# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
"""Run ephemeral Docker Swarm services."""
from __future__ import annotations
from typing import TYPE_CHECKING
from docker import types
from airflow.exceptions import AirflowException
from airflow.providers.docker.operators.docker import DockerOperator
from airflow.utils.strings import get_random_string
if TYPE_CHECKING:
from airflow.utils.context import Context
[docs]class DockerSwarmOperator(DockerOperator):
"""
Execute a command as an ephemeral docker swarm service.
Example use-case - Using Docker Swarm orchestration to make one-time
scripts highly available.
A temporary directory is created on the host and
mounted into a container to allow storing files
that together exceed the default disk size of 10GB in a container.
The path to the mounted directory can be accessed
via the environment variable ``AIRFLOW_TMP_DIR``.
If a login to a private registry is required prior to pulling the image, a
Docker connection needs to be configured in Airflow and the connection ID
be provided with the parameter ``docker_conn_id``.
:param image: Docker image from which to create the container.
If image tag is omitted, "latest" will be used.
:param api_version: Remote API version. Set to ``auto`` to automatically
detect the server's version.
:param auto_remove: Auto-removal of the container on daemon side when the
container's process exits.
The default is False.
:param command: Command to be run in the container. (templated)
:param docker_url: URL of the host running the docker daemon.
Default is unix://var/run/docker.sock
:param environment: Environment variables to set in the container. (templated)
:param force_pull: Pull the docker image on every run. Default is False.
:param mem_limit: Maximum amount of memory the container can use.
Either a float value, which represents the limit in bytes,
or a string like ``128m`` or ``1g``.
:param tls_ca_cert: Path to a PEM-encoded certificate authority
to secure the docker connection.
:param tls_client_cert: Path to the PEM-encoded certificate
used to authenticate docker client.
:param tls_client_key: Path to the PEM-encoded key used to authenticate docker client.
:param tls_hostname: Hostname to match against
the docker server certificate or False to disable the check.
:param tls_ssl_version: Version of SSL to use when communicating with docker daemon.
:param tmp_dir: Mount point inside the container to
a temporary directory created on the host by the operator.
The path is also made available via the environment variable
``AIRFLOW_TMP_DIR`` inside the container.
:param user: Default user inside the docker container.
:param docker_conn_id: The :ref:`Docker connection id <howto/connection:docker>`
:param tty: Allocate pseudo-TTY to the container of this service
This needs to be set see logs of the Docker container / service.
:param enable_logging: Show the application's logs in operator's logs.
Supported only if the Docker engine is using json-file or journald logging drivers.
The `tty` parameter should be set to use this with Python applications.
:param configs: List of docker configs to be exposed to the containers of the swarm service.
The configs are ConfigReference objects as per the docker api
[https://docker-py.readthedocs.io/en/stable/services.html#docker.models.services.ServiceCollection.create]_
:param secrets: List of docker secrets to be exposed to the containers of the swarm service.
The secrets are SecretReference objects as per the docker create_service api.
[https://docker-py.readthedocs.io/en/stable/services.html#docker.models.services.ServiceCollection.create]_
:param mode: Indicate whether a service should be deployed as a replicated or global service,
and associated parameters
:param networks: List of network names or IDs or NetworkAttachmentConfig to attach the service to.
:param placement: Placement instructions for the scheduler. If a list is passed instead,
it is assumed to be a list of constraints as part of a Placement object.
"""
def __init__(
self,
*,
image: str,
enable_logging: bool = True,
configs: list[types.ConfigReference] | None = None,
secrets: list[types.SecretReference] | None = None,
mode: types.ServiceMode | None = None,
networks: list[str | types.NetworkAttachmentConfig] | None = None,
placement: types.Placement | list[types.Placement] | None = None,
**kwargs,
) -> None:
super().__init__(image=image, **kwargs)
self.enable_logging = enable_logging
self.service = None
self.configs = configs
self.secrets = secrets
self.mode = mode
self.networks = networks
self.placement = placement
[docs] def execute(self, context: Context) -> None:
self.environment["AIRFLOW_TMP_DIR"] = self.tmp_dir
return self._run_service()
def _run_service(self) -> None:
self.log.info("Starting docker service from image %s", self.image)
self.service = self.cli.create_service(
types.TaskTemplate(
container_spec=types.ContainerSpec(
image=self.image,
command=self.format_command(self.command),
mounts=self.mounts,
env=self.environment,
user=self.user,
tty=self.tty,
configs=self.configs,
secrets=self.secrets,
),
restart_policy=types.RestartPolicy(condition="none"),
resources=types.Resources(mem_limit=self.mem_limit),
networks=self.networks,
placement=self.placement,
),
name=f"airflow-{get_random_string()}",
labels={"name": f"airflow__{self.dag_id}__{self.task_id}"},
mode=self.mode,
)
if self.service is None:
raise Exception("Service should be set here")
self.log.info("Service started: %s", str(self.service))
# wait for the service to start the task
while not self.cli.tasks(filters={"service": self.service["ID"]}):
continue
if self.enable_logging:
self._stream_logs_to_output()
while True:
if self._has_service_terminated():
self.log.info("Service status before exiting: %s", self._service_status())
break
self.log.info("auto_removeauto_removeauto_removeauto_removeauto_remove : %s", str(self.auto_remove))
if self.service and self._service_status() != "complete":
if self.auto_remove == "success":
self.cli.remove_service(self.service["ID"])
raise AirflowException(f"Service did not complete: {self.service!r}")
elif self.auto_remove == "success":
if not self.service:
raise Exception("The 'service' should be initialized before!")
self.cli.remove_service(self.service["ID"])
def _service_status(self) -> str | None:
if not self.service:
raise Exception("The 'service' should be initialized before!")
return self.cli.tasks(filters={"service": self.service["ID"]})[0]["Status"]["State"]
def _has_service_terminated(self) -> bool:
status = self._service_status()
return status in ["complete", "failed", "shutdown", "rejected", "orphaned", "remove"]
def _stream_logs_to_output(self) -> None:
if not self.service:
raise Exception("The 'service' should be initialized before!")
logs = self.cli.service_logs(
self.service["ID"], follow=True, stdout=True, stderr=True, is_tty=self.tty
)
line = ""
for log in logs:
try:
log = log.decode()
except UnicodeDecodeError:
continue
if log == "\n":
self.log.info(line)
line = ""
else:
line += log
# flush any remaining log stream
if line:
self.log.info(line)
[docs] def on_kill(self) -> None:
if self.hook.client_created and self.service is not None:
self.log.info("Removing docker service: %s", self.service["ID"])
self.cli.remove_service(self.service["ID"])
