MFLib

MFLib is the main class that is used to interact with the Measurement Framework set up in a user’s FABRIC Experiment.

class mflib.mflib.MFLib(slice='', local_storage_directory='/tmp/mflib', mf_repo_branch='main', optimize_repos=False)

Bases: Core

MFLib allows for adding and controlling the MeasurementFramework in a Fabric experiementers slice.

static addMeasNode(slice, cores=4, ram=16, disk=500, site='EDC', image='docker_ubuntu_20')

Adds Measurement node and measurement network to an unsubmitted slice object.

Parameters:

  • slice (fablib.slice) – Slice object already set with experiment topology.

  • cores (int, optional) – Cores for measurement node. Defaults to 4 cores.

  • ram (int, optional) – _description_. Defaults to 16 GB ram.

  • disk (int, optional) – _description_. Defaults to 500 GB disk.

  • network_type (string, optional) – _description_. Defaults to FABNetv4.

  • site (string, optional) – _description_. Defaults to NCSA.

add_mflib_log_handler(log_handler)

Adds the given log handler to the mflib_logger. Note log handler needs to be created with set_mflib_logger first.

Parameters:

log_handler (logging handler) – Log handler to add to the mflib_logger.

download_common_hosts()

Downloads hosts.ini file and returns file text. Downloaded hosts.ini file will be stored locally for future reference.

init(slice, optimize_repos)

Sets up the slice to ensure it can be monitored. Sets up basic software on Measurement Node and experiment nodes. Slice must already have a Measurement Node. See log file for details of init output.

Parameters:

slice_name (str) – The name of the slice to be monitored.

Returns:

False if no Measure Node found or a init process fails. True otherwise.

Return type:

Bool

instrumentize(services=['prometheus', 'elk'])

Instrumentize the slice. This is a convenience method that sets up & starts the monitoring of the slice. Sets up Prometheus, ELK & Grafana.

Parameters:

services (List of Strings) – Just add the listed components. Options are elk or prometheus.

Returns:

The output from each phase of instrumetizing.

Return type:

dict

mflib_class_version = '1.0.42'
remove_mflib_log_handler(log_handler)

Removes the given log handler from the mflib_logger.

Parameters:

log_handler (logging handler) – Log handler to remove from mflib_logger

set_mflib_logger()

Sets up the mflib logging file. The filename is created from the self.logging_filename. Note that the self.logging_filename will be set with the slice when the slice name is set.

This method uses the logging filename inherited from Core.