Role(value)Role specifies the tasks that will run on the node pool. Roles
can be specific to workloads. Exactly one
GkeNodePoolTarget
within the virtual cluster must have the DEFAULT role, which is
used to run all workloads that are not associated with a node pool.
Values:
ROLE_UNSPECIFIED (0):
Role is unspecified.
DEFAULT (1):
At least one node pool must have the DEFAULT role. Work
assigned to a role that is not associated with a node pool
is assigned to the node pool with the DEFAULT role. For
example, work assigned to the CONTROLLER role will be
assigned to the node pool with the DEFAULT role if no
node pool has the CONTROLLER role.
CONTROLLER (2):
Run work associated with the Dataproc control
plane (for example, controllers and webhooks).
Very low resource requirements.
SPARK_DRIVER (3):
Run work associated with a Spark driver of a
job.
SPARK_EXECUTOR (4):
Run work associated with a Spark executor of
a job.