Boot Management

@qubiq/core/boot ships a small orchestration layer that mirrors the native node’s boot flags. It’s ideal for validators, custodial services, or automation daemons that need to resume safely after downtime.

BootManager

import { BootManager, BootMode } from "@qubiq/core/boot";

const manager = new BootManager({
  persistPath: ".qubic/boot-state.json",
});

const decision = await manager.decide();
if (decision.mode === BootMode.Scratch) {
  console.log("Start clean sync", decision.reason);
} else {
  console.log("Resume from cached state");
}

• Persists the last successful epoch (FileBootStateStore).
• Queries live/query services to compare epochs and detect gaps.
• Honors overrides: forcedMode, desiredEpoch, requireCleanState.

Boot state stores

• FileBootStateStore (default) keeps JSON on disk.
• Implement BootStateStore to back Redis, S3, or database tables.
• Multiple workers can share one store to coordinate restarts.

Client hooks

Inject custom LiveServiceClient / QueryServiceClient instances for air-gapped setups:

const manager = new BootManager({
  liveClient: new LiveServiceClient({ baseUrl: process.env.LIVE_URL }),
  queryClient: new QueryServiceClient({ baseUrl: process.env.QUERY_URL }),
});

Putting it together

• Run BootManager.decide() before launching long-running connectors or automation runtimes.
• Store the resulting flag inside provisioning scripts (Kubernetes configmaps, systemd env vars, etc.).
• Emit telemetry whenever mode changes so you can alert on repeated scratch boots.