Compute-Unit

Compute-Unit#

Use one or more units to manage computational workloads on a chosen platform. These units can be provisioned in local or cloud environments, allowing tasks to be scaled and deployed efficiently. Configuration and connectivity typically occur through command-line tools or interfaces provided by the hosting platform.

ComputeManager Class#

The ComputeManager class is responsible for managing compute units. It provides methods to create, list, update, and retrieve compute units.

How to get is though Session.get_compute_manager()

from dxflow.session import Session

session = Session(username="your_username", password="your_password")
compute_manager = session.get_compute_manager()
compute_manager.list()

Name        | Status     | IP            | Cluster Type         | CPU | Memory(GB) | Disk                        
------------+------------+---------------+----------------------+-----+------------+-----------------------------
Demo        | READY      | 54.211.250.47 | AWS:g6.2xlarge       | 8   | 32         | {boot: 512, volume: 256}    
TEST        | TERMINATED | 54.235.60.203 | AWS:T2 Medium        | 2   | 4          | {boot: 16, volume: 48}

Methods:#

  • get_unit(self, name: str = None, ip_address: str = None) -> 'ComputeUnit' | None: Retrieves a compute unit by name or IP address. It will return ComputeUnit object if found, or None if not found.

  • list(self, table: bool = True, return_info: bool = False): Lists all compute units in a tabular format or returns their information.

ComputeUnit Class#

The ComputeUnit class represents an individual compute unit and provides methods to manage its lifecycle and resources.

an example of how to get a compute unit:

demo_unit = compute_manager.get_unit(name="Demo")

Methods:#

  • detail(): Retrieves detailed information about the compute unit.

  • start(): Starts the compute unit.

  • stop(): Stops the compute unit.

  • troubleshoot(): Troubleshoots the compute unit.

  • terminate(): Terminates the compute unit.

  • stats(): Retrieves statistics (CPU, memory, disk) about the compute unit.

  • get_status(): Returns the current status of the compute unit (e.g., running, stopped, terminated).

  • create_project(project_name, template): Creates a project within the compute unit.

Properties:#

  • storage: The storage system associated with the compute unit.

  • project: The project manager for handling projects within the compute unit.

  • container: The container manager for managing containers within the compute unit.

Each compute unit has its own UnitStorage, UnitContainersManager, and UnitProjectsManager, which can be accessed through self.storage, self.containers, and self.projects, respectively. These components provide specialized functionality for managing storage, containers, and projects.

Details about the UnitProjectsManager, UnitStorage, and UnitContainersManager will be discussed in the following pages.

Contents