Skip to content

Restoring from a backup

Something went wrong after a configuration change, an update, or a mistaken edit. Frigate will not start, cameras vanished from the UI, or behavior changed in a way you cannot fix quickly. If you saved a backup using Backing up your Frigate configuration, you can put the known-good files back and restart services.

This guide covers restoring configuration files on a running NVR. It does not cover full operating system recovery, ZFS pool rebuilds, or replacing a failed boot drive. Those scenarios need hardware work or LAN Foundry support. See Drive and pool errors for storage emergencies.

Before you start

This guide assumes:

  • You have a dated backup of config.yml (and ideally your Caddyfile) from another folder on the NVR or from an off-NVR copy
  • You can SSH into the NVR or use Portainer
  • The Frigate container exists but is misconfigured, or stops immediately after start

If the container problem is not config-related (disk full, network, image update), start with What to do when a container won't start.

Decide what to restore

Symptom Likely restore target
Frigate exits on start, YAML errors in logs config.yml
Cameras missing or wrong after editing config config.yml
Hostnames broken after editing proxy rules Caddyfile
Both Frigate and URLs broken config.yml and Caddyfile
Problem started after apt upgrade only OS issue; config restore may not help
Pool DEGRADED or recordings missing Storage issue; see drive-errors guide

When in doubt, restore config.yml first. It is the most common fix.

Step 1 — Stop Frigate

Restoring while Frigate is writing to config.yml can cause conflicts. Stop the container:

docker stop frigate

If Frigate is already stopped or crash-looping, skip this step.

Step 2 — Locate the live config path

FRIGATE_CONFIG=$(docker inspect frigate --format '{{ range .Mounts }}{{ if eq .Destination "/config" }}{{ .Source }}{{ end }}{{ end }}')
echo "$FRIGATE_CONFIG"

The active file is $FRIGATE_CONFIG/config.yml.

Step 3 — Preserve the broken file (optional)

Renaming instead of deleting lets you compare later:

cp "$FRIGATE_CONFIG/config.yml" "$FRIGATE_CONFIG/config.yml.broken-$(date +%Y%m%d)"

Step 4 — Copy the backup into place

If the backup is on the NVR (example date):

cp ~/nvr-backups/2026-06-15/config.yml "$FRIGATE_CONFIG/config.yml"

If the backup is on your admin PC, copy it to the NVR first:

scp ./lan-foundry-backups/2026-06-15/config.yml yourusername@192.168.1.100:~/config.yml.restore

Then on the NVR:

cp ~/config.yml.restore "$FRIGATE_CONFIG/config.yml"

Restore Caddyfile (if needed)

CADDYFILE=$(docker inspect caddy --format '{{ range .Mounts }}{{ if eq .Destination "/etc/caddy/Caddyfile" }}{{ .Source }}{{ end }}{{ end }}')
cp ~/nvr-backups/2026-06-15/Caddyfile "$CADDYFILE"
docker restart caddy

Step 5 — Start Frigate and verify

docker start frigate
docker logs frigate --tail 50

Look for a clean startup without YAML parse errors. Open https://frigate.internal and confirm:

  • Camera tiles show live video
  • Timeline and events look as expected
  • Settings match what you remember from before the bad change

If Frigate still fails, read the log line that mentions the error. You may have restored a backup from an older Frigate version that needs small edits for your current image. Compare with the .broken file you saved to see what changed.

Restore after a container recreate gone wrong

If an update removed volume mounts or recreated the container with empty config:

  1. Confirm where the old config still lives. Bind mounts often remain on the host even when the container was recreated. Check the path from docker inspect frigate.
  2. If the host path was wiped, copy config.yml from your off-NVR backup into the mount path from Step 2.
  3. Recreate the container using the same volumes and network as before. See Updating via Portainer and What to do when a container won't start.

Certificate data for Caddy usually survives in Docker volumes caddy-data and caddy-config unless those volumes were deleted. Restoring config.yml does not restore TLS certificates, and you typically do not need to.

When restore is not enough

Restore fixes configuration mistakes. It does not fix:

  • Full boot drive failure — the OS and Docker setup may need reinstallation or Argus support
  • ZFS pool failure or corruption — recordings may be lost independent of config.yml
  • Bad container image — roll back the image tag in Portainer or pull a previous Frigate version

Contact lanfoundry.com/support for Argus hardware if the NVR will not boot, the pool is FAULTED, or containers cannot be recreated after following the guides above.

Where to go from here

Was this page helpful?

Found something wrong or out of date? Let us know below, or open a GitHub issue if you have a GitHub account.