fabrictestbed 1.1.1

FABRIC Python Client Library with CLI

FABRIC TESTBED USER LIBRARY AND CLI

Fabric User CLI for experiments

Overview

This package supports User facing APIs as well as CLI.

• Tokens: Token management
• Slices: Slice management
• Slivers: Sliver management
• Resources: Resource management

CLI Commands

Command	SubCommand	Action	Input	Output
tokens	issue	Issue token with projectname and scope	projectname Project Name, scope Scope	Points user to Credential Manager to generate the tokens
token	refresh	Refresh token	projectname Project Name, scope Scope, refreshtoken Refresh Token	Returns new identity and refresh tokens
token	revoke	Revoke token	refreshtoken Refresh Token	Success or Failure status
slices	query	Query user slice(s)	idtoken Identity Token, refreshtoken Refresh Token, projectname Project Name, scope Scope, sliceid Slice Id	List of Slices or Graph ML representing slice identified by Slice Id
slices	create	Create user slice	idtoken Identity Token, refreshtoken Refresh Token, projectname Project Name, scope Scope, slicename Slice Name, slicegraph Slice graph	List of Slivers created for the Slice
slices	delete	Delete user slice	idtoken Identity Token, refreshtoken Refresh Token, projectname Project Name, scope Scope, sliceid Slice Id	Success or Failure Status
slivers	query	Query user sliver(s)	idtoken Identity Token, refreshtoken Refresh Token, projectname Project Name, scope Scope, sliceid Slice Id, sliverid Sliver Id	List of Slivers for the slice identified by Slice Id or Sliver identified by Sliver Id
resources	query	Query resources	idtoken Identity Token, refreshtoken Refresh Token, projectname Project Name, scope Scope	Graph ML representing the available resources

API

SliceManager class implements the API supporting the operations listed above. Check example in Usage below.

Requirements

Python 3.9+

Pre-requisites

Ensure that following are installed

• virtualenv
• virtualenvwrapper

Installation

Multiple installation options possible. For CF development the recommended method is to install from GitHub MASTER branch:

$ mkvirtualenv fabrictestbed
$ workon fabrictestbed
$ pip install git+https://github.com/fabric-testbed/fabric-cli.git

For inclusion in tools, etc, use PyPi

$ mkvirtualenv fabrictestbed
$ workon fabrictestbed
$ pip install fabrictestbed

Usage (API)

User API supports token and orchestrator commands:

from fabrictestbed.slice_manager import SliceManager, Status
from fabrictestbed.slice_editor import ExperimentTopology, Capacities, ComponentType

orchestrator_host = "<ORCHESTRATOR FQDN>"
credmgr_host = "<CREDENTIAL MANAGER FQDN>"
fabric_refresh_token = "<REFRESH TOKEN>"

# Create API Client object
# Users can request tokens with different Project and Scopes by altering `project_name` and `scope`
# parameters in the refresh call below.
client = SliceManager(oc_host=orchestrator_host, cm_host=credmgr_host,
                      refresh_token=fabric_refresh_token, project_name='all', scope='all')

# Get new Fabric Identity Token and update Fabric Refresh Token
try:
    id_token, refresh_token = client.refresh_tokens()
except Exception as e:
    print("Exception occurred while getting tokens:{}".format(e))

# User is expected to update the refresh token in JupyterHub environment such as below
# fabric_refresh_token=client.get_refresh_token()
# %store fabric_refresh_token

# Query Resources
status, advertised_topology = client.resources()

print(f"Status: {status}")
if status == Status.OK:
    print(f"Toplogy: {advertised_topology}")

# Create Slice
# Create topology
t = ExperimentTopology()

# Add node
n1 = t.add_node(name='n1', site='RENC')

# Set capacities
cap = Capacities()
cap.set_fields(core=4, ram=64, disk=500)

# Set Properties
n1.set_properties(capacities=cap, image_type='qcow2', image_ref='default_centos_8')

# Add PCI devices
n1.add_component(ctype=ComponentType.SmartNIC, model='ConnectX-5', name='nic1')

# Add node
n2 = t.add_node(name='n2', site='UKY')

# Set properties
n2.set_properties(capacities=cap, image_type='qcow2', image_ref='default_centos_8')

# Add PCI devices
n2.add_component(ctype=ComponentType.GPU, model='Tesla T4', name='nic2')

# Add node
n3 = t.add_node(name='n3', site='LBNL')

# Set properties
n3.set_properties(capacities=cap, image_type='qcow2', image_ref='default_centos_8')

# Add PCI devices
n3.add_component(ctype=ComponentType.GPU, model='Tesla T4', name='nic3')

# Generate Slice Graph
slice_graph = t.serialize()

ssh_key = None
with open("/home/fabric/.ssh/id_rsa.pub", "r") as myfile:
    ssh_key = myfile.read()
    ssh_key = ssh_key.strip()

# Request slice from Orchestrator
status, reservations = client.create(slice_name='JupyterSlice2', slice_graph=slice_graph, ssh_key=ssh_key)

print("Response Status {}".format(status))
if status == Status.OK:
    print("Reservations created {}".format(reservations))

slice_id = reservations[0].slice_id
# Delete Slice
status, result = client.delete(slice_id=slice_id)

print("Response Status {}".format(status))
if status == Status.OK:
    print("Response received {}".format(result))

Usage (CLI)

Configuration

User CLI expects the user to set FABRIC_ORCHESTRATOR_HOST and FABRIC_CREDMGR_HOST environment variables.

In addition, User is expected to pass either Fabric Identity Token or Fabric Refresh Token to all the orchestrator commands. Alternatively, user is expected to set atleast one of the environment variables FABRIC_ID_TOKEN and FABRIC_REFRESH_TOKEN.

Create config.yml with default content as shown below.

To enable CLI auto-completion, add following line to your ~/.bashrc

eval "$(_FABRIC_CLI_COMPLETE=source_bash fabric-cli)"

Open a new shell to enable completion. Or run the eval command directly in your current shell to enable it temporarily.

User CLI supports token and orchestrator commands:

(usercli) $ fabric-cli
Usage: fabric-cli [OPTIONS] COMMAND [ARGS]...

Options:
  -v, --verbose
  --help         Show this message and exit.

Commands:
  resources  Resource management (set $FABRIC_ORCHESTRATOR_HOST to the...
  slices     Slice management (set $FABRIC_ORCHESTRATOR_HOST to the...
  slivers    Sliver management (set $FABRIC_ORCHESTRATOR_HOST to the...
  tokens     Token management (set $FABRIC_CREDMGR_HOST to the Credential...

Token Management Commands

List of the token commands supported can be found below:

(usercli) $ fabric-cli tokens
Usage: fabric-cli tokens [OPTIONS] COMMAND [ARGS]...

  Token management (set $FABRIC_CREDMGR_HOST to the Credential Manager
  Server)

Options:
  --help  Show this message and exit.

Commands:
  issue    Issue token with projectname and scope
  refresh  Refresh token
  revoke   Revoke token

Resource Management Commands

List of the resource commands supported can be found below:

$ fabric-cli resources
Usage: fabric-cli resources [OPTIONS] COMMAND [ARGS]...

  Query Resources (set $FABRIC_ORCHESTRATOR_HOST to the Control Framework
  Orchestrator)

Options:
  --help  Show this message and exit.

Commands:
  query  issue token with projectname and scope

Slice Management Commands

(usercli) $ fabric-cli slices
Usage: fabric-cli slices [OPTIONS] COMMAND [ARGS]...

  Slice management (set $FABRIC_ORCHESTRATOR_HOST to the Orchestrator)

Options:
  --help  Show this message and exit.

Commands:
  create  Create user slice
  delete  Delete user slice
  query   Query user slice(s)

Sliver Management Commands

(usercli) $ fabric-cli slivers
Usage: fabric-cli slivers [OPTIONS] COMMAND [ARGS]...

  Sliver management (set $FABRIC_ORCHESTRATOR_HOST to the Orchestrator)

Options:
  --help  Show this message and exit.

Commands:
  query  Query user slice sliver(s)