chart
alues that are defined in the root of the values file.
Type: object
Path: ``
replicaCount¶
Number of replicas for the Spark-on-K8S deployment if autoscaling is disabled
Default: 1
Type: number
Path: replicaCount
workers¶
API worker processes in each pod (1-64), set as SPARK_ON_K8S_API_WORKERS. null keeps the
image default (4 in the published image). Each worker is a separate process with its own
memory and connection limit; size resources accordingly. An env entry takes precedence.
Default: null
Type: string
Path: workers
imagePullSecrets¶
List of imagePullSecrets to be used
Default: []
Type: list
Path: imagePullSecrets
nameOverride¶
Override the chart name in the release resources
Default: ""
Type: string
Path: nameOverride
fullnameOverride¶
Override the fullname of the release resources
Default: ""
Type: string
Path: fullnameOverride
podAnnotations¶
Annotations to be added to the Spark-on-K8S deployment pods
Default: {}
Type: string
Path: podAnnotations
podLabels¶
Labels to be added to the Spark-on-K8S deployment pods
Default: {}
Type: string
Path: podLabels
resources¶
The resource configuration for the Spark-on-K8S deployment
Default: {}
Type: string
Path: resources
volumes¶
The list of volumes to be added to the Spark-on-K8S deployment pods
Default: []
Type: list
Path: volumes
volumeMounts¶
The list of volume mounts to be added to the Spark-on-K8S deployment pods
Default: []
Type: list
Path: volumeMounts
nodeSelector¶
The node selector for the Spark-on-K8S deployment
Default: {}
Type: string
Path: nodeSelector
tolerations¶
The tolerations for the Spark-on-K8S deployment
Default: []
Type: list
Path: tolerations
affinity¶
The affinity for the Spark-on-K8S deployment
Default: {}
Type: string
Path: affinity
env¶
A list of environment variables to be added to the Spark-on-K8S deployment.
Entries override the chart defaults for in-cluster authentication, API host
(0.0.0.0), API port (service.port), and the optional history server address.
Default: []
Type: list
Path: env
namespaces¶
Namespaces the Spark-on-K8S API serves and, with chart-managed RBAC, can access.
Empty (the default) selects only the release namespace, with a Role and RoleBinding
there. "*" selects every namespace, with a ClusterRole and ClusterRoleBinding that
grant the same rules cluster-wide. A comma-separated string such as
"namespace1,namespace2" selects those namespaces, with a Role and RoleBinding in
each. The API refuses other namespaces, including in its Spark UI proxy and log
streams.
Default: ""
Type: string
Path: namespaces
readOnly¶
If true, the Spark-on-K8S service will be able to read pods, services, pods logs,
and if false, it will be able to exec into pods and create/delete pods and services.
With namespaces "*", false also requires rbac.allowClusterWideWrite.
Default: true
Type: boolean
Path: readOnly