> For the complete documentation index, see [llms.txt](https://asus-isg-aidc.gitbook.io/guide/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://asus-isg-aidc.gitbook.io/guide/latest/portal/overview/service-catalog/slurm.md).

# Slurm

| Developer | Last modified |
| --------- | ------------- |
| AIDC Team | 2026/08/21    |

## Service Introduction

**Slurm (Simple Linux Utility for Resource Management)** is an open-source workload manager designed for Linux clusters. It provides resource management, job scheduling, and job monitoring capabilities, widely used in High Performance Computing (HPC) environments.

![](https://pub-f334ff01208c4e6195b80133ac6e6030.r2.dev/portal/marketplace/service-catalog-slurm.png)

### Quick Summary

| Item                          | Details                                           |
| ----------------------------- | ------------------------------------------------- |
| ⏱️ **Estimated Time**         | 10-30 minutes (depends on node count)             |
| 💻 **Supported Platforms**    | See the **Supported Platforms Matrix**            |
| 🔧 **Minimum Nodes Required** | 3 nodes (Headnode × 1, Compute × 1, Database × 1) |
| 📦 **Included Components**    | Slurm Controller, Slurmd, Slurmdbd, MariaDB       |

***

## Prerequisites

Before starting the deployment, ensure the following requirements are met:

### Hardware Requirements

| Role              | Minimum Spec                  | Recommended Spec               |
| ----------------- | ----------------------------- | ------------------------------ |
| **Headnode**      | 4 vCPU / 8GB RAM / 50GB Disk  | 8 vCPU / 16GB RAM / 100GB Disk |
| **Compute Node**  | 4 vCPU / 8GB RAM / 50GB Disk  | Depends on workload            |
| **Database Node** | 2 vCPU / 4GB RAM / 100GB Disk | 4 vCPU / 8GB RAM / 200GB Disk  |

### Network Requirements

* ✅ All nodes must be in the **same network segment** or have routing enabled
* ✅ Firewall rules must allow inter-node communication

### Other Requirements

* ✅ Nodes have been provisioned and are in **Ready** state
* ✅ SSH access between nodes is configured

***

## Deployment Dialog Introduction

Slurm Cluster deployment is divided into **3 steps**:

```mermaid
flowchart LR
    Step1["Step 1: Headnode / Compute Node"]
    Step2["Step 2: Database Configuration"]
    Review["Step 3: Review and Deploy"]

    Step1 --> Step2
    Step2 --> Review
```

[Step 1: Headnode/Compute Node](#step-1-headnodecompute-node-configuration) | [Step 2: Database Configuration](#step-2-database-configuration) | [Step 3: Review and Deploy](#step-3-review-and-deploy)

***

### Step 1: Headnode/Compute Node Configuration

![](https://pub-f334ff01208c4e6195b80133ac6e6030.r2.dev/portal/marketplace/service-catalog-slurm-step1.png)

| Parameter Name     | Type   | Description                   |   Required  | Validation Rules                                                                    | Example         |
| ------------------ | ------ | ----------------------------- | :---------: | ----------------------------------------------------------------------------------- | --------------- |
| **Headnode Group** | string | Group for Slurm control nodes | 🔴 Required | **The selected group must contain exactly one node, Cannot duplicate other Groups** | `slurm-head`    |
| **Compute Group**  | string | Group for Slurm compute nodes | 🔴 Required | **Cannot duplicate other Groups**                                                   | `slurm-compute` |

***

### Step 2: Database Configuration

![](https://pub-f334ff01208c4e6195b80133ac6e6030.r2.dev/portal/marketplace/service-catalog-slurm-step2.png)

| Parameter Name    | Type   | Description                            |   Required  | Validation Rules                               | Example       |
| ----------------- | ------ | -------------------------------------- | :---------: | ---------------------------------------------- | ------------- |
| **DB Node Group** | string | Group for Slurm database nodes         | 🔴 Required | **Cannot duplicate other Groups**              | `slurm-db`    |
| **DB Username**   | string | Database username for Slurm accounting | 🔴 Required | **At least 6 characters**                      | `slurm_admin` |
| **DB Password**   | string | Database password                      | 🔴 Required | **At least 6 characters; use strong password** | `Slurm@2026!` |

{% hint style="warning" %}
**⚠️ Critical Constraint**

To ensure cluster stability, the three roles (**Headnode**, **Compute**, **Database**) **must be deployed on different Groups**.

The system will automatically disable Groups that have already been selected to prevent configuration errors.
{% endhint %}

***

### Step 3: Review and Deploy

![](https://pub-f334ff01208c4e6195b80133ac6e6030.r2.dev/portal/marketplace/service-catalog-slurm-step3.png)

***

## Post-Deployment Status

After successful deployment, you can view in **Status Monitor**:

* Task status progress
* Execution status of each node
* Error logs (if any)

### Deployment Verification

After the deployment status shows **Completed**, perform the following checks to confirm the service is running correctly:

**1. Check Cluster Status**

```bash
# SSH into the Headnode and run:
sinfo
```

✅ **Expected Result:** All nodes should appear with state `idle` or `alloc`

```
PARTITION   AVAIL  TIMELIMIT  NODES  STATE NODELIST
compute*    up     infinite   4      idle  node[01-04]
```

**2. Check Controller Status**

```bash
scontrol show node
```

✅ **Expected Result:** it should return compute node status

**3. Check Slurm DB**

```bash
sacctmgr list cluster
```

✅ **Expected Result:** displays information about clusters that are registered in the Slurm accounting database. it will show Cluster names that have been registered in the

***

## Frequently Asked Questions

### Q: Why can't I select certain Groups?

A: The system automatically disables Groups that have been selected by other roles to ensure the three roles are in different Groups.

### Q: What are the restrictions for DB Username and Password?

A: Both must be at least 6 characters. Strong passwords with mixed case, numbers, and special characters are recommended.

### Q: How long does deployment take?

A: Depends on the number of nodes and network conditions, typically 10-30 minutes.

### Q: What if deployment fails midway?

A: Check the **Status Monitor** for detailed error logs. You can retry the deployment after resolving the issues.

***

## Troubleshooting Checklist

If the cluster is not functioning as expected, use this checklist:

| Check Item               | Command                      | Expected Result  |
| ------------------------ | ---------------------------- | ---------------- |
| Controller running?      | `systemctl status slurmctld` | Active (running) |
| Compute daemon running?  | `systemctl status slurmd`    | Active (running) |
| Database daemon running? | `systemctl status slurmdbd`  | Active (running) |
| Munge running?           | `systemctl status munge`     | Active (running) |
| Node visible?            | `sinfo`                      | Nodes listed     |

***

## Related References

* [Service Catalog](/guide/latest/portal/overview/service-catalog.md)
* [Status Monitor User Guide](/guide/latest/portal/overview/status-monitor.md#slurm-status-card)
* [Slurm Official Documentation](https://slurm.schedmd.com/documentation.html)