To review the state please   Get the affected directory path from CM > Configuration > Scope: YARN (MR2 included) > NodeManager Log Directory. The default is /var/log/hadoop-yarn Verify the available disk space for this path: # df -h /var/log/hadoop-yarn Look up what are the biggest disk space consumers on this mount point Clean up where possible   If no clean up is possible and there is sufficient disk space available then you can in decrease the alarm threshold in the CM > YARN > Configuration > Scope: Node Manager > Category: Monitoring > Log Directory Free Space Monitoring Percentage Thresholds configuration property. ... View more