[MAPREDUCE-3072] NodeManager doesn't recognize kill -9 of AM container

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: 0.23.0
• Component/s: nodemanager

Environment

[criccomi@criccomi-ld trunk]$ svn info
Path: .
URL: http://svn.apache.org/repos/asf/hadoop/common/trunk
Repository Root: http://svn.apache.org/repos/asf
Repository UUID: 13f79535-47bb-0310-9956-ffa450edef68
Revision: 1174189
Node Kind: directory
Schedule: normal
Last Changed Author: szetszwo
Last Changed Rev: 1173990
Last Changed Date: 2011-09-22 01:25:20 -0700 (Thu, 22 Sep 2011)

• Assignee: Unassigned
• Reporter: Chris Riccomini
• Created: 22/Sep/11 16:39
• Updated: 22/Sep/11 16:45
• Resolved: 22/Sep/11 16:44

Description

If I kill -9 my application master's pid, the NM continues reporting that the container is running. I assume it should probably instead report back to the RM that the AM has died. Instead, it continues sending this status:

2011-09-22 09:33:13,352 INFO nodemanager.NodeStatusUpdaterImpl (NodeStatusUpdaterImpl.java:getNodeStatus(222)) - Sending out status for container: container_id {, app_attempt_id {, application_id

{, id: 1, cluster_timestamp: 1316707951832, }

, attemptId: 1, }, id: 1, }, state: C_RUNNING, diagnostics: "\n", exit_status: -1000,

2011-09-22 09:33:13,682 INFO monitor.ContainersMonitorImpl (ContainersMonitorImpl.java:run(402)) - Memory usage of ProcessTree 27263 for container-id container_1316707951832_0001_01_000001 : Virtual 0 bytes, limit : 2147483648 bytes; Physical 0 bytes, limit -1 bytes

This status keeps being sent forever.

Issue Links

• is duplicated by: MAPREDUCE-3031 Job Client goes into infinite loop when we kill AM