Movement Network's documentation PR #258 adds a callout warning under the "Bootstrapping from backups" section on the archival-node run page. The warning addresses a known issue affecting nodes restored from older public snapshots: the backup node's persisted network identity, stored in `db/ephemeral_identity_key`, is preserved during restoration.
When multiple nodes are restored from the same snapshot, they retain the same peer ID, causing validator fullnodes to detect duplicate connections and close them repeatedly, resulting in `NoConnectedPeers` errors and constant reconnection cycles. The fix is straightforward: operators must delete the `db/ephemeral_identity_key` file and restart the node, allowing it to generate a unique network identity.
The affected date ranges are specified: mainnet restores on or after May 5, 2026, and testnet restores from snapshots dated January 20 to February 6, 2026. A companion change to the `l1_restore.sh` script in the movement-networks repository now automatically removes this file during restoration, preventing the issue for future deployments.
This update helps operators quickly diagnose and resolve connectivity issues when restoring archival nodes from historical snapshots, reducing troubleshooting time and improving node reliability across the network.