[MAPREDUCE-4062] AM Launcher thread can hang forever

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.23.2
• Fix Version/s: 0.23.3, 2.0.2-alpha
• Component/s: mrv2
• Target Version/s: 0.23.3
• Assignee: Thomas Graves
• Reporter: Thomas Graves
• Created: 23/Mar/12 21:42
• Updated: 07/Sep/12 21:03
• Resolved: 03/Apr/12 17:08

Description

We saw an instance where the RM stopped launch Application masters. We found that the launcher thread was hung because something weird/bad happened to the NM node. Currently there is only 1 launcher thread (jira 4061 to fix that). We need this to not happen. Even once we increase the number of threads to > 1 if that many nodes go bad the RM would be stuck. Note that this was stuck like this for approximately 9 hours.

Stack trace on hung AM launcher:

"pool-1-thread-1" prio=10 tid=0x000000004343e800 nid=0x3a4c in Object.wait()
[0x000000004fad2000]
java.lang.Thread.State: WAITING (on object monitor)
at java.lang.Object.wait(Native Method)
at java.lang.Object.wait(Object.java:485)
at org.apache.hadoop.ipc.Client.call(Client.java:1076)

• locked <0x00002aab05a4f3f0> (a org.apache.hadoop.ipc.Client$Call)
  at
  org.apache.hadoop.yarn.ipc.ProtoOverHadoopRpcEngine$Invoker.invoke(ProtoOverHadoopRpcEngine.java:135)
  at $Proxy76.startContainer(Unknown Source)
  at
  org.apache.hadoop.yarn.api.impl.pb.client.ContainerManagerPBClientImpl.startContainer(ContainerManagerPBClientImpl.java:87)
  at
  org.apache.hadoop.yarn.server.resourcemanager.amlauncher.AMLauncher.launch(AMLauncher.java:118)
  at
  org.apache.hadoop.yarn.server.resourcemanager.amlauncher.AMLauncher.run(AMLauncher.java:265)
  at
  java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
  at
  java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
  at java.lang.Thread.run(Thread.java:619)

Issue Links

• is related to: MAPREDUCE-4061 RM only has 1 AM launcher thread