You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-issues@hadoop.apache.org by "Kaushal Khator (Jira)" <ji...@apache.org> on 2025/09/19 10:09:00 UTC

[jira] [Created] (YARN-11861) Containers killed despite disabled memory enforcement when elastic memory is enabled

Kaushal Khator created YARN-11861:
-------------------------------------

             Summary: Containers killed despite disabled memory enforcement when elastic memory is enabled
                 Key: YARN-11861
                 URL: https://issues.apache.org/jira/browse/YARN-11861
             Project: Hadoop YARN
          Issue Type: Improvement
    Affects Versions: 3.4.2
            Reporter: Kaushal Khator


When both {{yarn.nodemanager.resource.memory.enforced=false}} and {{yarn.nodemanager.elastic-memory-control.enabled=true}} are configured, containers are still being terminated by the method ContainersMonitorImpl.checkLimit() method with the message "Container killed by YARN for exceeding physical memory limits".

This behavior is inconsistent with user expectations. When memory enforcement is disabled and elastic memory control is enabled, users expect containers to be preempted only when the overall system memory reaches its limit, not based on individual container limits.

*Current Behavior:*
 # User sets {{yarn.nodemanager.resource.memory.enforced=false}}
 # User enables {{yarn.nodemanager.elastic-memory-control.enabled=true}}
 # Containers exceeding their individual memory limits are still killed by ContainersMonitorImpl.checkLimit()
 # Log shows: "Container killed by YARN for exceeding physical memory limits"



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org