Jobs — Job Management
The Jobs section of Open OnDemand provides two complementary tools to manage your computations on the MCMeSU cluster:
- Active Jobs — Monitor the status of your Slurm jobs in real time
- Job Composer — Create, edit, and submit Slurm job scripts through a graphical interface
Both tools are accessible from the Jobs menu in the top navigation bar.
Active Jobs
Active Jobs gives you a real-time overview of all the jobs currently known to Slurm — whether they are pending, running, or recently completed.
Interface Overview
The Active Jobs table displays the following information for each job:
| Column | Description |
|---|---|
| Job ID | Unique Slurm identifier for the job |
| Job Name | Name defined in your script via #SBATCH --job-name |
| User | Username that submitted the job |
| Account | Slurm account charged for the job |
| Time Used | Elapsed time since the job started |
| Queue | Partition the job is running or waiting on |
| Cluster | The cluster the jobs run on (Only MCMeSU for now) |
| Action | If pending, the reason why the job is not yet running |
Filtering and Sorting
By default, Active Jobs displays all jobs across all users visible to Slurm. You can:
- Sort by any column by clicking its header
- Toggle between showing only your jobs or all cluster jobs using the Your Jobs / All Jobs switch
Deleting a Job
To cancel a running or pending job:
- Locate the job in the table
- Click the Delete button (trash icon) on the right side of the row
- Confirm the cancellation
This is equivalent to running scancel <jobid> from the terminal.

Active Jobs auto-refreshes periodically. If you need an immediate update, click the Refresh button at the top of the table.
Job Composer
The Job Composer allows you to create and submit Slurm job scripts entirely through a graphical interface — no terminal required. It also keeps a history of all jobs you have submitted through it, making it easy to reuse or tweak previous submissions.
Interface Overview

The Job Composer interface is split into two panels:
- Left panel: the list of your jobs (your job history)
- Right panel: the details of the selected job, including the script editor, job metadata, and submission controls
Creating a New Job

- Click the New Job button at the top left
- You can select one option :
- From Default Template to start from a basic Slurm script skeleton.
- From Template shows basics default template, used in the different partitions.
- From selected path let you use your own template.
- From selected job duplicate the job selected in the job list
A new job entry will appear in the left panel, and its script will be loaded in the editor on the right.
Editing the Job Script
The script editor allows you to modify your job script directly in the browser.
Job Folder
Each job in the Job Composer has its own working directory on the cluster, created automatically under:
$HOME/ondemand/data/sys/myjobs/projects/default/<job_id>/
This is the directory from which your job will be submitted and executed. You can:
- Add input files to this directory via the Open Dir button, which opens it directly in the File Manager
- Place any data or configuration files your script depends on here before submitting
If your job requires large input datasets, do not copy them into the job folder in $HOME. Instead, reference them from $SCRATCH directly in your script using their full path.
Submitting a Job
Once your script is ready:
- Make sure the correct job is selected in the left panel
- Click the Submit button at the top of the right panel
- The job will be handed off to Slurm — its status will update to
PENDINGorRUNNINGshortly after
You can monitor it immediately in Active Jobs.
Managing Existing Jobs

The left panel lists all jobs you have created through the Job Composer. For each entry you can:
| Action | Description |
|---|---|
| Edit Files | Open the job folder in the File Manager |
| Open Terminal | Open a terminal in the job folder |
| Job Options | Edit job metadata (name, working directory) |
| Delete | Remove the job entry and its associated folder |
Job Templates
Templates are a core feature of the Job Composer. They allow you to save a job script — along with any associated files — as a reusable base for future submissions. Instead of rewriting your #SBATCH directives and module setup every time, you start from a template and only adjust what needs to change.
There are two types of templates available in the Job Composer:
| Type | Description |
|---|---|
| System templates | Provided by the MCMeSU team. Available to all users. |
| User templates | Created by you from any of your existing jobs. Personal and editable. |
System Templates
System templates are provided and maintained by the MCMeSU team. They appear in the template list for all users and are designed to cover the most common use cases on the cluster.
To use a system template:
- Click New Job in the Job Composer
- Select From Template
- A list of available templates will appear — system templates are listed alongside your personal ones
- Select the desired template and click Create New Job
A new job entry is created in your job list, pre-populated with the template's script and any associated files. You can then edit the script freely — your copy is independent from the original template.
System templates are a good starting point even for advanced users. They reflect the recommended practices for MCMeSU and include the correct partition names, default resource values, and module loading patterns.
Creating a User Template
You can save any of your existing jobs as a personal template to reuse it later.
- In the Job Composer, select the job you want to save as a template from the left panel
- Click Create Template in the toolbar
- Give your template a meaningful name and an optional description
- Click Save

Your template is now available in the template list whenever you create a new job.
Managing User Templates
All the templates are accessible from the Templates view, reachable via Templates in the navigation bar.

From this view, you can:
| Action | Description |
|---|---|
| New Tempalte | Create a new template |
| Copy Template | Copy the selected template |
| View Files | Inspect the template's associated files |
| Open Terminal | Open the terminal on the location of the template |
| Edit Files | Open the template folder in the File Manager to add, remove or modify files included in the template |
| Delete | Permanently remove the template |
Deleting a template does not affect jobs that were previously created from it. Each job has its own independent copy of the script.
What a Template Contains
A template is not just a script — it is a directory snapshot. When you create a job from a template, the entire contents of the template folder are copied into the new job's working directory.
This means you can include in a template:
- The main job script (
main_job.shor similar) - Secondary scripts called by the main script
- Configuration files required by your application
- Small reference files or parameter files