UPES-ECS Backup & Restore Procedure¶
Model: local-first. No cloud dependency. Sensitive data encrypted.
Names: snapshots ECS Snapshot YYYYMMDD-HHMM · export upes-ecs-export-YYYYMMDD.zip · git repo upes-ecs-config · restore doc UPES-ECS Restore Checklist.
A backup you have never restored is not a backup. Test restore quarterly (monthly during pilot).
1. What gets backed up¶
| Category | Includes |
|---|---|
| Asterisk config | pjsip.conf, extensions.conf + extensions_custom.conf, queues.conf, voicemail.conf, confbridge.conf, features.conf, rtp.conf, custom logger.conf/modules.conf |
| Identity | SAP-ID ↔ extension mappings, display names, roles/contexts, fixed-device mappings, caller-ID names, account status |
| Emergency dialplan | 111, 199, queue routing, escalation, voicemail fallback, 700–799 paging, 9000–9004 conferences, transfer rules |
| Queue config | members, strategy, timeouts, announcements, pause behaviour |
| Prompts | voicemail, hold, drill, paging-test, any pre-recorded messages |
| Recordings/voicemail | emergency recordings + voicemails (retention policy) |
| Logs/metadata | CDR, CEL, queue, paging, conference, missed-emergency, access-denied, registration, health logs |
| Health config | check scripts, thresholds, critical-device list, dashboard config |
| Docs | SOP + drill docs |
2. Backup types & schedule¶
| Type | When | Contents |
|---|---|---|
| Pre-change (mandatory) | Before any production change | Full config snapshot |
| Daily config | Every day | Config + SIP account data + SAP mappings + fixed mappings + dialplan + queues + voicemail config + prompts |
| Weekly export | Weekly | Responder directory, device list, role/context map, extension ownership, health config, relevant logs → upes-ecs-export-YYYYMMDD.zip |
| Policy recordings | Per policy | Emergency recordings + voicemails (encrypted) under retention |
Pre-change is mandatory for any change to: 111 routing, ERT queue, SIP accounts, contexts, paging codes, conference rooms, voicemail, recording paths, access control.
3. Retention of backups¶
- Config: keep 30 daily + 12 weekly snapshots (auto-rotate).
- Logs: 1 year.
- Recordings/voicemail: 90 days unless flagged for preservation.
4. Storage layout (local-first)¶
Primary: backup directory on the PBX server (separate disk if possible) or local NAS
Secondary: separate LAN machine / NAS
Offline: encrypted USB copy, physically locked, held by authorized admin
No cloud required.
5. Security¶
Backups can contain SIP credentials, SAP-ID identity data, recordings, and logs — treat as sensitive.
- Access: IT Admin + approved UPES-ECS owner only. Not students/staff/normal ERT.
- Encryption: config-only backups = restricted access; credential/recording backups = restricted access + encrypted.
- Credentials stored as secrets/hashes, never plain text in docs.
- Backup access is logged. Offline media physically locked.
- Deletion requires IT Admin + university authority / ERT Lead approval.
6. Versioning (git)¶
Keep extensions_custom.conf and all custom config in the LAN-local git repo upes-ecs-config with change notes and rollback tags:
v1.0 Initial 111 emergency hotline
v1.1 Added ERT queue
v1.2 Added fallback escalation
v1.3 Added paging restrictions
v1.4 Added SAP-ID user accounts
Change control — every production change records: what / who / when / why / backup taken / test result / rollback plan.
FreePBX's own Backup & Restore module covers the GUI-managed config; the git repo covers custom dialplan + scripts. Use both.
7. Restore¶
Who: IT Admin + approved UPES-ECS owner. Target time: config-only restore under 1 hour.
Restore order:
1. OS / network (static IP, hostname)
2. Asterisk/FreePBX config
3. SIP accounts + SAP-ID mappings
4. Dialplan + queues
5. Prompts + voicemail
6. Logs / recordings (as needed, encrypted)
Verification checklist (UPES-ECS Restore Checklist): - [ ] Service running - [ ] SIP users register - [ ] Student-to-student calling works - [ ] 199 test passes, then 111 reaches the queue - [ ] Recording works · [ ] Voicemail works · [ ] Queue works - [ ] Fixed phones (4200/4300) register - [ ] Unauthorized paging blocked · [ ] Unauthorized 9000 blocked - [ ] Health monitoring reports correctly · [ ] Backup status current
Sign-off: IT Admin + ERT Lead.
8. Failure handling¶
- Backup fails → Critical alert; fix before any config change or go-live.
- Restore fails → escalate to IT lead; use the previous snapshot or the manual rebuild plan.
9. Test cadence¶
- Restore test after every major change.
- Full restore drill: quarterly (monthly during pilot).
- SAP-ID ownership history must be preserved through any restore so call logs and identities stay accurate.