Cardano Stake Pool Setup
Cardano stake pool operators require two nodes — a relay node and a producer node. These are separate resources with different deployment methods and different provisioning timelines. This page explains the full operational model before you purchase anything.
The two-node architecture
| Node | Purpose | Deployment method |
|---|---|---|
| Relay node | Connects to the Cardano P2P network; receives and forwards blocks | CLI (cpctl) or server-type (SSH) |
| Producer node | Mints blocks; must not be publicly reachable | Server-type (SSH) only |
The producer node must be deployed via SSH onto a dedicated server. Cardano's security model requires block-signing keys to reside on a network-isolated machine under the operator's direct control — containerised or orchestrated deployment is not compatible with this requirement.
What happens after you purchase
Both the relay node (server-type path) and the producer node require a dedicated server to be provisioned before you can deploy. This is not instantaneous.
- Complete your order for a server-type compute resource at computeportal.io/start.
- CP provisions your dedicated server. This takes up to 24 business hours (Monday–Friday, excluding public holidays).
- You receive an email notification when the server is ready.
- SSH into the server and deploy your node software, keys, and firewall rules.
Nothing is broken during the provisioning window
You will not be able to SSH into your server until the email arrives. This is expected — provisioning is not instantaneous. Contacting support before the email arrives will not speed up the process.
If you are deploying your relay node via CLI (cpctl) rather than server-type, the provisioning wait does not apply — the relay deploys to an existing shared cluster and is available immediately after purchase.
Recommended order of operations
- Purchase your relay resource and deploy it first. CLI-path relay nodes are live immediately; server-type relay nodes require the provisioning wait.
- Purchase your producer resource and wait for the provisioning email before attempting to SSH in.
- Once both nodes are running, connect them: point your producer's
topology.jsonat your relay, and expose your relay's TCP port so the producer can reach it.
Deploying the relay first means you can register it on-chain and begin peering with the network while the producer server is being provisioned.
Provider portability
Your producer node is installed directly on a specific physical server — not in a container or cluster that CP manages. If the underlying infrastructure provider changes, you must:
- Repurchase a server-type resource
- Wait through the provisioning window again
- Redeploy and reconfigure all node software, keys, and firewall rules on the new server
This is different from all other CP workloads (websites, APIs, AI agents), which are portable and can be migrated transparently. Plan accordingly if long-term provider continuity matters to your pool operation.
Next steps
- Deploy your relay node → — full CLI walkthrough: ConfigMap upload, Mithril bootstrap, volume attach, TCP expose for the producer
- Quick Start → — if you are new to
cpctl - Support — computeportal.io