Manage Jobs

This guide provides an overview of how to view and list Slurm jobs on your cluster.

List Jobs

Use the fluidctl slurm clusters jobs list command to list jobs on a cluster:

fluidctl slurm clusters jobs list --region <region> --project <project-id> --cluster-id <cluster-id>

Filter Options:

--state - Filter jobs by state (e.g., RUNNING, PENDING, COMPLETED, FAILED)
--user - Filter jobs by username
--partition - Filter jobs by partition
--limit - Limit the number of results (default: no limit)

Examples:

List all running jobs:

fluidctl slurm clusters jobs list --region <region> --project <project-id> --cluster-id <cluster-id> --state RUNNING

List jobs for a specific user with limit:

fluidctl slurm clusters jobs list --region <region> --project <project-id> --cluster-id <cluster-id> --user myuser --limit 10

To list jobs via the API, send a GET request to the /slurm/clusters/{cluster_id}/proxy/slurm/v0.0.41/jobs endpoint:

GET /slurm/clusters/<cluster-id>/proxy/slurm/v0.0.41/jobs HTTP/1.1
Host: <region>.atlas.fluidstack.io
Authorization: Bearer <token>

Use the fluidctl slurm clusters jobs get command to get details of a specific job:

fluidctl slurm clusters jobs get <job-id> --region <region> --project <project-id> --cluster-id <cluster-id>

Example:

fluidctl slurm clusters jobs get 12345 --region <region> --project <project-id> --cluster-id <cluster-id>

To get job details via the API, send a GET request to the /slurm/clusters/{cluster_id}/proxy/slurm/v0.0.41/job/{job_id} endpoint:

GET /slurm/clusters/<cluster-id>/proxy/slurm/v0.0.41/job/<job-id> HTTP/1.1
Host: <region>.atlas.fluidstack.io
Authorization: Bearer <token>

Jobs can be in one of the following states:

State	Description
`PENDING`	Job is waiting to be scheduled
`RUNNING`	Job is currently executing
`COMPLETING`	Job is in the process of completing
`COMPLETED`	Job has finished successfully
`FAILED`	Job has terminated with an error
`TIMEOUT`	Job was terminated due to time limit
`CANCELLED`	Job was cancelled by user or administrator