Hello everyone, We're happy to share that the Slurm agent v1 (with

PythonFunctionTask

) has been implemented. It supports the following three core methods: 1. `create`: Use

srun

to run a Slurm job which executes Flyte entrypoints,

pyflyte-fast-execute

and

pyflyte-execute

2. `get`: Use

scontrol show job <job_id>

to monitor the Slurm job state 3. `delete`: Use

scancel <job_id>

to cancel the Slurm job (this method is still under test) We setup an environment to test it locally without running the agent gRPC server. The setup is divided into three components: a client (localhost), a remote tiny Slurm cluster, and an Amazon S3 bucket that facilitates communication between the two. The attached figure below illustrates the interaction between the client and the remote Slurm cluster.

@creamy-shampoo-53278 great job But @damp-lion-88352 we also wanted a simple agent that does not do pyflyte execute and is more like a script runner?

Hi Ketan, For faster development and testing on a remote Slurm cluster, I setup the controller and compute node on the same machine, following Slurm-101 instructions. As for the configuration, I use this official Slurm configurator to generate the config file!

When I say remote, can the user add the remote slurm host in the config

I see. Currently, I hard code ssh config in agent.py . I’ll add another ssh_config field in task_config to let users set ssh-related info!

That’s what we want to avoid

@task(
    task_config=Slurm(
        ssh_conf={
            "host": "<ssh_host>",
            "port": "<ssh_port>",
            "username": "<ssh_username>",
            "password": "<ssh_password>",
        },
        srun_conf={
            "partition": "debug",
            "job-name": "demo-slurm",
            # Remote working directory
            "chdir": "<your_remote_working_dir>"
        }
    )
)
def plus_one(x: int) -> int:
    return x + 1

It's now possible to define a task by passing in

ssh_conf

!

yes