Unable to declare custom resources for initContainers-server-config-init in k8ssandra values.yaml

[anihe85]

I am getting below error with initContainers-server-config-init default value

  server-config-init:
    Container ID:   containerd://9431efb3d765e0f0d32365c355f22f2ba6efb2afd219eb5bf2a5715419940e6c
    Image:          docker.io/datastax/cass-config-builder:1.0.4
    Image ID:       docker.io/datastax/cass-config-builder@sha256:7b12c7303356307f451380e11f13be12dfb3aa33e6dd5b9e1704d89bacf9226a
    Port:           <none>
    Host Port:      <none>
    State:          Waiting
      Reason:       CrashLoopBackOff
    Last State:     Terminated
      Reason:       OOMKilled
      Exit Code:    137
      Started:      Sat, 06 May 2023 00:58:06 +0530
      Finished:     Sat, 06 May 2023 00:58:12 +0530
    Ready:          False
    Restart Count:  2
    Limits:
      cpu:     1
      memory:  256M
    Requests:
      cpu:     1
      memory:  256M

I have tried to declare like this but it seems it is not picking up

  initContainers:
    name: server-config-init
    resources:
      limits:
        cpu: 2
        memory: 1024M
      requests:
        cpu: 2
        memory: 1024M

[olim7t]

Hi @anihe85,

I was able to customize the limits, here is a full cluster definition:

apiVersion: k8ssandra.io/v1alpha1
kind: K8ssandraCluster
metadata:
  namespace: k8ssandra-operator
  name: demo
spec:
  cassandra:
    serverVersion: 4.0.6

    initContainers:
      - name: server-config-init
        resources:
          limits:
            cpu: 2
            memory: 1024M
          requests:
            cpu: 2
            memory: 1024M

    datacenters:
      - metadata:
          name: dc1
        size: 1
        racks:
          - name: rack1
            nodeAffinityLabels:
              "topology.kubernetes.io/zone": region1-zone1
        storageConfig:
          cassandraDataVolumeClaimSpec:
            storageClassName: standard
            accessModes:
              - ReadWriteOnce
            resources:
              requests:
                storage: 100Mi
        # (you can also have a per-DC `initContainers` at this level)

Checking the values once the pod is created:

kubectl get pod demo-dc1-rack1-sts-0 -o yaml -nk8ssandra-operator \
  | yq '.spec.initContainers[0] | {"name": .name, "resources": .resources}'

name: server-config-init
resources:
  limits:
    cpu: "2"
    memory: 1024M
  requests:
    cpu: "2"
    memory: 1024M

It would also be interesting to find out why the container OOM'ed in the first place, did you capture any logs?