schluenz / pyslurm

Python Interface to Slurm

Home Page:https://pyslurm.github.io

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

PySlurm

PySlurm

Overview

PySlurm is the Python client library for the Slurm HPC Scheduler.

Prerequisites

  • Slurm - Slurm shared library and header files
  • Python - >= 3.6
  • Cython - >= 0.29.30 but < 3.0

This PySlurm branch is for the Slurm Major-Release 22.05

Installation

By default, it is searched inside /usr/include for the Header files and in /usr/lib64 for Slurms shared-library (libslurm.so) during Installation. For Slurm installations in different locations, you will need to provide the corresponding paths to the necessary files.

You can specify these Paths with environment variables, for example:

export SLURM_INCLUDE_DIR=/opt/slurm/22.05/include
export SLURM_LIB_DIR=/opt/slurm/22.05/lib

Then you can proceed to install PySlurm, for example:

pip install pyslurm==22.05.0

Or by cloning the repository:

git clone https://github.com/PySlurm/pyslurm.git && cd pyslurm
python setup.py install

# Or simply with pip
pip install .

Also see python setup.py --help

Release Versioning

PySlurm's versioning scheme follows the official Slurm versioning. The first two numbers (MAJOR.MINOR) always correspond to Slurms Major-Release, for example 22.05. The last number (MICRO) is however not tied in any way to Slurms MICRO version. For example, any PySlurm 22.05.X version should work with any Slurm 22.05.X release.

Documentation

The API documentation is hosted at https://pyslurm.github.io.

To build the docs locally, use Sphinx to generate the documentation from the reStructuredText based docstrings found in the pyslurm module once it is built:

cd doc
make clean
make html

Testing

PySlurm requires an installation of Slurm.

Using a Test Container

To run tests locally without an existing Slurm cluster, docker and docker-compose is required.

Clone the project:

git clone https://github.com/PySlurm/pyslurm.git
cd pyslurm

Start the Slurm container in the background:

docker-compose up -d

The cluster takes a few seconds to start all the required Slurm services. Tail the logs:

docker-compose logs -f

When the cluster is ready, you will see the following log message:

Cluster is now available

Press CTRL+C to stop tailing the logs. Slurm is now running in a container in detached mode. docker-compose also bind mounds the git directory inside the container at /pyslurm so that the container has access to the test cases.

Install test dependencies:

pipenv sync --dev

Execute the tests inside the container:

pipenv run pytest -sv scripts/run_tests_in_container.py

When testing is complete, stop the running Slurm container:

docker-compose down

Testing on an Existing Slurm Cluster

You may also choose to clone the project and run tests on a node where Slurm is already compiled and installed:

git clone https://github.com/PySlurm/pyslurm.git
cd pyslurm
pip install .
./scripts/configure.sh
pipenv sync --dev
pipenv run pytest -sv

Contributors

PySlurm is made by contributors like you.

Help

Ask questions on the PySlurm Google Group

About

Python Interface to Slurm

https://pyslurm.github.io

License:GNU General Public License v2.0


Languages

Language:Cython 89.7%Language:Python 9.5%Language:Shell 0.8%Language:C 0.0%