Fork Repair

Immediate procedure

1. Freeze rollout changes and collect block hashes from a quorum of validators.

2. Identify the highest common finalized height before restarting any validator.

3. Do not delete PVCs. Snapshot first, then repair from a signed operator runbook.

Capture network evidence

Record the endpoint, returned chain ID, node version, timestamps, current height, and observed error. Compare the affected node with healthy nodes on the same chain; retain logs and signed evidence before making changes.

curl --fail-with-body --show-error --silent --max-time 15 'https://sigil-velocity-canary-root-1.fly.dev/rpc' \
  -H 'content-type: application/json' \
  --data '{"jsonrpc":"2.0","id":1,"method":"sigil_getNodeInfo","params":{}}'

Verify recovery

Repeat the original failing check. Confirm expected synchronization and activation state, and compare the relevant block hashes or state roots at the same height. Record the recovery action, evidence, and remaining uncertainty in the incident record.

curl --fail-with-body --show-error --silent --max-time 15 'https://sigil-velocity-canary-root-1.fly.dev/rpc' \
  -H 'content-type: application/json' \
  --data '{"jsonrpc":"2.0","id":1,"method":"sigil_getSyncStatus","params":{}}'

• Operator guide
• RPC reference