[AMBARI-16225] Blueprint Deployed Clusters Throw Exceptions On Startup

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 2.4.0
• Fix Version/s: 2.4.0
• Component/s: ambari-server
• Assignee: Magyari Sandor Szilard
• Reporter: Magyari Sandor Szilard
• Due: 06/May/16
• Created: 03/May/16 13:11
• Updated: 20/Jun/16 19:12
• Resolved: 04/May/16 15:09

Description

After deploying a cluster with standard services (ZK, HDFS, etc) using blueprints, when user restarts Ambari, there are exceptions like below as the server is starting up.

java.lang.NullPointerException
	at org.apache.ambari.server.controller.internal.Stack.<init>(Stack.java:164)
	at org.apache.ambari.server.topology.BlueprintImpl.parseStack(BlueprintImpl.java:292)
	at org.apache.ambari.server.topology.BlueprintImpl.<init>(BlueprintImpl.java:66)
	at org.apache.ambari.server.topology.BlueprintFactory.getBlueprint(BlueprintFactory.java:87)
	at org.apache.ambari.server.topology.PersistedStateImpl$ReplayedTopologyRequest.<init>(PersistedStateImpl.java:334)
	at org.apache.ambari.server.topology.PersistedStateImpl.getAllRequests(PersistedStateImpl.java:154)
	at org.apache.ambari.server.topology.TopologyManager.ensureInitialized(TopologyManager.java:126)
	at org.apache.ambari.server.topology.TopologyManager.onHostHeartBeatLost(TopologyManager.java:386)
	at org.apache.ambari.server.state.host.HostImpl$HostHeartbeatLostTransition.transition(HostImpl.java:408)
	at org.apache.ambari.server.state.host.HostImpl$HostHeartbeatLostTransition.transition(HostImpl.java:396)
	at org.apache.ambari.server.state.fsm.StateMachineFactory$SingleInternalArc.doTransition(StateMachineFactory.java:354)
	at org.apache.ambari.server.state.fsm.StateMachineFactory.doTransition(StateMachineFactory.java:294)
	at org.apache.ambari.server.state.fsm.StateMachineFactory.access$300(StateMachineFactory.java:39)
	at org.apache.ambari.server.state.fsm.StateMachineFactory$InternalStateMachine.doTransition(StateMachineFactory.java:440)
	at org.apache.ambari.server.state.host.HostImpl.handleEvent(HostImpl.java:584)
	at org.apache.ambari.server.agent.HeartbeatMonitor.doWork(HeartbeatMonitor.java:159)
	at org.apache.ambari.server.agent.HeartbeatMonitor.run(HeartbeatMonitor.java:120)
	at java.lang.Thread.run(Thread.java:745)

Looks like a the AMC is not initialized yet as the heartbeat from agents is being handled. It doesn't seem to affect operation of the cluster, but it looks really bad.

Issue Links

• relates to: AMBARI-16013 Host_status stuck in UNKNOWN status after blueprint deploy with host in heartbeat-lost
• links to: ReviewBoard