MongoDB Service Terminated by Linux OOM Killer – Adjusting OOMScoreAdjust to Improve Stability

Original Question or Issue:

MongoDB service becomes unstable or stops unexpectedly due to the Linux OOM (Out-Of-Memory) Killer terminating the process.

Environment:

• Product - FileCloud Server
• Version - Any
• Platform - Linux

Steps to Reproduce:

Error or Log Message:

Kernel logs may show entries similar to:

Out of memory: Kill process xxxx (mongod) score xxx or sacrifice childKilled process xxxx (mongod) total-vm:xxxxkB

MongoDB service may stop unexpectedly.

Defect or Enhancement Number:

Cause:

Linux OOM Killer selects and terminates the MongoDB process during memory pressure due to its OOM score.

Resolution or Workaround:

Adjust the OOMScoreAdjust value for the MongoDB service to reduce the likelihood of it being terminated:

1. Run:

   systemctl edit mongod

2. Add:

   [Service]
   OOMScoreAdjust=-900

3. Reload and restart:

   systemctl daemon-reload
   systemctl restart mongod

4. Verify:

   systemctl show mongod | grep OOMScoreAdjust

Notes: