompleted = metrics.getAppsCompleted();
|
||||
this.appsPending = metrics.getAppsPending();
|
||||
this.appsRunning = metrics.getAppsRunning();
|
||||
this.appsFailed = metrics.getAppsFailed();
|
||||
this.appsKilled = metrics.getAppsKilled();
|
||||
|
||||
this.reservedMB = metrics.getReservedMB();
|
||||
this.availableMB = metrics.getAvailableMB();
|
||||
this.allocatedMB = metrics.getAllocatedMB();
|
||||
|
||||
this.containersAllocated = metrics.getAllocatedContainers();
|
||||
this.containersPending = metrics.getPendingContainers();
|
||||
this.containersReserved = metrics.getReservedContainers();
|
||||
|
||||
this.totalMB = availableMB + reservedMB + allocatedMB;
|
||||
this.activeNodes = clusterMetrics.getNumActiveNMs();
|
||||
this.lostNodes = clusterMetrics.getNumLostNMs();
|
||||
@ -71,6 +91,26 @@ public int getAppsSubmitted() {
|
||||
return this.appsSubmitted;
|
||||
}
|
||||
|
||||
public int getAppsCompleted() {
|
||||
return appsCompleted;
|
||||
}
|
||||
|
||||
public int getAppsPending() {
|
||||
return appsPending;
|
||||
}
|
||||
|
||||
public int getAppsRunning() {
|
||||
return appsRunning;
|
||||
}
|
||||
|
||||
public int getAppsFailed() {
|
||||
return appsFailed;
|
||||
}
|
||||
|
||||
public int getAppsKilled() {
|
||||
return appsKilled;
|
||||
}
|
||||
|
||||
public long getReservedMB() {
|
||||
return this.reservedMB;
|
||||
}
|
||||
@ -87,6 +127,14 @@ public int getContainersAllocated() {
|
||||
return this.containersAllocated;
|
||||
}
|
||||
|
||||
public int getReservedContainers() {
|
||||
return this.containersReserved;
|
||||
}
|
||||
|
||||
public int getPendingContainers() {
|
||||
return this.containersPending;
|
||||
}
|
||||
|
||||
public long getTotalMB() {
|
||||
return this.totalMB;
|
||||
}
|
||||
|
@ -32,6 +32,11 @@
|
||||
public class UserMetricsInfo {
|
||||
|
||||
protected int appsSubmitted;
|
||||
protected int appsCompleted;
|
||||
protected int appsPending;
|
||||
protected int appsRunning;
|
||||
protected int appsFailed;
|
||||
protected int appsKilled;
|
||||
protected int runningContainers;
|
||||
protected int pendingContainers;
|
||||
protected int reservedContainers;
|
||||
@ -54,10 +59,18 @@ public UserMetricsInfo(final ResourceManager rm, final RMContext rmContext,
|
||||
|
||||
if (userMetrics != null) {
|
||||
this.userMetricsAvailable = true;
|
||||
|
||||
this.appsSubmitted = userMetrics.getAppsSubmitted();
|
||||
this.appsCompleted = metrics.getAppsCompleted();
|
||||
this.appsPending = metrics.getAppsPending();
|
||||
this.appsRunning = metrics.getAppsRunning();
|
||||
this.appsFailed = metrics.getAppsFailed();
|
||||
this.appsKilled = metrics.getAppsKilled();
|
||||
|
||||
this.runningContainers = userMetrics.getAllocatedContainers();
|
||||
this.pendingContainers = userMetrics.getPendingContainers();
|
||||
this.reservedContainers = userMetrics.getReservedContainers();
|
||||
|
||||
this.reservedMB = userMetrics.getReservedMB();
|
||||
this.pendingMB = userMetrics.getPendingMB();
|
||||
this.allocatedMB = userMetrics.getAllocatedMB();
|
||||
@ -72,6 +85,26 @@ public int getAppsSubmitted() {
|
||||
return this.appsSubmitted;
|
||||
}
|
||||
|
||||
public int getAppsCompleted() {
|
||||
return appsCompleted;
|
||||
}
|
||||
|
||||
public int getAppsPending() {
|
||||
return appsPending;
|
||||
}
|
||||
|
||||
public int getAppsRunning() {
|
||||
return appsRunning;
|
||||
}
|
||||
|
||||
public int getAppsFailed() {
|
||||
return appsFailed;
|
||||
}
|
||||
|
||||
public int getAppsKilled() {
|
||||
return appsKilled;
|
||||
}
|
||||
|
||||
public long getReservedMB() {
|
||||
return this.reservedMB;
|
||||
}
|
||||
|
@ -142,7 +142,7 @@ public void testLimitsComputation() throws Exception {
|
||||
CapacityScheduler.parseQueue(csContext, csConf, null, "root",
|
||||
queues, queues,
|
||||
CapacityScheduler.queueComparator,
|
||||
CapacityScheduler.applicationComparator,
|
||||
CapacityScheduler.applicationComparator,
|
||||
TestUtils.spyHook);
|
||||
|
||||
LeafQueue queue = (LeafQueue)queues.get(A);
|
||||
@ -163,6 +163,10 @@ public void testLimitsComputation() throws Exception {
|
||||
expectedMaxActiveApps * (queue.getUserLimit() / 100.0f) *
|
||||
queue.getUserLimitFactor()),
|
||||
queue.getMaximumActiveApplicationsPerUser());
|
||||
assertEquals(
|
||||
(int)(clusterResource.getMemory() * queue.getAbsoluteCapacity()),
|
||||
queue.getMetrics().getAvailableMB()
|
||||
);
|
||||
|
||||
// Add some nodes to the cluster & test new limits
|
||||
clusterResource = Resources.createResource(120 * 16 * GB);
|
||||
@ -178,6 +182,10 @@ public void testLimitsComputation() throws Exception {
|
||||
(int)Math.ceil(expectedMaxActiveApps *
|
||||
(queue.getUserLimit() / 100.0f) * queue.getUserLimitFactor()),
|
||||
queue.getMaximumActiveApplicationsPerUser());
|
||||
assertEquals(
|
||||
(int)(clusterResource.getMemory() * queue.getAbsoluteCapacity()),
|
||||
queue.getMetrics().getAvailableMB()
|
||||
);
|
||||
}
|
||||
|
||||
@Test
|
||||
|
@ -48,7 +48,7 @@ public class TestNodesPage {
|
||||
|
||||
// Number of Actual Table Headers for NodesPage.NodesBlock might change in
|
||||
// future. In that case this value should be adjusted to the new value.
|
||||
final int numberOfThInMetricsTable = 10;
|
||||
final int numberOfThInMetricsTable = 13;
|
||||
final int numberOfActualTableHeaders = 10;
|
||||
|
||||
private Injector injector;
|
||||
|
@ -361,6 +361,7 @@ public void verifyClusterMetricsXML(String xml) throws JSONException,
|
||||
|
||||
verifyClusterMetrics(
|
||||
WebServicesTestUtils.getXmlInt(element, "appsSubmitted"),
|
||||
WebServicesTestUtils.getXmlInt(element, "appsCompleted"),
|
||||
WebServicesTestUtils.getXmlInt(element, "reservedMB"),
|
||||
WebServicesTestUtils.getXmlInt(element, "availableMB"),
|
||||
WebServicesTestUtils.getXmlInt(element, "allocatedMB"),
|
||||
@ -379,8 +380,9 @@ public void verifyClusterMetricsJSON(JSONObject json) throws JSONException,
|
||||
Exception {
|
||||
assertEquals("incorrect number of elements", 1, json.length());
|
||||
JSONObject clusterinfo = json.getJSONObject("clusterMetrics");
|
||||
assertEquals("incorrect number of elements", 12, clusterinfo.length());
|
||||
verifyClusterMetrics(clusterinfo.getInt("appsSubmitted"),
|
||||
assertEquals("incorrect number of elements", 19, clusterinfo.length());
|
||||
verifyClusterMetrics(
|
||||
clusterinfo.getInt("appsSubmitted"), clusterinfo.getInt("appsCompleted"),
|
||||
clusterinfo.getInt("reservedMB"), clusterinfo.getInt("availableMB"),
|
||||
clusterinfo.getInt("allocatedMB"),
|
||||
clusterinfo.getInt("containersAllocated"),
|
||||
@ -390,7 +392,8 @@ public void verifyClusterMetricsJSON(JSONObject json) throws JSONException,
|
||||
clusterinfo.getInt("rebootedNodes"),clusterinfo.getInt("activeNodes"));
|
||||
}
|
||||
|
||||
public void verifyClusterMetrics(int sub, int reservedMB, int availableMB,
|
||||
public void verifyClusterMetrics(int submittedApps, int completedApps,
|
||||
int reservedMB, int availableMB,
|
||||
int allocMB, int containersAlloc, int totalMB, int totalNodes,
|
||||
int lostNodes, int unhealthyNodes, int decommissionedNodes,
|
||||
int rebootedNodes, int activeNodes) throws JSONException, Exception {
|
||||
@ -404,7 +407,9 @@ public void verifyClusterMetrics(int sub, int reservedMB, int availableMB,
|
||||
+ metrics.getAllocatedMB();
|
||||
|
||||
assertEquals("appsSubmitted doesn't match",
|
||||
metrics.getAppsSubmitted(), sub);
|
||||
metrics.getAppsSubmitted(), submittedApps);
|
||||
assertEquals("appsCompleted doesn't match",
|
||||
metrics.getAppsCompleted(), completedApps);
|
||||
assertEquals("reservedMB doesn't match",
|
||||
metrics.getReservedMB(), reservedMB);
|
||||
assertEquals("availableMB doesn't match",
|
||||
|
Loading…
Reference in New Issue
Block a user