[YARN-10822] Containers going from New to Scheduled transition for killed container on recovery

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.4.0
• Fix Version/s: 3.4.0
• Component/s: container, nodemanager
• Labels: pull-request-available
• Target Version/s: 3.4.0
• Hadoop Flags: Reviewed
• Assignee: Minni Mittal
• Reporter: Minni Mittal
• Created: 14/Jun/21 16:34
• Updated: 11/Feb/24 23:30
• Resolved: 01/Feb/22 17:36

Description

INFO  [91] ContainerImpl: Container container_e1171_1623422468672_2229_01_000738 transitioned from NEW to LOCALIZING

INFO  [91] ContainerImpl: Container container_e1171_1623422468672_2229_01_000738 transitioned from LOCALIZING to SCHEDULED

INFO  [91] ContainerScheduler: Opportunistic container container_e1171_1623422468672_2229_01_000738 will be queued at the NM.

INFO  [127] ContainerManagerImpl: Stopping container with container Id: container_e1171_1623422468672_2229_01_000738

INFO  [91] ContainerImpl: Container container_e1171_1623422468672_2229_01_000738 transitioned from SCHEDULED to KILLING

INFO  [91] ContainerImpl: Container container_e1171_1623422468672_2229_01_000738 transitioned from KILLING to CONTAINER_CLEANEDUP_AFTER_KILL

INFO  [91] NMAuditLogger: USER=defaultcafor1stparty OPERATION=Container Finished - Killed TARGET=ContainerImpl RESULT=SUCCESS APPID=application_1623422468672_2229 CONTAINERID=container_e1171_1623422468672_2229_01_000738

INFO  [91] ApplicationImpl: Removing container_e1171_1623422468672_2229_01_000738 from application application_1623422468672_2229

INFO  [91] ContainersMonitorImpl: Stopping resource-monitoring for container_e1171_1623422468672_2229_01_000738

INFO  [163] NodeStatusUpdaterImpl: Removed completed containers from NM context:[container_e1171_1623422468672_2229_01_000738]

NM restart happened and recovery is attempted

 

INFO  [1] ContainerManagerImpl: Recovering container_e1171_1623422468672_2229_01_000738 in state QUEUED with exit code -1000

INFO  [1] ApplicationImpl: Adding container_e1171_1623422468672_2229_01_000738 to application application_1623422468672_2229

INFO  [89] ContainerImpl: Container container_e1171_1623422468672_2229_01_000738 transitioned from NEW to SCHEDULED

INFO  [89] ContainerImpl: Container container_e1171_1623422468672_2229_01_000738 transitioned from SCHEDULED to KILLING

INFO  [89] ContainerImpl: Container container_e1171_1623422468672_2229_01_000738 transitioned from KILLING to CONTAINER_CLEANEDUP_AFTER_KILL

Ideally, when container got killed before restart, it should finish the container immediately. 

Issue Links

• links to: GitHub Pull Request #3632