Disaster Recovery Architecture for
a Multi-Site Retail Operation
No tested recovery plan and backups that had never been verified
A retail business operating 8 locations across the UK and Europe came to GBWise after a near-miss: a server failure at their head office caused 6 hours of point-of-sale downtime across all locations before systems were manually restored. An investigation revealed that their backup solution — which had been running for three years — had never been tested for recovery. Several backup jobs had been silently failing for months.
- Backup solution had been silently failing on 3 of 8 sites — discovered only after the outage
- No documented RTO or RPO — the business had no idea how long recovery would take
- POS systems at all 8 locations depended on a single head-office server with no failover
- No offsite or cloud backup copy — all backup data stored on-site, vulnerable to fire or theft
- The 6-hour outage had cost the business an estimated £40,000 in lost sales and staff time
3-2-1-1 backup architecture with tested recovery procedures
Backup audit and failure remediation
Full audit of existing backup jobs across all 8 sites. Failed jobs identified and root-caused. Proxmox Backup Server deployed at head office as the primary backup target with verified job monitoring and alerting.
3-2-1-1 architecture implementation
Primary backups to on-site PBS, replicated to a secondary off-site PBS instance, with immutable cloud copy to Backblaze B2. Air-gapped copy rotated monthly. All backup jobs configured with verification — automatic restore test on every backup.
Head-office server HA and failover
Head-office infrastructure migrated to a 2-node Proxmox HA cluster. POS dependency on single server eliminated — any node failure triggers automatic VM restart within 60 seconds.
DR runbook and live recovery test
Documented DR runbook produced covering every failure scenario. Full DR test executed — simulated complete head-office server loss, recovered all systems from backup in 22 minutes. Results presented to the board.
Monitoring and quarterly DR testing
Centralised monitoring deployed across all 8 sites. Quarterly DR tests scheduled as part of ongoing managed services contract — every recovery test is documented and signed off.
Stack
(Name withheld at client request)
Outcomes at 90 days post-implementation
When did you last test your backups?
We offer a free backup and DR audit for qualified prospects.