Using the code bellow it is possible to create a dask kubernetes cluster in azure aks.
It uses a remote scheduler (dask.config.set({"kubernetes.scheduler-service-type": "LoadBalancer"})) and works perfectly.
To use virtual nodes, uncomment the line extra_pod_config=virtual_config (which follows this official example).
It doesn't work, with the following error:
ACI does not support providing args without specifying the command. Please supply both command and args to the pod spec.
This is tied to passing containers: args: [dask-scheduler]
Which containers: command: should I supply to fix this issue?
Thank you
import dask
from dask.distributed import Client
from dask_kubernetes import KubeCluster, KubeConfig, make_pod_spec
image = "daskdev/dask"
cluster = "aks-cluster1"
dask.config.set({"kubernetes.scheduler-service-type": "LoadBalancer"})
dask.config.set({"distributed.comm.timeouts.connect": 180})
virtual_config = {
"nodeSelector": {
"kubernetes.io/role": "agent",
"beta.kubernetes.io/os": "linux",
"type": "virtual-kubelet",
},
"tolerations": [
{"key": "virtual-kubelet.io/provider", "operator": "Exists"},
],
}
pod_spec = make_pod_spec(
image=image,
# extra_pod_config=virtual_config,
memory_limit="2G",
memory_request="2G",
cpu_limit=1,
cpu_request=1,
threads_per_worker=1, # same as cpu
)
# az aks get-credentials --name aks-cluster1 --resource-group resource_group1
# cp ~/.kube/config ./aksconfig.yaml
auth = KubeConfig(config_file="./aksconfig.yaml", context=cluster,)
cluster = KubeCluster(
pod_spec, auth=auth, deploy_mode="remote", scheduler_service_wait_timeout=180
)
client = Client(cluster)the reason comes from this virtual kubelet protection: in the pod configuration, dask uses args to start a scheduler or worker, but no command is supplied.
So I explicitly added the entrypoint command command_entrypoint_explicit and it works: pods are created sucessfully.
Second problem: network names resolution. workers fail to connect to the scheduler by network name: tcp://{name}.{namespace}:{port}
Although tcp://{name}.{namespace}.svc.cluster.local:{port} works. I edited this in dask_kubernetes.core.Scheduler.start and it works.
Another option is the virtual_config bellow. Code bellow is a complete solution.
import dask
from dask.distributed import Client
from dask_kubernetes import KubeCluster, KubeConfig, make_pod_spec
dask.config.set({"kubernetes.scheduler-service-type": "LoadBalancer"})
dask.config.set({"distributed.comm.timeouts.connect": 180})
image = "daskdev/dask"
cluster = "aks-cluster-prod3"
virtual_config = {
"nodeSelector": {
"kubernetes.io/role": "agent",
"beta.kubernetes.io/os": "linux",
"type": "virtual-kubelet",
},
"tolerations": [
{"key": "virtual-kubelet.io/provider", "operator": "Exists"},
{"key": "azure.com/aci", "effect": "NoSchedule"},
],
"dnsConfig": {
"options": [{"name": "ndots", "value": "5"}],
"searches": [
"default.svc.cluster.local",
"svc.cluster.local",
"cluster.local",
],
},
}
# copied from: https://github.com/dask/dask-docker/blob/master/base/Dockerfile#L25
command_entrypoint_explicit = {
"command": ["tini", "-g", "--", "/usr/bin/prepare.sh"],
}
pod_spec = make_pod_spec(
image=image,
extra_pod_config=virtual_config,
extra_container_config=command_entrypoint_explicit,
memory_limit="2G",
memory_request="2G",
cpu_limit=1,
cpu_request=1,
threads_per_worker=1, # same as cpu
)
# az aks get-credentials --name aks-cluster1 --resource-group resource_group1
# cp ~/.kube/config ./aksconfig.yaml
auth = KubeConfig(config_file="./aksconfig.yaml", context=cluster,)
cluster = KubeCluster(
pod_spec,
auth=auth,
deploy_mode="remote",
scheduler_service_wait_timeout=180,
n_workers=1,
)
client = Client(cluster)