Compute Templates¶
Runnable Compute¶
For ETL Workloads
compute:
dataplane: hub
purpose: runnable
nodePool:
nodeSelector:
{{"dataos.io/purpose": "runnable"}}
tolerations:
- key: {{"dedicated"}}
operator: {{"Equal"}}
value: {{"runnable"}}
effect: {{"NoSchedule"}}
Sample Runnable Compute YAML
# Resource Section
name: "runnable-default-01"
version: v1
type: compute
layer: system
description: "runnable compute"
# Compute-specific Section
compute:
dataplane: hub
purpose: runnable
nodePool:
nodeSelector:
"dataos.io/purpose": "runnable"
tolerations:
- key: "dedicated"
operator: "Equal"
value: "runnable"
effect: "NoSchedule"
Query Compute¶
For Query Workloads
compute:
dataplane: hub
purpose: query
nodePool:
nodeSelector:
{{"dataos.io/purpose": "query"}}
tolerations:
- key: {{"dedicated"}}
operator: {{"Equal"}}
value: {{"query"}}
effect: {{"NoSchedule"}}
Sample Query Compute YAML
# Resource Section
name: "query-default"
version: v1
type: compute
layer: system
description: "default query compute"
# Compute-specific Section
compute:
dataplane: hub
purpose: query
nodePool:
nodeSelector:
"dataos.io/purpose": "query"
tolerations:
- key: "dedicated"
operator: "Equal"
value: "query"
effect: "NoSchedule"
GPU Compute¶
For Machine Learning Workloads
compute:
dataplane: {{hub}}
purpose: gpu
nodePool:
nodeSelector:
{{"dataos.io/purpose": "gpu"}}
tolerations:
- key: {{"dedicated"}}
operator: {{"Equal"}}
value: {{"gpu"}}
effect: {{"NoSchedule"}}
Sample GPU Compute YAML
# Resource Section
name: gpu
version: v1
type: compute
layer: system
description: gpu compute for jobs
# Compute-specific Section
compute:
dataplane: hub
purpose: gpu
nodePool:
nodeSelector:
"dataos.io/purpose": "gpu"
tolerations:
- key: "dedicated"
operator: "Equal"
value: "gpu"
effect: "NoSchedule"