Cassandra Cluster CRD

Cassandra database clusters can be created and configuring using the clusters.cassandra.rook.io custom resource definition (CRD).

Please refer to the the user guide walk-through for complete instructions. This page will explain all the available configuration options on the Cassandra CRD.

Sample

  1. apiVersion: cassandra.rook.io/v1alpha1
  2. kind: Cluster
  3. metadata:
  4.   name: rook-cassandra
  5.   namespace: rook-cassandra
  6. spec:
  7.   version: 3.11.1
  8.   repository: my-private-repo.io/cassandra
  9.   mode: cassandra
  10.   datacenter:
  11.     name: us-east-1
  12.     racks:
  13.       - name: us-east-1a
  14.         members: 3
  15.         storage:
  16.           volumeClaimTemplates:
  17.             - metadata:
  18.                 name: rook-cassandra-data
  19.               spec:
  20.                 storageClassName: my-storage-class
  21.                 resources:
  22.                   requests:
  23.                     storage: 200Gi
  24.         resources:
  25.           requests:
  26.             cpu: 8
  27.             memory: 32Gi
  28.           limits:
  29.             cpu: 8
  30.             memory: 32Gi
  31.         placement:
  32.           nodeAffinity:
  33.             requiredDuringSchedulingIgnoredDuringExecution:
  34.               nodeSelectorTerms:
  35.                 - matchExpressions:
  36.                   - key: failure-domain.beta.kubernetes.io/region
  37.                     operator: In
  38.                     values: us-east-1
  39.                   - key: failure-domain.beta.kubernetes.io/zone
  40.                     operator: In
  41.                     value: us-east-1a

Settings Explanation

Cluster Settings:

  • version: The version of Cassandra to use. It is used as the image tag to pull.
  • repository: Optional field. Specifies a custom image repo. If left unset, the official docker hub repo is used.
  • mode: Optional field. Specifies if this is a Cassandra or Scylla cluster. If left unset, it defaults to cassandra. Values: {scylla, cassandra}

In the Cassandra model, each cluster contains datacenters and each datacenter contains racks. At the moment, the operator only supports single datacenter setups.

Datacenter Settings:

  • name: Name of the datacenter. Usually, a datacenter corresponds to a region.
  • racks: List of racks for the specific datacenter.

Rack Settings:

  • name: Name of the rack. Usually, a rack corresponds to an availability zone.
  • members: Number of Cassandra members for the specific rack. (In Cassandra documentation, they are called nodes. We don’t call them nodes to avoid confusion as a Cassandra Node corresponds to a Kubernetes Pod, not a Kubernetes Node).
  • storage: Defines the volumes to use for each Cassandra member. Currently, only 1 volume is supported.
  • resources: Defines the CPU and RAM resources for the Cassandra Pods.
  • placement: Defines the placement of Cassandra Pods. Has the following subfields: