public virtual void TestRPCResponseId()
{
string node = "localhost";
Resource capability = BuilderUtils.NewResource(1024, 1);
RegisterNodeManagerRequest request = recordFactory.NewRecordInstance <RegisterNodeManagerRequest
>();
nodeId = NodeId.NewInstance(node, 1234);
request.SetNodeId(nodeId);
request.SetHttpPort(0);
request.SetResource(capability);
RegisterNodeManagerRequest request1 = recordFactory.NewRecordInstance <RegisterNodeManagerRequest
>();
request1.SetNodeId(nodeId);
request1.SetHttpPort(0);
request1.SetResource(capability);
resourceTrackerService.RegisterNodeManager(request1);
NodeStatus nodeStatus = recordFactory.NewRecordInstance <NodeStatus>();
nodeStatus.SetNodeId(nodeId);
NodeHealthStatus nodeHealthStatus = recordFactory.NewRecordInstance <NodeHealthStatus
>();
nodeHealthStatus.SetIsNodeHealthy(true);
nodeStatus.SetNodeHealthStatus(nodeHealthStatus);
NodeHeartbeatRequest nodeHeartBeatRequest = recordFactory.NewRecordInstance <NodeHeartbeatRequest
>();
nodeHeartBeatRequest.SetNodeStatus(nodeStatus);
nodeStatus.SetResponseId(0);
NodeHeartbeatResponse response = resourceTrackerService.NodeHeartbeat(nodeHeartBeatRequest
);
NUnit.Framework.Assert.IsTrue(response.GetResponseId() == 1);
nodeStatus.SetResponseId(response.GetResponseId());
response = resourceTrackerService.NodeHeartbeat(nodeHeartBeatRequest);
NUnit.Framework.Assert.IsTrue(response.GetResponseId() == 2);
/* try calling with less response id */
response = resourceTrackerService.NodeHeartbeat(nodeHeartBeatRequest);
NUnit.Framework.Assert.IsTrue(response.GetResponseId() == 2);
nodeStatus.SetResponseId(0);
response = resourceTrackerService.NodeHeartbeat(nodeHeartBeatRequest);
NUnit.Framework.Assert.IsTrue(NodeAction.Resync.Equals(response.GetNodeAction()));
NUnit.Framework.Assert.AreEqual("Too far behind rm response id:2 nm response id:0"
, response.GetDiagnosticsMessage());
}
/// <exception cref="System.IO.IOException"/>
/// <exception cref="Org.Apache.Hadoop.Yarn.Exceptions.YarnException"/>
public virtual void Heartbeat()
{
NodeStatus nodeStatus = Org.Apache.Hadoop.Yarn.Server.Resourcemanager.NodeManager
.CreateNodeStatus(nodeId, GetContainerStatuses(containers));
nodeStatus.SetResponseId(responseID);
NodeHeartbeatRequest request = recordFactory.NewRecordInstance <NodeHeartbeatRequest
>();
request.SetNodeStatus(nodeStatus);
NodeHeartbeatResponse response = resourceTrackerService.NodeHeartbeat(request);
responseID = response.GetResponseId();
}
public void Run()
{
int lastHeartBeatID = 0;
while (!this._enclosing.isStopped)
{
try
{
NodeHeartbeatResponse response = null;
NodeStatus nodeStatus = this._enclosing.GetNodeStatus(lastHeartBeatID);
NodeHeartbeatRequest request = NodeHeartbeatRequest.NewInstance(nodeStatus, this.
_enclosing.context.GetContainerTokenSecretManager().GetCurrentKey(), this._enclosing
.context.GetNMTokenSecretManager().GetCurrentKey());
response = this._enclosing.resourceTracker.NodeHeartbeat(request);
this._enclosing.nextHeartBeatInterval = response.GetNextHeartBeatInterval();
this.UpdateMasterKeys(response);
if (response.GetNodeAction() == NodeAction.Shutdown)
{
Org.Apache.Hadoop.Yarn.Server.Nodemanager.NodeStatusUpdaterImpl.Log.Warn("Recieved SHUTDOWN signal from Resourcemanager as part of heartbeat,"
+ " hence shutting down.");
Org.Apache.Hadoop.Yarn.Server.Nodemanager.NodeStatusUpdaterImpl.Log.Warn("Message from ResourceManager: "
+ response.GetDiagnosticsMessage());
this._enclosing.context.SetDecommissioned(true);
this._enclosing.dispatcher.GetEventHandler().Handle(new NodeManagerEvent(NodeManagerEventType
.Shutdown));
break;
}
if (response.GetNodeAction() == NodeAction.Resync)
{
Org.Apache.Hadoop.Yarn.Server.Nodemanager.NodeStatusUpdaterImpl.Log.Warn("Node is out of sync with ResourceManager,"
+ " hence resyncing.");
Org.Apache.Hadoop.Yarn.Server.Nodemanager.NodeStatusUpdaterImpl.Log.Warn("Message from ResourceManager: "
+ response.GetDiagnosticsMessage());
this._enclosing.rmIdentifier = ResourceManagerConstants.RmInvalidIdentifier;
this._enclosing.dispatcher.GetEventHandler().Handle(new NodeManagerEvent(NodeManagerEventType
.Resync));
this._enclosing.pendingCompletedContainers.Clear();
break;
}
this._enclosing.RemoveOrTrackCompletedContainersFromContext(response.GetContainersToBeRemovedFromNM
());
lastHeartBeatID = response.GetResponseId();
IList <ContainerId> containersToCleanup = response.GetContainersToCleanup();
if (!containersToCleanup.IsEmpty())
{
this._enclosing.dispatcher.GetEventHandler().Handle(new CMgrCompletedContainersEvent
(containersToCleanup, CMgrCompletedContainersEvent.Reason.ByResourcemanager));
}
IList <ApplicationId> appsToCleanup = response.GetApplicationsToCleanup();
this._enclosing.TrackAppsForKeepAlive(appsToCleanup);
if (!appsToCleanup.IsEmpty())
{
this._enclosing.dispatcher.GetEventHandler().Handle(new CMgrCompletedAppsEvent(appsToCleanup
, CMgrCompletedAppsEvent.Reason.ByResourcemanager));
}
IDictionary <ApplicationId, ByteBuffer> systemCredentials = response.GetSystemCredentialsForApps
();
if (systemCredentials != null && !systemCredentials.IsEmpty())
{
((NodeManager.NMContext) this._enclosing.context).SetSystemCrendentialsForApps(Org.Apache.Hadoop.Yarn.Server.Nodemanager.NodeStatusUpdaterImpl
.ParseCredentials(systemCredentials));
}
}
catch (ConnectException e)
{
this._enclosing.dispatcher.GetEventHandler().Handle(new NodeManagerEvent(NodeManagerEventType
.Shutdown));
throw new YarnRuntimeException(e);
}
catch (Exception e)
{
Org.Apache.Hadoop.Yarn.Server.Nodemanager.NodeStatusUpdaterImpl.Log.Error("Caught exception in status-updater"
, e);
}
finally
{
lock (this._enclosing.heartbeatMonitor)
{
this._enclosing.nextHeartBeatInterval = this._enclosing.nextHeartBeatInterval <=
0 ? YarnConfiguration.DefaultRmNmHeartbeatIntervalMs : this._enclosing.nextHeartBeatInterval;
try
{
Sharpen.Runtime.Wait(this._enclosing.heartbeatMonitor, this._enclosing.nextHeartBeatInterval
);
}
catch (Exception)
{
}
}
}
}
}
/// <exception cref="Org.Apache.Hadoop.Yarn.Exceptions.YarnException"/>
/// <exception cref="System.IO.IOException"/>
public virtual NodeHeartbeatResponse NodeHeartbeat(NodeHeartbeatRequest request)
{
NodeStatus remoteNodeStatus = request.GetNodeStatus();
NodeId nodeId = remoteNodeStatus.GetNodeId();
// 1. Check if it's a valid (i.e. not excluded) node
if (!this.nodesListManager.IsValidNode(nodeId.GetHost()))
{
string message = "Disallowed NodeManager nodeId: " + nodeId + " hostname: " + nodeId
.GetHost();
Log.Info(message);
shutDown.SetDiagnosticsMessage(message);
return(shutDown);
}
// 2. Check if it's a registered node
RMNode rmNode = this.rmContext.GetRMNodes()[nodeId];
if (rmNode == null)
{
/* node does not exist */
string message = "Node not found resyncing " + remoteNodeStatus.GetNodeId();
Log.Info(message);
resync.SetDiagnosticsMessage(message);
return(resync);
}
// Send ping
this.nmLivelinessMonitor.ReceivedPing(nodeId);
// 3. Check if it's a 'fresh' heartbeat i.e. not duplicate heartbeat
NodeHeartbeatResponse lastNodeHeartbeatResponse = rmNode.GetLastNodeHeartBeatResponse
();
if (remoteNodeStatus.GetResponseId() + 1 == lastNodeHeartbeatResponse.GetResponseId
())
{
Log.Info("Received duplicate heartbeat from node " + rmNode.GetNodeAddress() + " responseId="
+ remoteNodeStatus.GetResponseId());
return(lastNodeHeartbeatResponse);
}
else
{
if (remoteNodeStatus.GetResponseId() + 1 < lastNodeHeartbeatResponse.GetResponseId
())
{
string message = "Too far behind rm response id:" + lastNodeHeartbeatResponse.GetResponseId
() + " nm response id:" + remoteNodeStatus.GetResponseId();
Log.Info(message);
resync.SetDiagnosticsMessage(message);
// TODO: Just sending reboot is not enough. Think more.
this.rmContext.GetDispatcher().GetEventHandler().Handle(new RMNodeEvent(nodeId, RMNodeEventType
.Rebooting));
return(resync);
}
}
// Heartbeat response
NodeHeartbeatResponse nodeHeartBeatResponse = YarnServerBuilderUtils.NewNodeHeartbeatResponse
(lastNodeHeartbeatResponse.GetResponseId() + 1, NodeAction.Normal, null, null, null
, null, nextHeartBeatInterval);
rmNode.UpdateNodeHeartbeatResponseForCleanup(nodeHeartBeatResponse);
PopulateKeys(request, nodeHeartBeatResponse);
ConcurrentMap <ApplicationId, ByteBuffer> systemCredentials = rmContext.GetSystemCredentialsForApps
();
if (!systemCredentials.IsEmpty())
{
nodeHeartBeatResponse.SetSystemCredentialsForApps(systemCredentials);
}
// 4. Send status to RMNode, saving the latest response.
this.rmContext.GetDispatcher().GetEventHandler().Handle(new RMNodeStatusEvent(nodeId
, remoteNodeStatus.GetNodeHealthStatus(), remoteNodeStatus.GetContainersStatuses
(), remoteNodeStatus.GetKeepAliveApplications(), nodeHeartBeatResponse));
return(nodeHeartBeatResponse);
}