Documentation Index
Fetch the complete documentation index at: https://docs.fuse.io/llms.txt
Use this file to discover all available pages before exploring further.
Upgrading Your Fuse Node to Nethermind v1.36.2
Fuse nodes are being upgraded from Nethermind v1.29.0 to v1.36.2 — the biggest node upgrade the network has seen in a while. The new version brings meaningfully faster block processing, lower resource usage, built-in monitoring, and smarter disk management. It also brings Fuse in line with the Fusaka Ethereum fork and the broader improvements the Ethereum ecosystem has been building toward.
The upgrade has been running cleanly on Spark testnet since rollout, with state sync completing in under two hours. Validators should follow the steps below when rolling out to mainnet.
What’s New
Faster block processing Block processing is significantly quicker across the board. RPC responses are snappier, and the node handles high traffic more efficiently under load. Under the hood, parallelised in-memory pruning no longer blocks the block processing thread, and new payload processing now averages under 20ms end-to-end.
Built-in monitoring dashboard Node operators now get a real-time interface to track sync status, logs, and network activity without any additional setup. You can also configure extended monitoring via Grafana and Prometheus.
Smarter disk management New history pruning tools keep storage growth under control at runtime, reducing disk usage by approximately 30%. That makes it cheaper and simpler to run a node long-term.
Expanded developer tooling This release includes eth_simulate for multi-block transaction simulation, and Era1 import/export as groundwork for Ethereum’s long-term data management standards (EIP-4444).
Improved stability DB corruption now triggers a clean exit instead of a hard crash. Memory regressions have been fixed, and numerous sync and RPC reliability improvements have landed across the release cycle.
\
Before You Begin
Note: The upgrade instructions below apply to nodes launched using the official quickstart.sh script. If you set up your node manually, refer to the Nethermind documentation for equivalent steps.
System requirements: Review the minimum hardware requirements and required disk speed before proceeding.
Database resync required: Starting with Nethermind v1.33.0, the internal database structure changed. Any node running a version prior to v1.33.0 — which includes all nodes currently on Fuse mainnet — must perform a full resync. Thanks to improvements to Nethermind’s state sync, this process typically takes 2–6 hours depending on hardware. To speed it up, you can use the DB snapshot instead of syncing from scratch.
Upgrade Steps
Important:_ Please upgrade in a staggered fashion - do not upgrade all nodes at the same time. You can coordinate with other validators in the Fuse Validators Telegram group to sequence your upgrades and maintain network stability throughout the rollout._\
1. Back up your keystore
Before making any changes, back up the keystore folder. This contains your node’s identity and should be preserved regardless of what happens to the database.
cp -r fusenet/keystore fusenet/keystore_backup
2. Stop the running node
sudo docker stop fuse
3. Remove or move the existing database
The old database is incompatible with v1.36.2 and must be cleared before resyncing.
Option A — Delete (faster):
rm -rf fusenet/database
Option B — Move aside (if you have disk space and want to keep a backup):
mv fusenet/database fusenet/database_backup
4. Re-run the quickstart script
Re-running the script with your original arguments will pull the updated version tags from GitHub and download the new containers from Docker Hub.
wget -O quickstart.sh https://raw.githubusercontent.com/fuseio/fuse-network/master/nethermind/quickstart.sh
chmod 755 quickstart.sh
./quickstart.sh -r [node_role] -n [network_name] -k [node_key]
Examples:
Validator on Fuse mainnet
./quickstart.sh -r validator -n fuse -k my-fuse-validator
Standard node on Fuse mainnet
./quickstart.sh -r node -n fuse -k my-fuse-node
Node on Spark testnet
./quickstart.sh -r node -n spark -k my-spark-node
Tip: If you already have quickstart.sh on your machine, re-running it with the same arguments is sufficient to pull the latest version — no need to re-download unless you want the freshest copy of the script itself.
5. Monitor the sync
Once the node restarts, confirm it is syncing correctly:
sudo docker logs fuse -f
You should see the node progressing through state sync. Sync typically completes in 2–6 hours. You can also monitor your node on the Fuse health dashboard (or Spark health dashboard for testnet).
Troubleshooting
Node not appearing on the health dashboard Allow a few minutes after startup. If the node still doesn’t appear after 10–15 minutes, check the logs for sync errors:
sudo docker logs fuse —tail 100
Sync appears stuck State sync can appear slow during the initial peer discovery phase. Wait at least 30 minutes before concluding there’s an issue. Check your firewall rules if peers are not connecting.
DB corruption error on startup This is expected if the old database was not fully cleared. Confirm fusenet/database is empty or removed, then restart the node.
Resources
