def generate_pod_spec_for_task(): # Primary containers do not require us to specify an image, the default image built for flyte tasks will get used. primary_container = V1Container(name="primary") # Note: for non-primary containers we must specify an image. secondary_container = V1Container(name="secondary", image="alpine",) secondary_container.command.extend(["/bin/sh"]) secondary_container.args.extend( ["-c", "echo hi pod world > {}".format(_SHARED_DATA_PATH)] ) resources = V1ResourceRequirements( requests={"cpu": "1", "memory": "100Mi"}, limits={"cpu": "1", "memory": "100Mi"} ) primary_container.resources = resources secondary_container = resources shared_volume_mount = V1VolumeMount(name="shared-data", mount_path="/data",) secondary_container.volumeMounts = [shared_volume_mount] primary_container.volumeMounts = [shared_volume_mount] pod_spec = V1PodSpec( containers=[primary_container, secondary_container], volumes=[ V1Volume( name="shared-data", empty_dir=V1EmptyDirVolumeSource(medium="Memory") ) ], ) return pod_spec
def get_pod_spec(): a_container = V1Container(name="a container", ) a_container.command = ["fee", "fi", "fo", "fum"] a_container.volume_mounts = [ V1VolumeMount( name="volume mount", mount_path="some/where", ) ] pod_spec = V1PodSpec( restart_policy="OnFailure", containers=[a_container, V1Container(name="another container")]) return pod_spec
def test_get_pod_manifest_tolerates_mixed_input(): """ Test that the get_pod_manifest function can handle a either a dictionary or an object both representing V1Container objects and that the function returns a V1Pod object containing V1Container objects. """ c = Config() dict_model = { 'name': 'mock_name_1', 'image': 'mock_image_1', 'command': ['mock_command_1'] } object_model = V1Container( name="mock_name_2", image="mock_image_2", command=['mock_command_2'], security_context=V1SecurityContext( privileged=True, run_as_user=0, capabilities=V1Capabilities(add=['NET_ADMIN']) ) ) c.KubeSpawner.init_containers = [dict_model, object_model] spawner = KubeSpawner(config=c, _mock=True) # this test ensures the following line doesn't raise an error manifest = sync_wait(spawner.get_pod_manifest()) # and tests the return value's types assert isinstance(manifest, V1Pod) assert isinstance(manifest.spec.init_containers[0], V1Container) assert isinstance(manifest.spec.init_containers[1], V1Container)
def _create_multus_init_container(image="lsstit/ddsnet4u:latest"): """Create the privileged container to allow CNI bridging for DDS multicast. """ return V1Container( name="multus-init", security_context=V1SecurityContext(privileged=True), image=image, )
def test_update_k8s_models_logger_warning(): """Ensure that the update_k8s_model function uses the logger to warn about overwriting previous values.""" target = V1Container( name="mock_name" ) source = {"name": "new_mock_name", "image_pull_policy": "Always"} mock_locker = MockLogger() update_k8s_model(target, source, logger=mock_locker, target_name="notebook_container", changes_name="extra_container_config") assert mock_locker.most_recent_warning.find("'notebook_container.name' current value: 'mock_name' is overridden with 'new_mock_name', which is the value of 'extra_container_config.name'") != -1 assert mock_locker.warning_count == 1
def _serialize_pod_spec(self, settings: SerializationSettings) -> Dict[str, Any]: containers = self.task_config.pod_spec.containers primary_exists = False for container in containers: if container.name == self.task_config.primary_container_name: primary_exists = True break if not primary_exists: # insert a placeholder primary container if it is not defined in the pod spec. containers.append( V1Container(name=self.task_config.primary_container_name)) final_containers = [] for container in containers: # In the case of the primary container, we overwrite specific container attributes with the default values # used in an SDK runnable task. if container.name == self.task_config.primary_container_name: sdk_default_container = super().get_container(settings) container.image = sdk_default_container.image # clear existing commands container.command = sdk_default_container.command # also clear existing args container.args = sdk_default_container.args limits, requests = {}, {} for resource in sdk_default_container.resources.limits: limits[_core_task.Resources.ResourceName.Name( resource.name).lower()] = resource.value for resource in sdk_default_container.resources.requests: requests[_core_task.Resources.ResourceName.Name( resource.name).lower()] = resource.value resource_requirements = V1ResourceRequirements( limits=limits, requests=requests) if len(limits) > 0 or len(requests) > 0: # Important! Only copy over resource requirements if they are non-empty. container.resources = resource_requirements container.env = [ V1EnvVar(name=key, value=val) for key, val in sdk_default_container.env.items() ] final_containers.append(container) self.task_config._pod_spec.containers = final_containers return ApiClient().sanitize_for_serialization( self.task_config.pod_spec)
def test_extract_pipelineparams_from_any(self): """Test extract_pipeleineparams.""" p1 = PipelineParam(name='param1', op_name='op1') p2 = PipelineParam(name='param2') p3 = PipelineParam(name='param3', value='value3') stuff_chars = ' between ' payload = str(p1) + stuff_chars + str(p2) + stuff_chars + str(p3) container = V1Container(name=p1, image=p2, env=[V1EnvVar(name="foo", value=payload)]) params = extract_pipelineparams_from_any(container) self.assertListEqual(sorted([p1, p2, p3]), sorted(params))
def test_fast_pod_task_serialization(): pod = Pod( pod_spec=V1PodSpec(restart_policy="OnFailure", containers=[V1Container(name="primary")]), primary_container_name="primary", ) @task(task_config=pod, environment={"FOO": "bar"}) def simple_pod_task(i: int): pass default_img = Image(name="default", fqn="test", tag="tag") serialization_settings = SerializationSettings( project="project", domain="domain", version="version", env={"FOO": "baz"}, image_config=ImageConfig(default_image=default_img, images=[default_img]), fast_serialization_settings=FastSerializationSettings(enabled=True), ) serialized = get_serializable(OrderedDict(), serialization_settings, simple_pod_task) assert serialized.template.k8s_pod.pod_spec["containers"][0]["args"] == [ "pyflyte-fast-execute", "--additional-distribution", "{{ .remote_package_path }}", "--dest-dir", "{{ .dest_dir }}", "--", "pyflyte-execute", "--inputs", "{{.input}}", "--output-prefix", "{{.outputPrefix}}", "--raw-output-data-prefix", "{{.rawOutputDataPrefix}}", "--checkpoint-path", "{{.checkpointOutputPrefix}}", "--prev-checkpoint", "{{.prevCheckpointPrefix}}", "--resolver", "flytekit.core.python_auto_container.default_task_resolver", "--", "task-module", "tests.test_pod", "task-name", "simple_pod_task", ]
def test_map_pod_task_serialization(): pod = Pod( pod_spec=V1PodSpec(restart_policy="OnFailure", containers=[V1Container(name="primary")]), primary_container_name="primary", ) @task(task_config=pod, environment={"FOO": "bar"}) def simple_pod_task(i: int): pass mapped_task = map_task(simple_pod_task, metadata=TaskMetadata(retries=1)) default_img = Image(name="default", fqn="test", tag="tag") serialization_settings = SerializationSettings( project="project", domain="domain", version="version", env={"FOO": "baz"}, image_config=ImageConfig(default_image=default_img, images=[default_img]), ) # Test that target is correctly serialized with an updated command pod_spec = mapped_task.get_k8s_pod(serialization_settings).pod_spec assert len(pod_spec["containers"]) == 1 assert pod_spec["containers"][0]["args"] == [ "pyflyte-map-execute", "--inputs", "{{.input}}", "--output-prefix", "{{.outputPrefix}}", "--raw-output-data-prefix", "{{.rawOutputDataPrefix}}", "--checkpoint-path", "{{.checkpointOutputPrefix}}", "--prev-checkpoint", "{{.prevCheckpointPrefix}}", "--resolver", "flytekit.core.python_auto_container.default_task_resolver", "--", "task-module", "tests.test_pod", "task-name", "simple_pod_task", ] assert { "primary_container_name": "primary" } == mapped_task.get_config(serialization_settings)
def test_update_k8s_model(): """Ensure update_k8s_model does what it should. The test is first updating attributes using the function and then and manually verifies that the correct changes have been made.""" manually_updated_target = V1Container( name="mock_name", image="mock_image", command=['iptables'], security_context=V1SecurityContext( privileged=True, run_as_user=0, capabilities=V1Capabilities(add=['NET_ADMIN']))) target = copy.deepcopy(manually_updated_target) source = {"name": "new_mock_name"} update_k8s_model(target, source) manually_updated_target.name = "new_mock_name" assert target == manually_updated_target
def make_task(operator: str, task_params: Dict[str, Any]) -> BaseOperator: """ Takes an operator and params and creates an instance of that operator. :returns: instance of operator object """ try: # class is a Callable https://stackoverflow.com/a/34578836/3679900 operator_obj: Callable[..., BaseOperator] = import_string(operator) except Exception as err: raise Exception(f"Failed to import operator: {operator}") from err try: if operator_obj in [PythonOperator, BranchPythonOperator]: if not task_params.get( "python_callable_name") and not task_params.get( "python_callable_file"): raise Exception( "Failed to create task. PythonOperator and BranchPythonOperator requires \ `python_callable_name` and `python_callable_file` parameters." ) task_params[ "python_callable"]: Callable = utils.get_python_callable( task_params["python_callable_name"], task_params["python_callable_file"], ) # remove dag-factory specific parameters # Airflow 2.0 doesn't allow these to be passed to operator del task_params["python_callable_name"] del task_params["python_callable_file"] # Check for the custom success and failure callables in SqlSensor. These are considered # optional, so no failures in case they aren't found. Note: there's no reason to # declare both a callable file and a lambda function for success/failure parameter. # If both are found the object will not throw and error, instead callable file will # take precedence over the lambda function if operator_obj in [SqlSensor]: # Success checks if task_params.get("success_check_file") and task_params.get( "success_check_name"): task_params[ "success"]: Callable = utils.get_python_callable( task_params["success_check_name"], task_params["success_check_file"], ) del task_params["success_check_name"] del task_params["success_check_file"] elif task_params.get("success_check_lambda"): task_params[ "success"]: Callable = utils.get_python_callable_lambda( task_params["success_check_lambda"]) del task_params["success_check_lambda"] # Failure checks if task_params.get("failure_check_file") and task_params.get( "failure_check_name"): task_params[ "failure"]: Callable = utils.get_python_callable( task_params["failure_check_name"], task_params["failure_check_file"], ) del task_params["failure_check_name"] del task_params["failure_check_file"] elif task_params.get("failure_check_lambda"): task_params[ "failure"]: Callable = utils.get_python_callable_lambda( task_params["failure_check_lambda"]) del task_params["failure_check_lambda"] if operator_obj in [HttpSensor]: if not (task_params.get("response_check_name") and task_params.get("response_check_file") ) and not task_params.get("response_check_lambda"): raise Exception( "Failed to create task. HttpSensor requires \ `response_check_name` and `response_check_file` parameters \ or `response_check_lambda` parameter.") if task_params.get("response_check_file"): task_params[ "response_check"]: Callable = utils.get_python_callable( task_params["response_check_name"], task_params["response_check_file"], ) # remove dag-factory specific parameters # Airflow 2.0 doesn't allow these to be passed to operator del task_params["response_check_name"] del task_params["response_check_file"] else: task_params[ "response_check"]: Callable = utils.get_python_callable_lambda( task_params["response_check_lambda"]) # remove dag-factory specific parameters # Airflow 2.0 doesn't allow these to be passed to operator del task_params["response_check_lambda"] # KubernetesPodOperator if operator_obj == KubernetesPodOperator: task_params["secrets"] = ([ Secret(**v) for v in task_params.get("secrets") ] if task_params.get("secrets") is not None else None) task_params["ports"] = ([ Port(**v) for v in task_params.get("ports") ] if task_params.get("ports") is not None else None) task_params["volume_mounts"] = ([ VolumeMount(**v) for v in task_params.get("volume_mounts") ] if task_params.get("volume_mounts") is not None else None) task_params["volumes"] = ([ Volume(**v) for v in task_params.get("volumes") ] if task_params.get("volumes") is not None else None) task_params["pod_runtime_info_envs"] = ([ PodRuntimeInfoEnv(**v) for v in task_params.get("pod_runtime_info_envs") ] if task_params.get("pod_runtime_info_envs") is not None else None) task_params["full_pod_spec"] = ( V1Pod(**task_params.get("full_pod_spec")) if task_params.get("full_pod_spec") is not None else None) task_params["init_containers"] = ([ V1Container(**v) for v in task_params.get("init_containers") ] if task_params.get("init_containers") is not None else None) if utils.check_dict_key(task_params, "execution_timeout_secs"): task_params["execution_timeout"]: timedelta = timedelta( seconds=task_params["execution_timeout_secs"]) del task_params["execution_timeout_secs"] if utils.check_dict_key(task_params, "sla_secs"): task_params["sla"]: timedelta = timedelta( seconds=task_params["sla_secs"]) del task_params["sla_secs"] if utils.check_dict_key(task_params, "execution_delta_secs"): task_params["execution_delta"]: timedelta = timedelta( seconds=task_params["execution_delta_secs"]) del task_params["execution_delta_secs"] if utils.check_dict_key( task_params, "execution_date_fn_name") and utils.check_dict_key( task_params, "execution_date_fn_file"): task_params[ "execution_date_fn"]: Callable = utils.get_python_callable( task_params["execution_date_fn_name"], task_params["execution_date_fn_file"], ) del task_params["execution_date_fn_name"] del task_params["execution_date_fn_file"] # on_execute_callback is an Airflow 2.0 feature if utils.check_dict_key( task_params, "on_execute_callback" ) and version.parse(AIRFLOW_VERSION) >= version.parse("2.0.0"): task_params["on_execute_callback"]: Callable = import_string( task_params["on_execute_callback"]) if utils.check_dict_key(task_params, "on_failure_callback"): task_params["on_failure_callback"]: Callable = import_string( task_params["on_failure_callback"]) if utils.check_dict_key(task_params, "on_success_callback"): task_params["on_success_callback"]: Callable = import_string( task_params["on_success_callback"]) if utils.check_dict_key(task_params, "on_retry_callback"): task_params["on_retry_callback"]: Callable = import_string( task_params["on_retry_callback"]) # use variables as arguments on operator if utils.check_dict_key(task_params, "variables_as_arguments"): variables: List[Dict[str, str]] = task_params.get( "variables_as_arguments") for variable in variables: if Variable.get(variable["variable"], default_var=None) is not None: task_params[variable["attribute"]] = Variable.get( variable["variable"], default_var=None) del task_params["variables_as_arguments"] task: BaseOperator = operator_obj(**task_params) except Exception as err: raise Exception(f"Failed to create {operator_obj} task") from err return task
def make_pod(name, cmd, port, image_spec, image_pull_policy, image_pull_secret=None, node_selector=None, run_as_uid=None, run_as_gid=None, fs_gid=None, supplemental_gids=None, run_privileged=False, env={}, working_dir=None, volumes=[], volume_mounts=[], labels={}, annotations={}, cpu_limit=None, cpu_guarantee=None, mem_limit=None, mem_guarantee=None, extra_resource_limits=None, extra_resource_guarantees=None, lifecycle_hooks=None, init_containers=None, service_account=None, extra_container_config=None, extra_pod_config=None, extra_containers=None, scheduler_name=None): """ Make a k8s pod specification for running a user notebook. Parameters ---------- name: Name of pod. Must be unique within the namespace the object is going to be created in. Must be a valid DNS label. image_spec: Image specification - usually a image name and tag in the form of image_name:tag. Same thing you would use with docker commandline arguments image_pull_policy: Image pull policy - one of 'Always', 'IfNotPresent' or 'Never'. Decides when kubernetes will check for a newer version of image and pull it when running a pod. image_pull_secret: Image pull secret - Default is None -- set to your secret name to pull from private docker registry. port: Port the notebook server is going to be listening on cmd: The command used to execute the singleuser server. node_selector: Dictionary Selector to match nodes where to launch the Pods run_as_uid: The UID used to run single-user pods. The default is to run as the user specified in the Dockerfile, if this is set to None. run_as_gid: The GID used to run single-user pods. The default is to run as the primary group of the user specified in the Dockerfile, if this is set to None. fs_gid The gid that will own any fresh volumes mounted into this pod, if using volume types that support this (such as GCE). This should be a group that the uid the process is running as should be a member of, so that it can read / write to the volumes mounted. supplemental_gids: A list of GIDs that should be set as additional supplemental groups to the user that the container runs as. You may have to set this if you are deploying to an environment with RBAC/SCC enforced and pods run with a 'restricted' SCC which results in the image being run as an assigned user ID. The supplemental group IDs would need to include the corresponding group ID of the user ID the image normally would run as. The image must setup all directories/files any application needs access to, as group writable. run_privileged: Whether the container should be run in privileged mode. env: Dictionary of environment variables. volumes: List of dictionaries containing the volumes of various types this pod will be using. See k8s documentation about volumes on how to specify these volume_mounts: List of dictionaries mapping paths in the container and the volume( specified in volumes) that should be mounted on them. See the k8s documentaiton for more details working_dir: String specifying the working directory for the notebook container labels: Labels to add to the spawned pod. annotations: Annotations to add to the spawned pod. cpu_limit: Float specifying the max number of CPU cores the user's pod is allowed to use. cpu_guarentee: Float specifying the max number of CPU cores the user's pod is guaranteed to have access to, by the scheduler. mem_limit: String specifying the max amount of RAM the user's pod is allowed to use. String instead of float/int since common suffixes are allowed mem_guarantee: String specifying the max amount of RAM the user's pod is guaranteed to have access to. String ins loat/int since common suffixes are allowed lifecycle_hooks: Dictionary of lifecycle hooks init_containers: List of initialization containers belonging to the pod. service_account: Service account to mount on the pod. None disables mounting extra_container_config: Extra configuration (e.g. envFrom) for notebook container which is not covered by parameters above. extra_pod_config: Extra configuration (e.g. tolerations) for pod which is not covered by parameters above. extra_containers: Extra containers besides notebook container. Used for some housekeeping jobs (e.g. crontab). scheduler_name: A custom scheduler's name. """ pod = V1Pod() pod.kind = "Pod" pod.api_version = "v1" pod.metadata = V1ObjectMeta(name=name, labels=labels.copy(), annotations=annotations.copy()) pod.spec = V1PodSpec(containers=[]) pod.spec.restartPolicy = 'Never' security_context = V1PodSecurityContext() if fs_gid is not None: security_context.fs_group = int(fs_gid) if supplemental_gids is not None and supplemental_gids: security_context.supplemental_groups = [ int(gid) for gid in supplemental_gids ] if run_as_uid is not None: security_context.run_as_user = int(run_as_uid) if run_as_gid is not None: security_context.run_as_group = int(run_as_gid) pod.spec.security_context = security_context if image_pull_secret is not None: pod.spec.image_pull_secrets = [] image_secret = V1LocalObjectReference() image_secret.name = image_pull_secret pod.spec.image_pull_secrets.append(image_secret) if node_selector: pod.spec.node_selector = node_selector notebook_container = V1Container( name='notebook', image=image_spec, working_dir=working_dir, ports=[V1ContainerPort(name='notebook-port', container_port=port)], env=[V1EnvVar(k, v) for k, v in env.items()], args=cmd, image_pull_policy=image_pull_policy, lifecycle=lifecycle_hooks, resources=V1ResourceRequirements()) if service_account is None: # Add a hack to ensure that no service accounts are mounted in spawned pods # This makes sure that we don"t accidentally give access to the whole # kubernetes API to the users in the spawned pods. # Note: We don't simply use `automountServiceAccountToken` here since we wanna be compatible # with older kubernetes versions too for now. hack_volume = V1Volume(name='no-api-access-please', empty_dir={}) hack_volumes = [hack_volume] hack_volume_mount = V1VolumeMount( name='no-api-access-please', mount_path="/var/run/secrets/kubernetes.io/serviceaccount", read_only=True) hack_volume_mounts = [hack_volume_mount] # Non-hacky way of not mounting service accounts pod.spec.automount_service_account_token = False else: hack_volumes = [] hack_volume_mounts = [] pod.spec.service_account_name = service_account if run_privileged: notebook_container.security_context = V1SecurityContext( privileged=True) notebook_container.resources.requests = {} if cpu_guarantee: notebook_container.resources.requests['cpu'] = cpu_guarantee if mem_guarantee: notebook_container.resources.requests['memory'] = mem_guarantee if extra_resource_guarantees: for k in extra_resource_guarantees: notebook_container.resources.requests[ k] = extra_resource_guarantees[k] notebook_container.resources.limits = {} if cpu_limit: notebook_container.resources.limits['cpu'] = cpu_limit if mem_limit: notebook_container.resources.limits['memory'] = mem_limit if extra_resource_limits: for k in extra_resource_limits: notebook_container.resources.limits[k] = extra_resource_limits[k] notebook_container.volume_mounts = volume_mounts + hack_volume_mounts pod.spec.containers.append(notebook_container) if extra_container_config: for key, value in extra_container_config.items(): setattr(notebook_container, _map_attribute(notebook_container.attribute_map, key), value) if extra_pod_config: for key, value in extra_pod_config.items(): setattr(pod.spec, _map_attribute(pod.spec.attribute_map, key), value) if extra_containers: pod.spec.containers.extend(extra_containers) pod.spec.init_containers = init_containers pod.spec.volumes = volumes + hack_volumes if scheduler_name: pod.spec.scheduler_name = scheduler_name return pod
} test_dep = V1Deployment( 'apps/v1', 'Deployment', V1ObjectMeta(name='web-main', labels={'app': 'web'}, annotations={'opsani-current-mode': 'canary'}), V1DeploymentSpec( replicas=3, selector=V1LabelSelector(match_labels={'app': 'web', 'role': 'main'}), template=V1PodTemplateSpec( metadata=V1ObjectMeta(labels={'app': 'web', 'role': 'main'}), spec=V1PodSpec(containers=[V1Container( name='main', image='opsani/co-http:latest', args=['busy=400'], resources=V1ResourceRequirements( limits={'cpu': '125m', 'memory': '512Mi'}, requests={'cpu': '125m', 'memory': '512Mi'} ), ports=[V1ContainerPort(container_port=8080)] )]) ) ) ) def setup_module(module): if os.getenv('KUBERNETES_SERVICE_HOST'): # If running in a kubernetes cluster kubernetes.config.load_incluster_config() else: kubernetes.config.load_kube_config() # setup desired ns
def get_minimal_pod_task_config() -> Pod: primary_container = V1Container(name="flytetask") pod_spec = V1PodSpec(containers=[primary_container]) return Pod(pod_spec=pod_spec, primary_container_name="flytetask")
def build_management_cmd(config_dict: dict, cmd: str = "", tag: str = "") -> str: from kubernetes import client, config from kubernetes.client.models import V1Container from kubernetes.client.rest import ApiException from kubernetes.config.kube_config import KubeConfigMerger, KubeConfigLoader, KUBE_CONFIG_DEFAULT_LOCATION from kubernetes.config.config_exception import ConfigException config.load_kube_config() # Workaround to read the proxy-url as it is not currently read by load_kube_config() # TODO: submit as pull request to kubernetes python try: kcfg = KubeConfigMerger(KUBE_CONFIG_DEFAULT_LOCATION) k = KubeConfigLoader(config_dict=kcfg.config) proxy_url = k._cluster['proxy-url'] except ConfigException: proxy_url = None if tag is not None: print(tag) if proxy_url is not None: client.Configuration._default.proxy = proxy_url app_v1_api = client.AppsV1Api() try: deployment = app_v1_api.read_namespaced_deployment( config_dict['project_name'], config_dict['namespace'], pretty=False) except ApiException as e: raise ExecuteError('AppsV1Api deployment not installed') container_v1 = deployment.spec.template.spec.containers[0] image_pull_secrets = None try: image_pull_secrets = deployment.spec.template.spec.image_pull_secrets[ 0] except ApiException as e: raise ExecuteError( 'Exception when calling AppsV1Api->image_pull_secret \n') except TypeError as e: print( f'image_pull_secrets does not exist and its ok, carrying on. Info: {e}' ) pass if tag: image_tag = container_v1.image.rsplit(':', 1)[0] image_tag = f'{image_tag}:{tag}' container = V1Container(env_from=container_v1.env_from, env=container_v1.env, image=image_tag, command=cmd.split(), args=[], name="management", stdin=True, tty=True) else: container = V1Container(env_from=container_v1.env_from, env=container_v1.env, image=container_v1.image, command=cmd.split(), args=[], name="management", stdin=True, tty=True) def create_dict_json_attributes(obj): if not hasattr(obj, 'to_dict'): return obj obj_dict = obj.to_dict() ret = dict() for key, value in obj_dict.items(): attrib = getattr(obj, key) if attrib is None: # Don't patch with null values continue if isinstance(attrib, str) \ or isinstance(attrib, int) \ or isinstance(attrib, float) \ or isinstance(attrib, bool): ret[obj.attribute_map[key]] = value elif isinstance(attrib, list): ret[obj.attribute_map[key]] = [ create_dict_json_attributes(a) for a in attrib ] else: ret[obj.attribute_map[key]] = create_dict_json_attributes( attrib) return ret # Convert V1_k8s_object to dictionary container_dict = create_dict_json_attributes(container) # Create dictionary overrides = dict(spec=dict(containers=[container_dict])) if image_pull_secrets: image_p_s_dict = create_dict_json_attributes(image_pull_secrets) overrides["spec"]["imagePullSecrets"] = [image_p_s_dict] overrides_str = dumps(overrides) return f'kubectl run management --rm --tty=true --stdin=true '\ f'--image={container.image} '\ f'--overrides=\'{overrides_str}\' '\ f'--output yaml --command -- \'\''
def make_pod( name, cmd, port, image_spec, image_pull_policy, image_pull_secret=None, node_selector=None, run_as_uid=None, run_as_gid=None, fs_gid=None, supplemental_gids=None, run_privileged=False, env=None, working_dir=None, volumes=None, volume_mounts=None, labels=None, annotations=None, cpu_limit=None, cpu_guarantee=None, mem_limit=None, mem_guarantee=None, extra_resource_limits=None, extra_resource_guarantees=None, lifecycle_hooks=None, init_containers=None, service_account=None, extra_container_config=None, extra_pod_config=None, extra_containers=None, scheduler_name=None, tolerations=None, node_affinity_preferred=None, node_affinity_required=None, pod_affinity_preferred=None, pod_affinity_required=None, pod_anti_affinity_preferred=None, pod_anti_affinity_required=None, priority_class_name=None, logger=None, userdir={}, ): """ Make a k8s pod specification for running a user notebook. Parameters ---------- name: Name of pod. Must be unique within the namespace the object is going to be created in. Must be a valid DNS label. image_spec: Image specification - usually a image name and tag in the form of image_name:tag. Same thing you would use with docker commandline arguments image_pull_policy: Image pull policy - one of 'Always', 'IfNotPresent' or 'Never'. Decides when kubernetes will check for a newer version of image and pull it when running a pod. image_pull_secret: Image pull secret - Default is None -- set to your secret name to pull from private docker registry. port: Port the notebook server is going to be listening on cmd: The command used to execute the singleuser server. node_selector: Dictionary Selector to match nodes where to launch the Pods run_as_uid: The UID used to run single-user pods. The default is to run as the user specified in the Dockerfile, if this is set to None. run_as_gid: The GID used to run single-user pods. The default is to run as the primary group of the user specified in the Dockerfile, if this is set to None. fs_gid The gid that will own any fresh volumes mounted into this pod, if using volume types that support this (such as GCE). This should be a group that the uid the process is running as should be a member of, so that it can read / write to the volumes mounted. supplemental_gids: A list of GIDs that should be set as additional supplemental groups to the user that the container runs as. You may have to set this if you are deploying to an environment with RBAC/SCC enforced and pods run with a 'restricted' SCC which results in the image being run as an assigned user ID. The supplemental group IDs would need to include the corresponding group ID of the user ID the image normally would run as. The image must setup all directories/files any application needs access to, as group writable. run_privileged: Whether the container should be run in privileged mode. env: Dictionary of environment variables. volumes: List of dictionaries containing the volumes of various types this pod will be using. See k8s documentation about volumes on how to specify these volume_mounts: List of dictionaries mapping paths in the container and the volume( specified in volumes) that should be mounted on them. See the k8s documentaiton for more details working_dir: String specifying the working directory for the notebook container labels: Labels to add to the spawned pod. annotations: Annotations to add to the spawned pod. cpu_limit: Float specifying the max number of CPU cores the user's pod is allowed to use. cpu_guarentee: Float specifying the max number of CPU cores the user's pod is guaranteed to have access to, by the scheduler. mem_limit: String specifying the max amount of RAM the user's pod is allowed to use. String instead of float/int since common suffixes are allowed mem_guarantee: String specifying the max amount of RAM the user's pod is guaranteed to have access to. String ins loat/int since common suffixes are allowed lifecycle_hooks: Dictionary of lifecycle hooks init_containers: List of initialization containers belonging to the pod. service_account: Service account to mount on the pod. None disables mounting extra_container_config: Extra configuration (e.g. envFrom) for notebook container which is not covered by parameters above. extra_pod_config: Extra configuration (e.g. tolerations) for pod which is not covered by parameters above. extra_containers: Extra containers besides notebook container. Used for some housekeeping jobs (e.g. crontab). scheduler_name: The pod's scheduler explicitly named. tolerations: Tolerations can allow a pod to schedule or execute on a tainted node. To learn more about pod tolerations, see https://kubernetes.io/docs/concepts/configuration/taint-and-toleration/. Pass this field an array of "Toleration" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#nodeselectorterm-v1-core node_affinity_preferred: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "PreferredSchedulingTerm" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#preferredschedulingterm-v1-core node_affinity_required: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "NodeSelectorTerm" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#nodeselectorterm-v1-core pod_affinity_preferred: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "WeightedPodAffinityTerm" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#weightedpodaffinityterm-v1-core pod_affinity_required: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "PodAffinityTerm" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#podaffinityterm-v1-core pod_anti_affinity_preferred: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "WeightedPodAffinityTerm" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#weightedpodaffinityterm-v1-core pod_anti_affinity_required: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "PodAffinityTerm" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#podaffinityterm-v1-core priority_class_name: The name of the PriorityClass to be assigned the pod. This feature is Beta available in K8s 1.11. """ pod = V1Pod() pod.kind = "Pod" pod.api_version = "v1" pod.metadata = V1ObjectMeta(name=name, labels=(labels or {}).copy(), annotations=(annotations or {}).copy()) pod.spec = V1PodSpec(containers=[]) pod.spec.restartPolicy = 'Never' security_context = V1PodSecurityContext() if fs_gid is not None: security_context.fs_group = int(fs_gid) if supplemental_gids is not None and supplemental_gids: security_context.supplemental_groups = [ int(gid) for gid in supplemental_gids ] if run_as_uid is not None: security_context.run_as_user = int(run_as_uid) if run_as_gid is not None: security_context.run_as_group = int(run_as_gid) pod.spec.security_context = security_context if image_pull_secret is not None: pod.spec.image_pull_secrets = [] image_secret = V1LocalObjectReference() image_secret.name = image_pull_secret pod.spec.image_pull_secrets.append(image_secret) if node_selector: pod.spec.node_selector = node_selector if lifecycle_hooks: lifecycle_hooks = get_k8s_model(V1Lifecycle, lifecycle_hooks) notebook_container = V1Container( name='notebook', image=image_spec, working_dir=working_dir, ports=[V1ContainerPort(name='notebook-port', container_port=port)], env=[V1EnvVar(k, v) for k, v in (env or {}).items()], args=cmd, image_pull_policy=image_pull_policy, lifecycle=lifecycle_hooks, resources=V1ResourceRequirements(), volume_mounts=[ get_k8s_model(V1VolumeMount, obj) for obj in (volume_mounts or []) ], ) if service_account is None: # This makes sure that we don't accidentally give access to the whole # kubernetes API to the users in the spawned pods. pod.spec.automount_service_account_token = False else: pod.spec.service_account_name = service_account if run_privileged: notebook_container.security_context = V1SecurityContext( privileged=True) notebook_container.resources.requests = {} if userdir.get('jp_cpu_request'): notebook_container.resources.requests['cpu'] = userdir.get( 'jp_cpu_request') elif cpu_guarantee: notebook_container.resources.requests['cpu'] = cpu_guarantee if userdir.get('jp_mem_request'): notebook_container.resources.requests['memory'] = userdir.get( 'jp_mem_request') elif mem_guarantee: notebook_container.resources.requests['memory'] = mem_guarantee if extra_resource_guarantees: notebook_container.resources.requests.update(extra_resource_guarantees) notebook_container.resources.limits = {} if userdir.get('jp_cpu_limit'): notebook_container.resources.limits['cpu'] = userdir.get( 'jp_cpu_limit') elif cpu_limit: notebook_container.resources.limits['cpu'] = cpu_limit if userdir.get('jp_mem_limit'): notebook_container.resources.limits['memory'] = userdir.get( 'jp_mem_limit') elif mem_limit: notebook_container.resources.limits['memory'] = mem_limit if extra_resource_limits: notebook_container.resources.limits.update(extra_resource_limits) if userdir.get("jp_gpu_enable", False): notebook_container.resources.limits[r'nvidia.com/gpu'] = userdir.get( 'jp_gpu_number', 0) if extra_container_config: notebook_container = update_k8s_model( target=notebook_container, changes=extra_container_config, logger=logger, target_name="notebook_container", changes_name="extra_container_config", ) pod.spec.containers.append(notebook_container) if extra_containers: pod.spec.containers.extend( [get_k8s_model(V1Container, obj) for obj in extra_containers]) if tolerations: pod.spec.tolerations = [ get_k8s_model(V1Toleration, obj) for obj in tolerations ] if init_containers: pod.spec.init_containers = [ get_k8s_model(V1Container, obj) for obj in init_containers ] if volumes: pod.spec.volumes = [get_k8s_model(V1Volume, obj) for obj in volumes] else: # Keep behaving exactly like before by not cleaning up generated pod # spec by setting the volumes field even though it is an empty list. pod.spec.volumes = [] if scheduler_name: pod.spec.scheduler_name = scheduler_name node_affinity = None if node_affinity_preferred or node_affinity_required: node_selector = None if node_affinity_required: node_selector = V1NodeSelector(node_selector_terms=[ get_k8s_model(V1NodeSelectorTerm, obj) for obj in node_affinity_required ], ) preferred_scheduling_terms = None if node_affinity_preferred: preferred_scheduling_terms = [ get_k8s_model(V1PreferredSchedulingTerm, obj) for obj in node_affinity_preferred ] node_affinity = V1NodeAffinity( preferred_during_scheduling_ignored_during_execution= preferred_scheduling_terms, required_during_scheduling_ignored_during_execution=node_selector, ) pod_affinity = None if pod_affinity_preferred or pod_affinity_required: weighted_pod_affinity_terms = None if pod_affinity_preferred: weighted_pod_affinity_terms = [ get_k8s_model(V1WeightedPodAffinityTerm, obj) for obj in pod_affinity_preferred ] pod_affinity_terms = None if pod_affinity_required: pod_affinity_terms = [ get_k8s_model(V1PodAffinityTerm, obj) for obj in pod_affinity_required ] pod_affinity = V1PodAffinity( preferred_during_scheduling_ignored_during_execution= weighted_pod_affinity_terms, required_during_scheduling_ignored_during_execution= pod_affinity_terms, ) pod_anti_affinity = None if pod_anti_affinity_preferred or pod_anti_affinity_required: weighted_pod_affinity_terms = None if pod_anti_affinity_preferred: weighted_pod_affinity_terms = [ get_k8s_model(V1WeightedPodAffinityTerm, obj) for obj in pod_anti_affinity_preferred ] pod_affinity_terms = None if pod_anti_affinity_required: pod_affinity_terms = [ get_k8s_model(V1PodAffinityTerm, obj) for obj in pod_anti_affinity_required ] pod_anti_affinity = V1PodAffinity( preferred_during_scheduling_ignored_during_execution= weighted_pod_affinity_terms, required_during_scheduling_ignored_during_execution= pod_affinity_terms, ) affinity = None if (node_affinity or pod_affinity or pod_anti_affinity): affinity = V1Affinity( node_affinity=node_affinity, pod_affinity=pod_affinity, pod_anti_affinity=pod_anti_affinity, ) if affinity: pod.spec.affinity = affinity if priority_class_name: pod.spec.priority_class_name = priority_class_name if extra_pod_config: pod.spec = update_k8s_model( target=pod.spec, changes=extra_pod_config, logger=logger, target_name="pod.spec", changes_name="extra_pod_config", ) return pod
def make_pod( name, cmd, port, image, image_pull_policy, image_pull_secrets=None, node_selector=None, uid=None, gid=None, fs_gid=None, supplemental_gids=None, privileged=False, allow_privilege_escalation=True, container_security_context=None, pod_security_context=None, env=None, working_dir=None, volumes=None, volume_mounts=None, labels=None, annotations=None, cpu_limit=None, cpu_guarantee=None, mem_limit=None, mem_guarantee=None, extra_resource_limits=None, extra_resource_guarantees=None, lifecycle_hooks=None, init_containers=None, service_account=None, automount_service_account_token=None, extra_container_config=None, extra_pod_config=None, extra_containers=None, scheduler_name=None, tolerations=None, node_affinity_preferred=None, node_affinity_required=None, pod_affinity_preferred=None, pod_affinity_required=None, pod_anti_affinity_preferred=None, pod_anti_affinity_required=None, priority_class_name=None, ssl_secret_name=None, ssl_secret_mount_path=None, logger=None, ): """ Make a k8s pod specification for running a user notebook. Parameters ---------- name: Name of pod. Must be unique within the namespace the object is going to be created in. Must be a valid DNS label. image: Image specification - usually a image name and tag in the form of image_name:tag. Same thing you would use with docker commandline arguments image_pull_policy: Image pull policy - one of 'Always', 'IfNotPresent' or 'Never'. Decides when kubernetes will check for a newer version of image and pull it when running a pod. image_pull_secrets: Image pull secrets - a list of references to Kubernetes Secret resources with credentials to pull images from image registries. This list can either have strings in it or objects with the string value nested under a name field. port: Port the notebook server is going to be listening on cmd: The command used to execute the singleuser server. node_selector: Dictionary Selector to match nodes where to launch the Pods uid: The UID used to run single-user pods. The default is to run as the user specified in the Dockerfile, if this is set to None. gid: The GID used to run single-user pods. The default is to run as the primary group of the user specified in the Dockerfile, if this is set to None. Setting this parameter requires that *feature-gate* **RunAsGroup** be enabled, otherwise the effective GID of the pod will be 0 (root). In addition, not setting `gid` once feature-gate RunAsGroup is enabled will also result in an effective GID of 0 (root). fs_gid The gid that will own any fresh volumes mounted into this pod, if using volume types that support this (such as GCE). This should be a group that the uid the process is running as should be a member of, so that it can read / write to the volumes mounted. supplemental_gids: A list of GIDs that should be set as additional supplemental groups to the user that the container runs as. You may have to set this if you are deploying to an environment with RBAC/SCC enforced and pods run with a 'restricted' SCC which results in the image being run as an assigned user ID. The supplemental group IDs would need to include the corresponding group ID of the user ID the image normally would run as. The image must setup all directories/files any application needs access to, as group writable. privileged: Whether the container should be run in privileged mode. allow_privilege_escalation: Controls whether a process can gain more privileges than its parent process. container_security_context: A kubernetes securityContext to apply to the container. pod_security_context: A kubernetes securityContext to apply to the pod. env: Dictionary of environment variables. volumes: List of dictionaries containing the volumes of various types this pod will be using. See k8s documentation about volumes on how to specify these volume_mounts: List of dictionaries mapping paths in the container and the volume( specified in volumes) that should be mounted on them. See the k8s documentaiton for more details working_dir: String specifying the working directory for the notebook container labels: Labels to add to the spawned pod. annotations: Annotations to add to the spawned pod. cpu_limit: Float specifying the max number of CPU cores the user's pod is allowed to use. cpu_guarantee: Float specifying the max number of CPU cores the user's pod is guaranteed to have access to, by the scheduler. mem_limit: String specifying the max amount of RAM the user's pod is allowed to use. String instead of float/int since common suffixes are allowed mem_guarantee: String specifying the max amount of RAM the user's pod is guaranteed to have access to. String ins loat/int since common suffixes are allowed lifecycle_hooks: Dictionary of lifecycle hooks init_containers: List of initialization containers belonging to the pod. service_account: Service account to mount on the pod. None disables mounting extra_container_config: Extra configuration (e.g. envFrom) for notebook container which is not covered by parameters above. extra_pod_config: Extra configuration (e.g. tolerations) for pod which is not covered by parameters above. extra_containers: Extra containers besides notebook container. Used for some housekeeping jobs (e.g. crontab). scheduler_name: The pod's scheduler explicitly named. tolerations: Tolerations can allow a pod to schedule or execute on a tainted node. To learn more about pod tolerations, see https://kubernetes.io/docs/concepts/configuration/taint-and-toleration/. Pass this field an array of "Toleration" objects. * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.20/#toleration-v1-core node_affinity_preferred: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "PreferredSchedulingTerm" objects. * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.20#preferredschedulingterm-v1-core node_affinity_required: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "NodeSelectorTerm" objects. * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.20/#nodeselectorterm-v1-core pod_affinity_preferred: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "WeightedPodAffinityTerm" objects. * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.20/#weightedpodaffinityterm-v1-core pod_affinity_required: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "PodAffinityTerm" objects. * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.20/#podaffinityterm-v1-core pod_anti_affinity_preferred: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "WeightedPodAffinityTerm" objects. * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.20/#weightedpodaffinityterm-v1-core pod_anti_affinity_required: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "PodAffinityTerm" objects. * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.20/#podaffinityterm-v1-core priority_class_name: The name of the PriorityClass to be assigned the pod. This feature is Beta available in K8s 1.11 and GA in 1.14. ssl_secret_name: Specifies the name of the ssl secret ssl_secret_mount_path: Specifies the name of the ssl secret mount path for the pod """ pod = V1Pod() pod.kind = "Pod" pod.api_version = "v1" pod.metadata = V1ObjectMeta( name=name, labels=(labels or {}).copy(), annotations=(annotations or {}).copy(), ) pod.spec = V1PodSpec(containers=[]) pod.spec.restart_policy = 'OnFailure' if image_pull_secrets is not None: # image_pull_secrets as received by the make_pod function should always # be a list, but it is allowed to have "a-string" elements or {"name": # "a-string"} elements. pod.spec.image_pull_secrets = [ V1LocalObjectReference(name=secret_ref) if type(secret_ref) == str else get_k8s_model(V1LocalObjectReference, secret_ref) for secret_ref in image_pull_secrets ] if ssl_secret_name and ssl_secret_mount_path: if not volumes: volumes = [] volumes.append( { 'name': 'jupyterhub-internal-certs', 'secret': {'secretName': ssl_secret_name, 'defaultMode': 511}, } ) env['JUPYTERHUB_SSL_KEYFILE'] = ssl_secret_mount_path + "ssl.key" env['JUPYTERHUB_SSL_CERTFILE'] = ssl_secret_mount_path + "ssl.crt" env['JUPYTERHUB_SSL_CLIENT_CA'] = ( ssl_secret_mount_path + "notebooks-ca_trust.crt" ) if not volume_mounts: volume_mounts = [] volume_mounts.append( { 'name': 'jupyterhub-internal-certs', 'mountPath': ssl_secret_mount_path, } ) if node_selector: pod.spec.node_selector = node_selector if lifecycle_hooks: lifecycle_hooks = get_k8s_model(V1Lifecycle, lifecycle_hooks) # Security contexts can be configured on Pod and Container level. The # Dedicated KubeSpawner API will bootstraps the container_security_context # except for if can only be configured on the Pod level, then it bootstraps # pod_security_context. # # The pod|container_security_context configuration is given a higher # priority than the dedicated KubeSpawner API options. # # Note that validation against the Python kubernetes-client isn't made as # the security contexts has evolved significantly and kubernetes-client is # too outdated. # # | Dedicated KubeSpawner API | Kubernetes API | Security contexts | # | -------------------------- | ------------------------ | ----------------- | # | supplemental_gids | supplementalGroups | Pod only | # | fs_gid | fsGroup | Pod only | # | - | fsGroupChangePolicy | Pod only | # | - | sysctls | Pod only | # | privileged | privileged | Container only | # | allow_privilege_escalation | allowPrivilegeEscalation | Container only | # | - | capabilities | Container only | # | - | procMount | Container only | # | - | readOnlyRootFilesystem | Container only | # | uid | runAsUser | Pod and Container | # | gid | runAsGroup | Pod and Container | # | - | runAsNonRoot | Pod and Container | # | - | seLinuxOptions | Pod and Container | # | - | seccompProfile | Pod and Container | # | - | windowsOptions | Pod and Container | # # ref: https://kubernetes.io/docs/tasks/configure-pod-container/security-context/ # ref: https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.20/#securitycontext-v1-core (container) # ref: https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.20/#podsecuritycontext-v1-core (pod) # psc = {} # populate with fs_gid / supplemental_gids if fs_gid is not None: psc["fsGroup"] = int(fs_gid) if supplemental_gids: psc["supplementalGroups"] = [int(gid) for gid in supplemental_gids] if pod_security_context: for key in pod_security_context.keys(): if "_" in key: raise ValueError( f"pod_security_context's keys should have k8s camelCase names, got '{key}'" ) psc.update(pod_security_context) if not psc: psc = None pod.spec.security_context = psc csc = {} # populate with uid / gid / privileged / allow_privilege_escalation if uid is not None: csc["runAsUser"] = int(uid) if gid is not None: csc["runAsGroup"] = int(gid) if privileged: # false as default csc["privileged"] = True if not allow_privilege_escalation: # true as default csc["allowPrivilegeEscalation"] = False if container_security_context: for key in container_security_context.keys(): if "_" in key: raise ValueError( f"container_security_context's keys should have k8s camelCase names, got '{key}'" ) csc.update(container_security_context) if not csc: csc = None def _get_env_var_deps(env): # only consider env var objects with an explicit string value if not env.value: return set() # $(MY_ENV) pattern: $( followed by non-)-characters to be captured, followed by ) re_k8s_env_reference_pattern = r"\$\(([^\)]+)\)" deps = set(re.findall(re_k8s_env_reference_pattern, env.value)) return deps - {env.name} unsorted_env = {} for key, env in (env or {}).items(): # Normalize KubeSpawners env input to valid Kubernetes EnvVar Python # representations. They should have a "name" field as well as either a # "value" field or "value_from" field. For examples see the # test_make_pod_with_env function. if type(env) == dict: if not "name" in env: env["name"] = key env = get_k8s_model(V1EnvVar, env) else: env = V1EnvVar(name=key, value=env) # Extract information about references to other envs as we want to use # those to make an intelligent sorting before we render this into a list # with an order that matters. unsorted_env[env.name] = { "deps": _get_env_var_deps(env), "key": key, "env": env, } # We sort environment variables in a way that allows dependencies to other # env to resolve as much as possible. There could be circular dependencies # so we will just do our best and settle with that. # # Algorithm description: # # - loop step: # - pop all unsorted_env entries with dependencies in sorted_env # - sort popped env based on key and extend the sorted_env list # - loop exit: # - exit if loop step didn't pop anything from unsorted_env # - before exit, sort what remains and extending the sorted_env list # sorted_env = [] while True: already_resolved_env_names = [e.name for e in sorted_env] extracted_env = {} for k, v in unsorted_env.copy().items(): if v["deps"].issubset(already_resolved_env_names): extracted_env[k] = unsorted_env.pop(k) if extracted_env: extracted_env = [ d["env"] for d in sorted(extracted_env.values(), key=operator.itemgetter("key")) ] sorted_env.extend(extracted_env) else: remaining_env = [ d["env"] for d in sorted(unsorted_env.values(), key=operator.itemgetter("key")) ] sorted_env.extend(remaining_env) break notebook_container = V1Container( name='notebook', image=image, working_dir=working_dir, ports=[V1ContainerPort(name='notebook-port', container_port=port)], env=sorted_env, args=cmd, image_pull_policy=image_pull_policy, lifecycle=lifecycle_hooks, resources=V1ResourceRequirements(), volume_mounts=[ get_k8s_model(V1VolumeMount, obj) for obj in (volume_mounts or []) ], security_context=csc, ) if service_account is not None: pod.spec.service_account_name = service_account if automount_service_account_token is None: if service_account is None: # This makes sure that we don't accidentally give access to the whole # kubernetes API to the users in the spawned pods. pod.spec.automount_service_account_token = False else: pod.spec.automount_service_account_token = automount_service_account_token notebook_container.resources.requests = {} if cpu_guarantee: notebook_container.resources.requests['cpu'] = cpu_guarantee if mem_guarantee: notebook_container.resources.requests['memory'] = mem_guarantee if extra_resource_guarantees: notebook_container.resources.requests.update(extra_resource_guarantees) notebook_container.resources.limits = {} if cpu_limit: notebook_container.resources.limits['cpu'] = cpu_limit if mem_limit: notebook_container.resources.limits['memory'] = mem_limit if extra_resource_limits: notebook_container.resources.limits.update(extra_resource_limits) if extra_container_config: notebook_container = update_k8s_model( target=notebook_container, changes=extra_container_config, logger=logger, target_name="notebook_container", changes_name="extra_container_config", ) pod.spec.containers.append(notebook_container) if extra_containers: pod.spec.containers.extend( [get_k8s_model(V1Container, obj) for obj in extra_containers] ) if tolerations: pod.spec.tolerations = [get_k8s_model(V1Toleration, obj) for obj in tolerations] if init_containers: pod.spec.init_containers = [ get_k8s_model(V1Container, obj) for obj in init_containers ] if volumes: pod.spec.volumes = [get_k8s_model(V1Volume, obj) for obj in volumes] else: # Keep behaving exactly like before by not cleaning up generated pod # spec by setting the volumes field even though it is an empty list. pod.spec.volumes = [] if scheduler_name: pod.spec.scheduler_name = scheduler_name node_affinity = None if node_affinity_preferred or node_affinity_required: node_selector = None if node_affinity_required: node_selector = V1NodeSelector( node_selector_terms=[ get_k8s_model(V1NodeSelectorTerm, obj) for obj in node_affinity_required ], ) preferred_scheduling_terms = None if node_affinity_preferred: preferred_scheduling_terms = [ get_k8s_model(V1PreferredSchedulingTerm, obj) for obj in node_affinity_preferred ] node_affinity = V1NodeAffinity( preferred_during_scheduling_ignored_during_execution=preferred_scheduling_terms, required_during_scheduling_ignored_during_execution=node_selector, ) pod_affinity = None if pod_affinity_preferred or pod_affinity_required: weighted_pod_affinity_terms = None if pod_affinity_preferred: weighted_pod_affinity_terms = [ get_k8s_model(V1WeightedPodAffinityTerm, obj) for obj in pod_affinity_preferred ] pod_affinity_terms = None if pod_affinity_required: pod_affinity_terms = [ get_k8s_model(V1PodAffinityTerm, obj) for obj in pod_affinity_required ] pod_affinity = V1PodAffinity( preferred_during_scheduling_ignored_during_execution=weighted_pod_affinity_terms, required_during_scheduling_ignored_during_execution=pod_affinity_terms, ) pod_anti_affinity = None if pod_anti_affinity_preferred or pod_anti_affinity_required: weighted_pod_affinity_terms = None if pod_anti_affinity_preferred: weighted_pod_affinity_terms = [ get_k8s_model(V1WeightedPodAffinityTerm, obj) for obj in pod_anti_affinity_preferred ] pod_affinity_terms = None if pod_anti_affinity_required: pod_affinity_terms = [ get_k8s_model(V1PodAffinityTerm, obj) for obj in pod_anti_affinity_required ] pod_anti_affinity = V1PodAffinity( preferred_during_scheduling_ignored_during_execution=weighted_pod_affinity_terms, required_during_scheduling_ignored_during_execution=pod_affinity_terms, ) affinity = None if node_affinity or pod_affinity or pod_anti_affinity: affinity = V1Affinity( node_affinity=node_affinity, pod_affinity=pod_affinity, pod_anti_affinity=pod_anti_affinity, ) if affinity: pod.spec.affinity = affinity if priority_class_name: pod.spec.priority_class_name = priority_class_name if extra_pod_config: pod.spec = update_k8s_model( target=pod.spec, changes=extra_pod_config, logger=logger, target_name="pod.spec", changes_name="extra_pod_config", ) return pod
def test_pod_resources(self): cases = [ { "pod": V1Pod(spec=V1PodSpec( init_containers=[ V1Container(name="1", resources=V1ResourceRequirements( limits={}, requests={})) ], containers=[ V1Container(name="1", resources=V1ResourceRequirements( limits={}, requests={})), V1Container(name="2", resources=V1ResourceRequirements( limits={}, requests={})) ])), # req_cpu, req_memory, lim_cpu, lim_memory, gpu_spec "expected": (0, 0, 0, 0, '') }, { "pod": V1Pod( spec=V1PodSpec( init_containers=[ V1Container(name="1", resources=V1ResourceRequirements( limits={ "cpu": "1", "memory": "512Mi" }, requests={ "cpu": "0.5", "memory": "256Mi" })) ], containers=[ V1Container(name="1", resources=V1ResourceRequirements( limits={ "cpu": "6", "memory": "6Gi" }, requests={ "cpu": "3", "memory": "0.5Gi" })), V1Container(name="2", resources=V1ResourceRequirements( limits={ "cpu": "2", "memory": "2Gi" }, requests={ "cpu": "1.5", "memory": "1Gi" })) ])), # req_cpu, req_memory, lim_cpu, lim_memory, gpu_spec "expected": (4.5, 1.5, 8.0, 8.0, '') } ] for case in cases: got = k8s_pod_resource_requirements(case['pod']) self.assertEqual(got, case['expected'])
def make_pod_spec(self, tls_secret, worker_name=None): annotations = self.common_annotations env = self.get_env() if worker_name is not None: # Worker name = "dask-gateway-worker-%s" % worker_name container_name = "dask-gateway-worker" labels = self.get_labels_for("dask-gateway-worker", worker_name=worker_name) mem_req = self.worker_memory mem_lim = self.worker_memory_limit cpu_req = self.worker_cores cpu_lim = self.worker_cores_limit env["DASK_GATEWAY_WORKER_NAME"] = worker_name cmd = self.worker_command extra_pod_config = self.worker_extra_pod_config extra_container_config = self.worker_extra_container_config else: # Scheduler name = "dask-gateway-scheduler-%s" % self.cluster_name container_name = "dask-gateway-scheduler" labels = self.get_labels_for("dask-gateway-scheduler") mem_req = self.scheduler_memory mem_lim = self.scheduler_memory_limit cpu_req = self.scheduler_cores cpu_lim = self.scheduler_cores_limit cmd = self.scheduler_command extra_pod_config = self.scheduler_extra_pod_config extra_container_config = self.scheduler_extra_container_config volume = V1Volume(name="dask-credentials", secret=V1SecretVolumeSource(secret_name=tls_secret)) container = V1Container( name=container_name, image=self.image, args=cmd, env=[V1EnvVar(k, v) for k, v in env.items()], image_pull_policy=self.image_pull_policy, resources=V1ResourceRequirements( requests={ "cpu": cpu_req, "memory": mem_req }, limits={ "cpu": cpu_lim, "memory": mem_lim }, ), volume_mounts=[ V1VolumeMount( name=volume.name, mount_path="/etc/dask-credentials/", read_only=True, ) ], ) if extra_container_config: container = merge_kube_objects(container, extra_container_config) pod = V1Pod( kind="Pod", api_version="v1", metadata=V1ObjectMeta(name=name, labels=labels, annotations=annotations), spec=V1PodSpec(containers=[container], volumes=[volume], restart_policy="Never"), ) # Ensure we don't accidentally give access to the kubernetes API pod.spec.automount_service_account_token = False if extra_pod_config: pod.spec = merge_kube_objects(pod.spec, extra_pod_config) return pod
# ==================== # # To use pod task as part of map task, we send pod task definition to :py:func:`~flytekit:flytekit.map_task`. # This will run pod task across a collection of inputs. @task(task_config=Pod( pod_spec=V1PodSpec( containers=[ V1Container( name="primary", resources=V1ResourceRequirements( requests={ "cpu": ".5", "memory": "500Mi" }, limits={ "cpu": ".5", "memory": "500Mi" }, ), ) ], init_containers=[ V1Container( image="alpine", name="init", command=["/bin/sh"], args=["-c", 'echo "I\'m a customizable init container"'], resources=V1ResourceRequirements(limits={ "cpu": ".5",
def make_pod( name, cmd, port, image, image_pull_policy, image_pull_secrets=None, node_selector=None, run_as_uid=None, run_as_gid=None, fs_gid=None, supplemental_gids=None, run_privileged=False, allow_privilege_escalation=True, env=None, working_dir=None, volumes=None, volume_mounts=None, labels=None, annotations=None, cpu_limit=None, cpu_guarantee=None, mem_limit=None, mem_guarantee=None, extra_resource_limits=None, extra_resource_guarantees=None, lifecycle_hooks=None, init_containers=None, service_account=None, extra_container_config=None, extra_pod_config=None, extra_containers=None, scheduler_name=None, tolerations=None, node_affinity_preferred=None, node_affinity_required=None, pod_affinity_preferred=None, pod_affinity_required=None, pod_anti_affinity_preferred=None, pod_anti_affinity_required=None, priority_class_name=None, logger=None, ): """ Make a k8s pod specification for running a user notebook. Parameters ---------- name: Name of pod. Must be unique within the namespace the object is going to be created in. Must be a valid DNS label. image: Image specification - usually a image name and tag in the form of image_name:tag. Same thing you would use with docker commandline arguments image_pull_policy: Image pull policy - one of 'Always', 'IfNotPresent' or 'Never'. Decides when kubernetes will check for a newer version of image and pull it when running a pod. image_pull_secrets: Image pull secrets - a list of references to Kubernetes Secret resources with credentials to pull images from image registries. This list can either have strings in it or objects with the string value nested under a name field. port: Port the notebook server is going to be listening on cmd: The command used to execute the singleuser server. node_selector: Dictionary Selector to match nodes where to launch the Pods run_as_uid: The UID used to run single-user pods. The default is to run as the user specified in the Dockerfile, if this is set to None. run_as_gid: The GID used to run single-user pods. The default is to run as the primary group of the user specified in the Dockerfile, if this is set to None. Setting this parameter requires that *feature-gate* **RunAsGroup** be enabled, otherwise the effective GID of the pod will be 0 (root). In addition, not setting `run_as_gid` once feature-gate RunAsGroup is enabled will also result in an effective GID of 0 (root). fs_gid The gid that will own any fresh volumes mounted into this pod, if using volume types that support this (such as GCE). This should be a group that the uid the process is running as should be a member of, so that it can read / write to the volumes mounted. supplemental_gids: A list of GIDs that should be set as additional supplemental groups to the user that the container runs as. You may have to set this if you are deploying to an environment with RBAC/SCC enforced and pods run with a 'restricted' SCC which results in the image being run as an assigned user ID. The supplemental group IDs would need to include the corresponding group ID of the user ID the image normally would run as. The image must setup all directories/files any application needs access to, as group writable. run_privileged: Whether the container should be run in privileged mode. allow_privilege_escalation: Controls whether a process can gain more privileges than its parent process. env: Dictionary of environment variables. volumes: List of dictionaries containing the volumes of various types this pod will be using. See k8s documentation about volumes on how to specify these volume_mounts: List of dictionaries mapping paths in the container and the volume( specified in volumes) that should be mounted on them. See the k8s documentaiton for more details working_dir: String specifying the working directory for the notebook container labels: Labels to add to the spawned pod. annotations: Annotations to add to the spawned pod. cpu_limit: Float specifying the max number of CPU cores the user's pod is allowed to use. cpu_guarentee: Float specifying the max number of CPU cores the user's pod is guaranteed to have access to, by the scheduler. mem_limit: String specifying the max amount of RAM the user's pod is allowed to use. String instead of float/int since common suffixes are allowed mem_guarantee: String specifying the max amount of RAM the user's pod is guaranteed to have access to. String ins loat/int since common suffixes are allowed lifecycle_hooks: Dictionary of lifecycle hooks init_containers: List of initialization containers belonging to the pod. service_account: Service account to mount on the pod. None disables mounting extra_container_config: Extra configuration (e.g. envFrom) for notebook container which is not covered by parameters above. extra_pod_config: Extra configuration (e.g. tolerations) for pod which is not covered by parameters above. extra_containers: Extra containers besides notebook container. Used for some housekeeping jobs (e.g. crontab). scheduler_name: The pod's scheduler explicitly named. tolerations: Tolerations can allow a pod to schedule or execute on a tainted node. To learn more about pod tolerations, see https://kubernetes.io/docs/concepts/configuration/taint-and-toleration/. Pass this field an array of "Toleration" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#nodeselectorterm-v1-core node_affinity_preferred: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "PreferredSchedulingTerm" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#preferredschedulingterm-v1-core node_affinity_required: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "NodeSelectorTerm" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#nodeselectorterm-v1-core pod_affinity_preferred: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "WeightedPodAffinityTerm" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#weightedpodaffinityterm-v1-core pod_affinity_required: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "PodAffinityTerm" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#podaffinityterm-v1-core pod_anti_affinity_preferred: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "WeightedPodAffinityTerm" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#weightedpodaffinityterm-v1-core pod_anti_affinity_required: Affinities describe where pods prefer or require to be scheduled, they may prefer or require a node to have a certain label or be in proximity / remoteness to another pod. To learn more visit https://kubernetes.io/docs/concepts/configuration/assign-pod-node/ Pass this field an array of "PodAffinityTerm" objects.* * https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.10/#podaffinityterm-v1-core priority_class_name: The name of the PriorityClass to be assigned the pod. This feature is Beta available in K8s 1.11. """ pod = V1Pod() pod.kind = "Pod" pod.api_version = "v1" pod.metadata = V1ObjectMeta(name=name, labels=(labels or {}).copy(), annotations=(annotations or {}).copy()) pod.spec = V1PodSpec(containers=[]) pod.spec.restart_policy = 'OnFailure' if image_pull_secrets is not None: # image_pull_secrets as received by the make_pod function should always # be a list, but it is allowed to have "a-string" elements or {"name": # "a-string"} elements. pod.spec.image_pull_secrets = [ V1LocalObjectReference(name=secret_ref) if type(secret_ref) == str else get_k8s_model(V1LocalObjectReference, secret_ref) for secret_ref in image_pull_secrets ] if node_selector: pod.spec.node_selector = node_selector if lifecycle_hooks: lifecycle_hooks = get_k8s_model(V1Lifecycle, lifecycle_hooks) # There are security contexts both on the Pod level or the Container level. # The security settings that you specify for a Pod apply to all Containers # in the Pod, but settings on the container level can override them. # # We configure the pod to be spawned on the container level unless the # option is only available on the pod level, such as for those relating to # the volumes as compared to the running user of the container. Volumes # belong to the pod and are only mounted by containers after all. # # ref: https://kubernetes.io/docs/tasks/configure-pod-container/security-context/ # ref: https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.16/#securitycontext-v1-core (container) # ref: https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.16/#podsecuritycontext-v1-core (pod) pod_security_context = V1PodSecurityContext() if fs_gid is not None: pod_security_context.fs_group = int(fs_gid) if supplemental_gids is not None and supplemental_gids: pod_security_context.supplemental_groups = [ int(gid) for gid in supplemental_gids ] # Only clutter pod spec with actual content if not all([e is None for e in pod_security_context.to_dict().values()]): pod.spec.security_context = pod_security_context container_security_context = V1SecurityContext() if run_as_uid is not None: container_security_context.run_as_user = int(run_as_uid) if run_as_gid is not None: container_security_context.run_as_group = int(run_as_gid) if run_privileged: container_security_context.privileged = True if not allow_privilege_escalation: container_security_context.allow_privilege_escalation = False # Only clutter container spec with actual content if all([e is None for e in container_security_context.to_dict().values()]): container_security_context = None # Transform a dict into valid Kubernetes EnvVar Python representations. This # representation shall always have a "name" field as well as either a # "value" field or "value_from" field. For examples see the # test_make_pod_with_env function. prepared_env = [] for k, v in (env or {}).items(): if type(v) == dict: if not "name" in v: v["name"] = k prepared_env.append(get_k8s_model(V1EnvVar, v)) else: prepared_env.append(V1EnvVar(name=k, value=v)) notebook_container = V1Container( name='notebook', image=image, working_dir=working_dir, ports=[V1ContainerPort(name='notebook-port', container_port=port)], env=prepared_env, args=cmd, image_pull_policy=image_pull_policy, lifecycle=lifecycle_hooks, resources=V1ResourceRequirements(), volume_mounts=[ get_k8s_model(V1VolumeMount, obj) for obj in (volume_mounts or []) ], security_context=container_security_context, ) if service_account is None: # This makes sure that we don't accidentally give access to the whole # kubernetes API to the users in the spawned pods. pod.spec.automount_service_account_token = False else: pod.spec.service_account_name = service_account notebook_container.resources.requests = {} if cpu_guarantee: notebook_container.resources.requests['cpu'] = cpu_guarantee if mem_guarantee: notebook_container.resources.requests['memory'] = mem_guarantee if extra_resource_guarantees: notebook_container.resources.requests.update(extra_resource_guarantees) notebook_container.resources.limits = {} if cpu_limit: notebook_container.resources.limits['cpu'] = cpu_limit if mem_limit: notebook_container.resources.limits['memory'] = mem_limit if extra_resource_limits: notebook_container.resources.limits.update(extra_resource_limits) if extra_container_config: notebook_container = update_k8s_model( target=notebook_container, changes=extra_container_config, logger=logger, target_name="notebook_container", changes_name="extra_container_config", ) pod.spec.containers.append(notebook_container) if extra_containers: pod.spec.containers.extend( [get_k8s_model(V1Container, obj) for obj in extra_containers]) if tolerations: pod.spec.tolerations = [ get_k8s_model(V1Toleration, obj) for obj in tolerations ] if init_containers: pod.spec.init_containers = [ get_k8s_model(V1Container, obj) for obj in init_containers ] if volumes: pod.spec.volumes = [get_k8s_model(V1Volume, obj) for obj in volumes] else: # Keep behaving exactly like before by not cleaning up generated pod # spec by setting the volumes field even though it is an empty list. pod.spec.volumes = [] if scheduler_name: pod.spec.scheduler_name = scheduler_name node_affinity = None if node_affinity_preferred or node_affinity_required: node_selector = None if node_affinity_required: node_selector = V1NodeSelector(node_selector_terms=[ get_k8s_model(V1NodeSelectorTerm, obj) for obj in node_affinity_required ], ) preferred_scheduling_terms = None if node_affinity_preferred: preferred_scheduling_terms = [ get_k8s_model(V1PreferredSchedulingTerm, obj) for obj in node_affinity_preferred ] node_affinity = V1NodeAffinity( preferred_during_scheduling_ignored_during_execution= preferred_scheduling_terms, required_during_scheduling_ignored_during_execution=node_selector, ) pod_affinity = None if pod_affinity_preferred or pod_affinity_required: weighted_pod_affinity_terms = None if pod_affinity_preferred: weighted_pod_affinity_terms = [ get_k8s_model(V1WeightedPodAffinityTerm, obj) for obj in pod_affinity_preferred ] pod_affinity_terms = None if pod_affinity_required: pod_affinity_terms = [ get_k8s_model(V1PodAffinityTerm, obj) for obj in pod_affinity_required ] pod_affinity = V1PodAffinity( preferred_during_scheduling_ignored_during_execution= weighted_pod_affinity_terms, required_during_scheduling_ignored_during_execution= pod_affinity_terms, ) pod_anti_affinity = None if pod_anti_affinity_preferred or pod_anti_affinity_required: weighted_pod_affinity_terms = None if pod_anti_affinity_preferred: weighted_pod_affinity_terms = [ get_k8s_model(V1WeightedPodAffinityTerm, obj) for obj in pod_anti_affinity_preferred ] pod_affinity_terms = None if pod_anti_affinity_required: pod_affinity_terms = [ get_k8s_model(V1PodAffinityTerm, obj) for obj in pod_anti_affinity_required ] pod_anti_affinity = V1PodAffinity( preferred_during_scheduling_ignored_during_execution= weighted_pod_affinity_terms, required_during_scheduling_ignored_during_execution= pod_affinity_terms, ) affinity = None if (node_affinity or pod_affinity or pod_anti_affinity): affinity = V1Affinity( node_affinity=node_affinity, pod_affinity=pod_affinity, pod_anti_affinity=pod_anti_affinity, ) if affinity: pod.spec.affinity = affinity if priority_class_name: pod.spec.priority_class_name = priority_class_name if extra_pod_config: pod.spec = update_k8s_model( target=pod.spec, changes=extra_pod_config, logger=logger, target_name="pod.spec", changes_name="extra_pod_config", ) return pod
def make_task(operator: str, task_params: Dict[str, Any]) -> BaseOperator: """ Takes an operator and params and creates an instance of that operator. :returns: instance of operator object """ try: # class is a Callable https://stackoverflow.com/a/34578836/3679900 operator_obj: Callable[..., BaseOperator] = import_string(operator) except Exception as err: raise f"Failed to import operator: {operator}" from err try: if operator_obj == PythonOperator: if not task_params.get( "python_callable_name") and not task_params.get( "python_callable_file"): raise Exception( "Failed to create task. PythonOperator requires `python_callable_name` \ and `python_callable_file` parameters.") task_params[ "python_callable"]: Callable = utils.get_python_callable( task_params["python_callable_name"], task_params["python_callable_file"], ) # KubernetesPodOperator if operator_obj == KubernetesPodOperator: task_params["secrets"] = ([ Secret(**v) for v in task_params.get("secrets") ] if task_params.get("secrets") is not None else None) task_params["ports"] = ([ Port(**v) for v in task_params.get("ports") ] if task_params.get("ports") is not None else None) task_params["volume_mounts"] = ([ VolumeMount(**v) for v in task_params.get("volume_mounts") ] if task_params.get("volume_mounts") is not None else None) task_params["volumes"] = ([ Volume(**v) for v in task_params.get("volumes") ] if task_params.get("volumes") is not None else None) task_params["pod_runtime_info_envs"] = ([ PodRuntimeInfoEnv(**v) for v in task_params.get("pod_runtime_info_envs") ] if task_params.get("pod_runtime_info_envs") is not None else None) task_params["full_pod_spec"] = ( V1Pod(**task_params.get("full_pod_spec")) if task_params.get("full_pod_spec") is not None else None) task_params["init_containers"] = ([ V1Container(**v) for v in task_params.get("init_containers") ] if task_params.get("init_containers") is not None else None) if utils.check_dict_key(task_params, "execution_timeout_secs"): task_params["execution_timeout"]: timedelta = timedelta( seconds=task_params["execution_timeout_secs"]) del task_params["execution_timeout_secs"] # use variables as arguments on operator if utils.check_dict_key(task_params, "variables_as_arguments"): variables: List[Dict[str, str]] = task_params.get( "variables_as_arguments") for variable in variables: if Variable.get(variable["variable"], default_var=None) is not None: task_params[variable["attribute"]] = Variable.get( variable["variable"], default_var=None) del task_params["variables_as_arguments"] task: BaseOperator = operator_obj(**task_params) except Exception as err: raise f"Failed to create {operator_obj} task" from err return task
def make_task(operator: str, task_params: Dict[str, Any], af_vars: Dict[str, Any]) -> BaseOperator: """ Takes an operator and params and creates an instance of that operator. :returns: instance of operator object """ try: # class is a Callable https://stackoverflow.com/a/34578836/3679900 operator_obj: Callable[..., BaseOperator] = import_string(operator) except Exception as err: raise Exception(f"Failed to import operator: {operator}") from err try: if operator_obj in [ PythonOperator, BranchPythonOperator, PythonSensor ]: if (not task_params.get("python_callable") and not task_params.get("python_callable_name") and not task_params.get("python_callable_file")): # pylint: disable=line-too-long raise Exception( "Failed to create task. PythonOperator, BranchPythonOperator and PythonSensor requires \ `python_callable_name` and `python_callable_file` " "parameters.\nOptionally you can load python_callable " "from a file. with the special pyyaml notation:\n" " python_callable_file: !!python/name:my_module.my_func" ) if not task_params.get("python_callable"): task_params[ "python_callable"]: Callable = utils.get_python_callable( task_params["python_callable_name"], task_params["python_callable_file"], ) # remove dag-factory specific parameters # Airflow 2.0 doesn't allow these to be passed to operator del task_params["python_callable_name"] del task_params["python_callable_file"] # Check for the custom success and failure callables in SqlSensor. These are considered # optional, so no failures in case they aren't found. Note: there's no reason to # declare both a callable file and a lambda function for success/failure parameter. # If both are found the object will not throw and error, instead callable file will # take precedence over the lambda function if operator_obj in [SqlSensor]: # Success checks if task_params.get("success_check_file") and task_params.get( "success_check_name"): task_params[ "success"]: Callable = utils.get_python_callable( task_params["success_check_name"], task_params["success_check_file"], ) del task_params["success_check_name"] del task_params["success_check_file"] elif task_params.get("success_check_lambda"): task_params[ "success"]: Callable = utils.get_python_callable_lambda( task_params["success_check_lambda"]) del task_params["success_check_lambda"] # Failure checks if task_params.get("failure_check_file") and task_params.get( "failure_check_name"): task_params[ "failure"]: Callable = utils.get_python_callable( task_params["failure_check_name"], task_params["failure_check_file"], ) del task_params["failure_check_name"] del task_params["failure_check_file"] elif task_params.get("failure_check_lambda"): task_params[ "failure"]: Callable = utils.get_python_callable_lambda( task_params["failure_check_lambda"]) del task_params["failure_check_lambda"] if operator_obj in [HttpSensor]: if not (task_params.get("response_check_name") and task_params.get("response_check_file") ) and not task_params.get("response_check_lambda"): raise Exception( "Failed to create task. HttpSensor requires \ `response_check_name` and `response_check_file` parameters \ or `response_check_lambda` parameter.") if task_params.get("response_check_file"): task_params[ "response_check"]: Callable = utils.get_python_callable( task_params["response_check_name"], task_params["response_check_file"], ) # remove dag-factory specific parameters # Airflow 2.0 doesn't allow these to be passed to operator del task_params["response_check_name"] del task_params["response_check_file"] else: task_params[ "response_check"]: Callable = utils.get_python_callable_lambda( task_params["response_check_lambda"]) # remove dag-factory specific parameters # Airflow 2.0 doesn't allow these to be passed to operator del task_params["response_check_lambda"] # KubernetesPodOperator if operator_obj == KubernetesPodOperator: task_params["secrets"] = ([ Secret(**v) for v in task_params.get("secrets") ] if task_params.get("secrets") is not None else None) task_params["ports"] = ([ Port(**v) for v in task_params.get("ports") ] if task_params.get("ports") is not None else None) task_params["volume_mounts"] = ([ VolumeMount(**v) for v in task_params.get("volume_mounts") ] if task_params.get("volume_mounts") is not None else None) task_params["volumes"] = ([ Volume(**v) for v in task_params.get("volumes") ] if task_params.get("volumes") is not None else None) task_params["pod_runtime_info_envs"] = ([ PodRuntimeInfoEnv(**v) for v in task_params.get("pod_runtime_info_envs") ] if task_params.get("pod_runtime_info_envs") is not None else None) task_params["full_pod_spec"] = ( V1Pod(**task_params.get("full_pod_spec")) if task_params.get("full_pod_spec") is not None else None) task_params["init_containers"] = ([ V1Container(**v) for v in task_params.get("init_containers") ] if task_params.get("init_containers") is not None else None) if operator_obj == DockerOperator: if task_params.get("environment") is not None: task_params["environment"] = { k: os.environ.get(v, v) for k, v in task_params["environment"].items() } if operator_obj == EcsOperator: for c in task_params["overrides"]["containerOverrides"]: if c.get('environment') is not None: for env in c['environment']: env['value'] = os.environ.get( env['value'], env['value']) if 'ECS_SECURITY_GROUPS' in af_vars and 'network_configuration' in task_params: task_params["network_configuration"]["awsvpcConfiguration"]['securityGroups'] \ = af_vars['ECS_SECURITY_GROUPS'] if 'ECS_SUBNETS' in af_vars and 'network_configuration' in task_params: task_params['network_configuration'][ "awsvpcConfiguration"]["subnets"] = af_vars[ "ECS_SUBNETS"] if af_vars.get('ECS_CLUSTER'): task_params['cluster'] = af_vars["ECS_CLUSTER"] task_params['task_definition'] = ( af_vars.get('ECS_CLUSTER') + '_' + task_params['task_definition']).lower() task_params['awslogs_group'] = \ task_params['awslogs_group'] + '/' + af_vars.get('ECS_CLUSTER').lower() if utils.check_dict_key(task_params, "execution_timeout_secs"): task_params["execution_timeout"]: timedelta = timedelta( seconds=task_params["execution_timeout_secs"]) del task_params["execution_timeout_secs"] if utils.check_dict_key(task_params, "sla_secs"): task_params["sla"]: timedelta = timedelta( seconds=task_params["sla_secs"]) del task_params["sla_secs"] if utils.check_dict_key(task_params, "execution_delta_secs"): task_params["execution_delta"]: timedelta = timedelta( seconds=task_params["execution_delta_secs"]) del task_params["execution_delta_secs"] if utils.check_dict_key( task_params, "execution_date_fn_name") and utils.check_dict_key( task_params, "execution_date_fn_file"): task_params[ "execution_date_fn"]: Callable = utils.get_python_callable( task_params["execution_date_fn_name"], task_params["execution_date_fn_file"], ) del task_params["execution_date_fn_name"] del task_params["execution_date_fn_file"] # on_execute_callback is an Airflow 2.0 feature if utils.check_dict_key( task_params, "on_execute_callback" ) and version.parse(AIRFLOW_VERSION) >= version.parse("2.0.0"): task_params["on_execute_callback"]: Callable = import_string( task_params["on_execute_callback"]) if utils.check_dict_key(task_params, "on_failure_callback"): task_params["on_failure_callback"]: Callable = import_string( task_params["on_failure_callback"]) if utils.check_dict_key(task_params, "on_success_callback"): task_params["on_success_callback"]: Callable = import_string( task_params["on_success_callback"]) if utils.check_dict_key(task_params, "on_retry_callback"): task_params["on_retry_callback"]: Callable = import_string( task_params["on_retry_callback"]) # use variables as arguments on operator if utils.check_dict_key(task_params, "variables_as_arguments"): variables: List[Dict[str, str]] = task_params.get( "variables_as_arguments") for variable in variables: if Variable.get(variable["variable"], default_var=None) is not None: task_params[variable["attribute"]] = Variable.get( variable["variable"], default_var=None) del task_params["variables_as_arguments"] # use variables as arguments on operator if utils.check_dict_key(task_params, "af_vars_as_arguments"): variables: List[Dict[str, str]] = task_params.get( "af_vars_as_arguments") for variable in variables: if af_vars.get(variable["variable"], None) is not None: task_params[variable["attribute"]] = af_vars.get( variable["variable"], None) del task_params["af_vars_as_arguments"] task: BaseOperator = operator_obj(**task_params) except Exception as err: raise Exception(f"Failed to create {operator_obj} task") from err return task