Home Assistant Recovery¶
Last Updated: 2026-07-19 Status: ✅ Scenario B tested for real on 2026-07-19 (Pacific's boot disk failed, HA failed over to Lotus for the duration of the rebuild) — worked as documented, MAC/IP handoff was seamless. Scenario A remains unverified. Purpose: How to recover Home Assistant depending on what has failed.
Before You Start¶
HA has two independent backup methods — choose the right scenario below.
| Situation | Use |
|---|---|
| HA is broken / won't load, but Pacific is running fine | Scenario A — Restore from HA backup |
| Pacific has failed completely | Scenario B — Migrate HA to Lotus |
Scenario A — Restore from HA Backup¶
Use this when Home Assistant is broken but the Pacific NUC is still running.
You have two options — try Option 1 first (faster, no VM reinstall needed).
Option 1 — Restore from HA built-in backup¶
HA automatically backs up its configuration. Backups are stored in two places: - Nabu Casa Cloud (offsite — accessible from anywhere) - Lotus (local share — accessible on home network)
Steps: 1. Open Home Assistant: http://192.168.1.12:8123 or https://ha.djchome.uk 2. Go to Settings → System → Backups 3. Select the most recent backup 4. Click Restore and follow the prompts 5. HA will restart — wait 5 minutes 6. Verify automations, integrations, and devices are working
Option 2 — Restore from Proxmox Backup Server¶
Use this if the HA VM itself is corrupt or the built-in restore fails.
Steps: 1. Open Proxmox: https://192.168.1.10:8006 2. Select the HomeAssistant VM (112) in the left panel 3. Click Backup in the top menu 4. Select the most recent backup from the PBS storage 5. Click Restore 6. Start the VM once restored 7. Verify HA loads at http://192.168.1.12:8123
Scenario B — Migrate HA to Lotus¶
Use this if Pacific has died and cannot be recovered quickly.
A standby Home Assistant VM exists on Lotus. It is configured with the same MAC address as the primary VM on Pacific, so when it comes online it will receive the same IP (192.168.1.12) and all devices and integrations will connect to it automatically.
Important caveat¶
The Lotus VM cannot be kept running at the same time as the primary on Pacific — they would conflict on the network. This means the Lotus VM may be behind on updates and recent config changes. After migrating, restore the latest HA backup to bring it up to date.
If Pacific needs an extended rebuild (not just a quick recovery)¶
Confirmed working 2026-07-19: if Pacific is going to be down for a while (e.g. a full disk replacement + OS reinstall), don't switch HA back and forth multiple times. Fail over to Lotus once, let it run as primary for as long as the rebuild takes (days/weeks is fine), fully soak-test Pacific before touching HA at all, then do a single controlled failback at the end (fresh backup taken from the Lotus instance right before shutting it down, restored into Pacific's VM once it's back). This removes all time pressure from the rebuild and avoids a second unnecessary outage.
Standby VM disk space¶
The Lotus standby VM's virtual disk may be smaller than the primary's (32GB vs Pacific's 42GB in one instance) since it isn't grown automatically over time the way the primary might be. If HAOS reports low storage after failover, grow the qcow2 live (virsh blockresize 'Home Assistant' <path> 64G on Lotus, no data risk — this only extends available space), then reboot the HA host (Settings → System → Hardware → Reboot Host, not just a core restart) — HAOS auto-expands its data partition into the new space on boot.
Steps¶
- Confirm Pacific is down and will not be coming back quickly
- On Lotus Unraid UI (http://192.168.1.80):
- Go to the VMs tab
- Start the standby Home Assistant VM
- Wait 5 minutes for it to boot
- Verify HA is accessible at http://192.168.1.12:8123
- Restore the latest HA backup (Settings → System → Backups) to get the most recent config
- Verify automations, integrations, and devices are working
After Pacific is repaired¶
When Pacific is restored, the Lotus VM must be shut down before Pacific comes back online to avoid IP conflict.
Post-Recovery Checklist¶
- [ ] HA loads at http://192.168.1.12:8123
- [ ] Remote access works at https://ha.djchome.uk
- [ ] Zigbee devices responding (lights, TRVs, blind)
- [ ] Frigate (doorbell camera) visible in HA
- [ ] Shelly Wall Display dashboard loading in living room
- [ ] Alexa integration working
- [ ] Automations running (check logbook for recent triggers)