MongoDB does not respect docker compose limits

Maulin_Tolia · November 28, 2022, 11:50am

I use MongoDB v5.0.3 via docker on a single AWS instance. The container kept crashing on long queries exactly after hitting 100% memory.
Then I noticed this:
This is the docker stat. You can notice that the memory is limited to 4 GB.

This is the MongoDB host info:

{
  system: {
    currentTime: 2022-11-28T10:23:25.772Z,
    hostname: 'mongo',
    cpuAddrSize: 64,
    memSizeMB: Long("31304"),
    memLimitMB: Long("31304"),
    numCores: 8,
    cpuArch: 'x86_64',
    numaEnabled: false
  }

You can see that the memLimitMB is set to 31 GB.

I rebuilt the container several times, but it still has the same issue.

What could be the reason?

John_Moser1 · December 27, 2022, 3:42pm

See my answer in

github.com/mongodb/mongodb-kubernetes-operator

MongoDB not respecting limits

opened 02:56PM - 05 Apr 22 UTC

AleCo3lho

CLOUDP-118989

Hello guys, I have a MongoDB Operator deployed in HA, and I'm facing this proble…m that even limiting the amount of memory that mongod must use, this limit is not being respected, which ends up taking a high memory consumption and consequently the pod is dropped. The node itself has 32gb of ram and we are reserving 30gb for the mongod container and limiting consumption to 23GB I have here some information from Grafana, showing the consumption of the container before a failure. ![memory](https://user-images.githubusercontent.com/20954739/161768821-f190f044-8fe2-432f-ac6d-c047db488f42.png) Just like the mongodb CRD yaml. ```yaml apiVersion: v1 items: - apiVersion: mongodbcommunity.mongodb.com/v1 kind: MongoDBCommunity metadata: annotations: mongodb.com/v1.lastAppliedMongoDBVersion: 5.0.6 generation: 9 name: prd-mongodb namespace: shared-database spec: additionalMongodConfig: storage.wiredTiger.engineConfig.cacheSizeGB: 23 storage.wiredTiger.engineConfig.journalCompressor: zlib members: 3 security: authentication: modes: - SCRAM statefulSet: spec: template: spec: containers: - env: - name: MANAGED_SECURITY_CONTEX value: "true" name: mongod resources: limits: cpu: "3.5" memory: 30Gi requests: cpu: "2" memory: 14Gi - env: - name: MANAGED_SECURITY_CONTEX value: "true" name: mongodb-agent resources: limits: cpu: "0.5" memory: 512M requests: cpu: "0.2" memory: 200M volumeClaimTemplates: - metadata: name: data-volume spec: accessModes: - ReadWriteOnce resources: requests: storage: 50Gi storageClassName: managed-premium - metadata: name: logs-volume spec: accessModes: - ReadWriteOnce resources: requests: storage: 10Gi storageClassName: managed-premium type: ReplicaSet users: - db: admin name: prd-mongodb-user passwordSecretRef: name: prd-mongodb-user-password roles: - db: admin name: clusterAdmin - db: admin name: userAdminAnyDatabase - db: admin name: root scramCredentialsSecretName: my-scram version: 5.0.6 status: currentMongoDBMembers: 3 currentStatefulSetReplicas: 3 mongoUri: <HIDE> phase: Running kind: List metadata: resourceVersion: "" selfLink: "" ``` Thank you so much everyone for your attention.