Backup & Recovery
Restore
Restore and quarantine
Restore
Restore returns your ERP to a previous backup state. Soviez distinguishes trusted backups (created and verified on your managed host) from untrusted sources (external files, old exports, or unknown origins).
Trusted vs untrusted
| Source | Behavior |
|---|---|
| Trusted / managed | Restore with integrity verification |
| External / untrusted | Must enter quarantine first |
untrusted restore
→ quarantine
→ scan
→ validation
→ explicit promotion (your approval)
A restored database is treated as potentially hostile until scanned. It may contain scheduled jobs, mail servers, webhooks, or other persistence mechanisms that should not run on Production until you approve promotion.
Commands
soviez.sh --restore <production-id> --backup <backup-id> [--confirm]
soviez.sh --restore-test <backup-id>
soviez.sh --restore-as-stage <backup-id> --stage-domain FQDN [--confirm]
soviez.sh --restore-status|cancel|retry|recover|rollback|cleanup <operation-id>
soviez.sh --security-quarantine-status|scan|promote|reject [ID]
Restore to Stage
Use --restore-as-stage to validate an old backup in an isolated Stage environment before touching Production. Stage requires its own domain (for example stage.example.com).
Full restore
A complete restore covers the database, filestore, manifest, and configuration with verification at each step. Do not skip integrity checks or force promotion when scans report FAIL.
Related
Source: canonical operator documentation · Permalink