User Tools

Site Tools


Sidebar

openbach:manuals:2.x:developer_manual:openbach_api:scenario_builder:index

Scenario Builder

The scenario builder is a Python tool allowing to build scenarios programmatically in Python and generate JSON files that can be imported into a project of an existing OpenBACH installation.

Please refer to the Introduction to OpenBACH-extra and OpenBACH API page to see how to install and set up your platform to use the API described in this page.

In summary, the scenario builder allows to create a Scenario capable of holding, initializing and configuring the OpenBACH functions:

  • Start Job Instance
  • Start Scenario Instance
  • Stop Job Instance
  • Stop Scenario Instance
  • If
  • While
  • Push File

Scenario

Contruction

An OpenBACH scenario can be built using the Scenario class.

You need to provide a name for your scenario, and an optional description:

DESCRIPTION = """Simple scenario
 
This simple scenario example aims at showing the capabilities
of the scenario_builder API.
"""
scenario = Scenario('tutorial', DESCRIPTION)

Managing content

You can add or remove arguments, constants or OpenBACH functions to your scenario:

  • Arguments
scenario.add_argument(argument_name, argument_description)
scenario.add_arguments(**arguments)
scenario.remove_argument(argument_name)
scenario.remove_arguments(*arguments_names)

Pretty straightforward, the add_arguments version uses keyword arguments as argument_name=argument_description to call the add_argument version.

Arguments are placeholders that can be used as values for OpenBACH functions whose real value won't be known until the scenario is started.

  • Constants
scenario.add_constant(constant_name, constant_value)
scenario.add_constants(**constants)
scenario.remove_constant(constant_name)
scenario.remove_constants(*constants_names)

Similar to arguments, the add_constants version uses keyword arguments as constant_name=constant_value to call the add_constant version.

Constants are placeholders that can be used as values for OpenBACH functions whose value is already defined. Useful to factor in values common to several OpenBACH functions and change it in a single place if need be.

  • OpenBACH functions
scenario.add_function()
scenario.remove_function(openbach_function_instance)

Defining and configuring OpenBACH functions is explained in details later in this guide. Removing OpenBACH functions must be done using the object returned by the add_function call.

Exporting to OpenBACH

Once your scenario is ready, you can call the build method to get a dictionary representation of this scenario. This dictionary is suitable to be written in a file as JSON data ready to be imported into OpenBACH.

with open('tutorial.json', 'w') as scenario_file:
    json.dump(scenario.build(), scenario_file)

As a shortcut, the write method does exactly that; so this is equivalent:

scenario.write('tutorial.json')

It is up to you to then send your JSON file to OpenBACH using either the auditorium scripts or the web interface.

Utilities

The extract_function_id method will let you retrieve OpenBACH function's paths from their names.

scenario.extract_function_id(*job_names, include_subscenarios=False, **filtered_jobs)

It is a generator that will yield paths corresponding to all the given job_names; these paths are suitable to use as jobs arguments to post-processing jobs, for instance. The keyword arguments should assign a filtering function to a job name. Once a matching start_job_instance OpenBACH function is found by its name, it is passed as the sole argument to the filtering function which should return a boolean value indicating if this job should be considered or not. This allows to discriminate job by their configuration parameters (think iperf client vs. iperf server), for instance.

The find_openbach_function will return an OpenBachFunction instance for one of these paths.

OpenBACH Functions

Initialization

In order to initialize the OpenBACH function:

name_of_function = scenario.add_function(type_of_function, wait_delay=wait_delay, wait_launched=[...], wait_finished=[...])

The type_of_function is the only mandatory parameter, and the possible values are:

  • start_job_instance
  • stop_scenario_instance
  • start_job_instance
  • stop_scenario_instance
  • if
  • while
  • push_file

Wait_launched, wait_finished and wait_delay are used for the scheduling conditions (as in the web interface case):

  • wait_launched: a list of other OpenBACH functions that must be completed before this one starts (defaults to None);
  • wait_finished: a list of other start_job_instance or start_scenario_instance OpenBACH functions; each job/scenario started by these functions must reach completion before this function starts (defaults to None);
  • wait_delay: an amount of time, in seconds, to wait with respect to the wait_launched/wait_finished functions, before executing its action (defaults to 0.0).

Configuration

In order to configure an OpenBACH function, you need to call its configure method. Each type of function has its own configure parameters:

  • If start_job_instance:
name_of_function.configure(job_name, entity_name, offset=0, **parameters_of_job)

This function will start the OpenBACH job job_name on the agent associated to the entity entity_name offset seconds after the OpenBACH function is scheduled. The job will be configured with the values of the parameters provided in the parameters_of_job keyword arguments.

Note that when configuring a job whose accept sub-commands, you need to provide a dictionary to be able to configure the parameters of the sub-command. Example: iperf3_srv.configure('iperf3', srv_entity, offset=0, interval=1.0, server={'exit':True, 'bind':srv_ip}). If you want to select a sub-command without parameters, provide an empty dictionary to the sub-command choice.

  • If start_scenario_instance:
name_of_function.configure(scenario_name, **arguments_of_scenario)

This function will start the OpenBACH scenario scenario_name when it is scheduled. Extra arguments provided as keyword arguments will be passed to the launched scenario.

scenario_name can be a string to start an existing scenario in your OpenBACH project; or a Scenario instance to configure a sub-scenario alongside the one holding this OpenBACH function.

Note that using a Scenario instance will allow you to navigate through them all using the Scenario.subscenarios iterator on your main scenario. This could help building them all at once, for instance.

  • If stop_job_instance:
name_of_function.configure(*functions_to_stop)

This function will stop the provided OpenBACH functions as soon as it is scheduled; granted they are of type start_job_instance.

  • If stop_scenario_instance:
name_of_function.configure(scenario_to_stop)

This function will stop the provided OpenBACH function as soon as it is scheduled; granted it is of type start_scenario_instance.

  • If if:
name_of_function.configure(condition)
name_of_function.configure_if_true(*openbach_functions)
name_of_function.configure_if_false(*openbach_functions)

This function will evaluate the condition when it is scheduled and launch one or the others configured OpenBACH functions depending on the result of this evaluation.

Note that only the configure call is mandatory.

  • If while:
name_of_function.configure(condition)
name_of_function.configure_while_body(*openbach_functions)
name_of_function.configure_while_end(*openbach_functions)

As for the if OpenBACH function, only the configure call is mandatory. As expected, the OpenBACH functions of the body will be executed as long as the condition holds true; the end functions will always be evaluated once the condition turns false.

  • If push_file:
name_of_function.configure(entity_name, path_on_the_controller, path_on_the_agent, users=(), groups=())

When this OpenBACH function is scheduled, the file referenced by path_on_the_controller will be copied to path_on_the_agent on the agent associated to the entity entity_name. You can provide a single string for each parameters or a list of strings to copy several files at once. The length of both lists must match.

If provided, the users and groups parameters will control ownership of the files on the agent; their length must also match with the other parameters.

Example

This example creates and returns a scenario containing one argument, ip_dst, and three functions:

  • hping_function, that starts the job hping to ping $ip_dst
  • fping_function, that starts the job fping to ping $ip_dst
  • stop_function, that stops hping_function and fping_function 20s after they started
example.py
#Import the Scenario class from scenario_builder and the ScenarioObserver from the auditorium-scripts.
from scenario_builder import Scenario
 
def build_delay_scenario(client, server, scenario_name):
    scenario = Scenario(scenario_name, 'Comparison of 2 types of RTT measurements')
    scenario.add_argument('ip_dst', 'Target of the pings and server ip adress')
 
    hping_function = scenario.add_function('start_job_instance')
    hping_function.configure('hping', client, offset=0, destination_ip='$ip_dst')
    fping_function = scenario.add_function('start_job_instance')
    fping_function.configure('fping', client, offset=0, destination_ip='$ip_dst')
    stop_function = scenario.add_function('stop_job_instance', wait_launched=[hping_function, fping_function], wait_delay=20)
    stop_function.configure(hping_function, fping_function)
    return scenario
openbach/manuals/2.x/developer_manual/openbach_api/scenario_builder/index.txt · Last modified: 2020/11/05 16:44 by mettinger