Celestia
- installation guide
- bridge node setup
- full node setup
- operating boundaries
- multiplexer
- light node setup
- monitoring
- keys
- validator signaling
- blobstream orchestrator
- storage optimization
- snapshots
- decentralize map
- one liner
- skill
- contributions
Celestia Full Node Terminology and Legacy DA Role Retirement
There is no current full role in celestia-node. The supported data
availability roles are bridge and light. Older guides used “full storage
node” for a retired DA role; do not use those obsolete commands with current
software.
Two different meanings
Consensus full node
A consensus full node runs celestia-appd from the celestia-app repository. It validates the consensus chain, executes application state, and can provide RPC and gRPC services. A validator is a consensus full node with an active consensus signer.
Current celestia-app mainnet reference:
- Version:
v9.0.6 - Commit:
6f4b596e47f80683adb1a161ca7cb640dcd9d206 - Chain ID:
celestia
| Consensus profile | CPU | Memory | NVMe | Network |
|---|---|---|---|---|
| Validator or non-archival consensus node | 32 cores | 32 GB | 12 TiB | 1 Gbps |
| Archival consensus node | 32 cores | 64 GB | 624 TiB | 1 Gbps |
The non-archival estimate uses a conservative 7-day planning window; archival uses one year. Both use the 128 MB per 6 seconds maximum-throughput envelope. Actual consumption may be lower. Validators must pass the official CPU benchmark; prefer GFNI and SHA-NI support.
A bridge node's initial sync needs an archival consensus gRPC source. Running a consensus full node does not automatically make it archival: block retention and application pruning must be configured and verified deliberately.
Data availability node
A DA node runs celestia-node. Choose one of the current roles:
| Need | Current role | Guide |
|---|---|---|
| Import, erasure-code, retain, and serve blocks | Bridge | Bridge node setup |
| Verify availability through DAS with low resources | Light | Light node setup |
Do not relabel an old DA “full” store as a bridge store. Reinitialize the chosen current role and apply the correct key-custody and network checks.
Migration checklist for a legacy DA deployment
- Record the installed binary version, configured network, store path, peer ID, retention intent, and upstream consensus source without exposing key data.
- Decide whether the workload requires bridge retention/serving or light DAS.
- Provision the current role in a separate store using its current guide.
- Keep mainnet RPC
26658on loopback and expose bridge P2P2121over TCP and UDP when choosing the bridge role. - Prove new-role sync, peer connectivity, storage headroom, and independent height freshness before retiring the legacy process.
- Handle any identity or wallet transfer only through a separately approved key procedure.
This page intentionally contains no legacy full-role command examples, reset steps, key deletion, or transaction actions.
Sources
Evidence reviewed from the official Celestia docs repository at commit
8fbaa868a323c13d3edae2875d9b27765eb29c45:
operate/getting-started/hardware-requirementsoperate/getting-started/overviewoperate/consensus-validators/consensus-nodeoperate/data-availability/bridge-nodeoperate/data-availability/light-node/quickstart