Recovering Billing Operations After a Host Restart

WayaDesk / Operations and recovery

A host restart should be part of acceptance testing, not an unexpected first experiment during a power interruption. Verify which processes return, what unfinished work remains and how the operator confirms service has recovered.

Check every required component

The application worker, public customer endpoint and any RADIUS bridge or daemon may have different startup arrangements. Record how each is started and who maintains it. Opening the admin page alone does not prove that all background and public workflows are available.

Inspect unfinished work

Review pending payments, durable access jobs and notification attempts. LinkDesk’s recovery design preserves payment identifiers and access work, while ambiguous SMS sends may need owner review. Do not assume every interrupted operation should be repeated automatically.

• Confirm the correct data directory and installed release are in use.
• Check fresh router observations after startup.
• Review the latest backup and unresolved warnings.

Test a representative subscriber

Observe valid authentication and the expected paid period. If the deployment uses public checkout, test its reachability separately. Avoid creating another charge merely to establish whether a service recovered.

Improve the restart runbook

Write the shortest reliable sequence that the operator can perform without the original installer. Include where to look for evidence and when to escalate. Recovery is successful when the relevant service and records are verified, not simply when the computer reaches its desktop.