firefrost-operations-manual/docs/migration-rollback-plan.md

Migration Rollback Plan

If Things Go Wrong

Immediate Rollback (During Migration)

If service fails on new VPS:

1. DO NOT stop TX1 service
2. Revert DNS in Cloudflare (instant)
3. TX1 service continues working
4. Debug VPS issue offline

Maximum Downtime: DNS TTL (5 minutes)

---

Service-Specific Rollback

Gitea Rollback

# On Cloudflare
git.firefrostgaming.com A record → 38.68.14.26 (TX1)

# On TX1
systemctl start gitea

Uptime Kuma Rollback

# On Cloudflare
status.firefrostgaming.com A record → 38.68.14.26 (TX1)

# On TX1
systemctl start uptime-kuma

MkDocs Rollback

# On Cloudflare
docs.firefrostgaming.com A record → 38.68.14.26 (TX1)

# On TX1
systemctl start mkdocs

---

Data Recovery

If Backup Needed

All backups stored in /root/migration-backups/ on TX1:

• gitea-backup.tar.gz + gitea-dump.sql
• kuma-backup.tar.gz
• mkdocs-source.tar.gz

Retention: 7 days after successful migration

---

Lessons Learned Log

What went wrong:

Root cause:

How we fixed it:

How to prevent next time:

---

Remember: Don't panic. TX1 services are still running. DNS changes are instant.