> 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/weka.md).

# WEKA

| Developer | Last modified |
| --------- | ------------- |
| AIDC Team | 2026/01/30    |

## Service Introduction

> **WEKA** is a high-performance parallel file system designed for AI, machine learning, and high-performance computing workloads. It provides extremely low latency and high throughput data access.
>
> WEKA Features:
>
> * **High Performance**: Provides GB/s-level throughput and microsecond-level latency
> * **Scalable**: Linearly scales to hundreds of nodes
> * **Reliability**: Built-in data protection, supports node failure tolerance
> * **Easy Management**: No need for complex RAID, Volume, and other configurations

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

### Quick Summary

| Item                               | Details                                     |
| ---------------------------------- | ------------------------------------------- |
| ⏱️ **Estimated Time**              | 30-60 minutes (depends on cluster size)     |
| 💻 **Supported Deployment Method** | WEKA OS–based deployment only               |
| 🔧 **Minimum Nodes Required**      | 6 nodes (for data protection)               |
| 📦 **Included Components**         | WEKA Server (cluster deployment on WEKA OS) |

***

## Prerequisites

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

### Hardware Requirements

| Component         | Requirement / Recommendation                                       | Notes                                                                                                    |
| ----------------- | ------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------- |
| **Cluster Size**  | ≥ 6                                                                | Minimum for production-ready cluster                                                                     |
| **CPU**           | Supported architectures: Intel Xeon, AMD EPYC, ARM (AArch64)       | Core count and allocation must be calculated according to cluster design and WEKA official guidance      |
| **Memory (RAM)**  | Sufficient to support WEKA processes, metadata, and SSD management | Exact size depends on cluster size and SSD configuration; follow WEKA official guidance                  |
| **Storage (SSD)** | Local SSDs only (NVMe/SAS/SATA)                                    | Capacity, redundancy, and hot spare must be planned based on cluster design and official recommendations |
| **Network**       | InfiniBand                                                         | Currently, cluster nodes only support deployment over InfiniBand                                         |

### Network Notes

* Required ports: `14000-14100` (WEKA communication)

### SSD Notes

* Maintain a capacity ratio of 8:1 or less between the smallest and largest SSDs in the cluster.
* Maintain a ratio of `8000:1` or less between the total SSD capacity and the total RAM of the cluster.

### Other Notes

* WEKA ISO image accessible

***

## Deployment Dialog Introduction

WEKA deployment is divided into **2 steps**:

```mermaid
flowchart LR
    Step1["Step 1: Node & ISO Configuration"]
    Review["Step 2: Review and Deploy"]

    Step1 --> Review
```

[Step 1: Basic Configuration](#step-1-basic-configuration) | [Step 2: Review and Deploy](#step-2-review-and-deploy)

***

## Node Selection

### Step 1: Basic Configuration

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

| Parameter Name              | Type   | Description                        |    Required   | Validation Rules                               | Example          |
| --------------------------- | ------ | ---------------------------------- | :-----------: | ---------------------------------------------- | ---------------- |
| **WEKA Installation Group** | string | Target Group for WEKA installation | ⚪ Conditional | **Either Group or Node required; min 6 nodes** | `weka-cluster`   |
| **WEKA Installation Node**  | string | Target Node for WEKA installation  | ⚪ Conditional | **Either Group or Node required**              | `weka-node-01`   |
| **ISO Name**                | string | WEKA ISO image file name           |  🔴 Required  | **Valid ISO file path**                        | `weka-4.2.1.iso` |

{% hint style="warning" %}
**⚠️ Minimum Node Requirement**

WEKA requires deployment on **at least 6 nodes** to ensure data protection and cluster stability.

* Nodes should have high-speed network connections (InfiniBand or 100 GbE recommended)
* NVMe SSDs are required as the storage backend
* All nodes must meet the hardware specifications
  {% endhint %}

***

### Step 2: Review and Deploy

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

***

## Post-Deployment Status

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

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

### Deployment Verification

After the deployment status shows **Completed**, perform the following checks:

**1. Check Cluster Status**

```bash
weka status
```

✅ **Expected Result:** Cluster shows as `OK` with all nodes healthy

**2. Check Cluster Details**

```bash
weka cluster
```

✅ **Expected Result:** All nodes listed with `UP` status

***

## Troubleshooting Checklist

| Check Item           | Command                | Expected Result |
| -------------------- | ---------------------- | --------------- |
| Network healthy?     | `weka cluster network` | No errors       |
| File system mounted? | `mount \| grep wekafs` | Mount listed    |

***

## Frequently Asked Questions

### Q: What's the minimum number of nodes required?

A: At least 6 nodes are required to provide basic data protection with N+2 redundancy.

### Q: What network environment does AIDC support for WEKA cluster deployment?

A: Currently, only deployment over InfiniBand is supported.

### Q: What workloads is WEKA suitable for?

A: Particularly suitable for scenarios requiring high IOPS and low latency, such as AI/ML training, HPC simulations, genomics, and big data analytics.

### Q: What happens if a node fails?

A: WEKA automatically rebuilds data within the configured disk failure tolerance.

***

## Related References

* [Service Catalog](/guide/latest/portal/overview/service-catalog.md)
* [Status Monitor User Guide](/guide/latest/portal/overview/status-monitor.md#weka-status-card)
* [WEKA Official Documentation](https://docs.weka.io/)