Deploy iQ Studio on a GPU‑enabled Kubernetes cluster. It is strongly recommended to deploy iQ Studio on a multi‑node Kubernetes cluster with GPU support rather than on a single‑node cluster. A single node cannot realistically support the required pod density, particularly for GPU workloads. A multi‑node cluster provides scalability and stability and eliminates the need for complex system‑level hardening.
Your environment must meet the installation requirements for iQ Studio, so review these requirements before continuing with the installation.
Kubernetes requirements
| Component | Minimum version | Notes |
|---|---|---|
| Operating system |
Ubuntu 24.04 LTS RHEL 9.5 |
Ubuntu or RHEL is the required operating system. |
| Kubernetes | 1.32.10 | Requires CSI‑compatible storage provider. |
| etcd | 3.5.20 | Kubernetes cluster state store. |
| Container runtime | containerd | Recommended for stable GPU workloads. |
| Container Network Interface | Calico | Preferred for multi-namespace networking and isolation. |
| NVIDIA GPU Operator | 25.10.1 | Supports GPU provisioning and monitoring. |