added backup script

This commit is contained in:
2026-02-25 01:47:35 +03:00
parent 9f36a48864
commit 9c0a3c19f6
4 changed files with 155 additions and 1 deletions

View File

@@ -81,3 +81,87 @@ docker compose -f stacks/media/compose.yaml down
| **Storage** | | |
| Gitea | 3000, 222 | storage |
| Copyparty | 3923 | storage |
## Backup
Backups use [Restic](https://restic.net/). The script backs up `/srv/homelab` (including `data/`), `/etc/caddy/Caddyfile`, and `/etc/unbound/unbound.conf`. It **stops all Docker stacks** before backup for consistent data, then starts them again—run it manually when you can afford a few minutes of downtime.
### One-time setup
1. Install restic (e.g. `pacman -S restic` or from [restic.net](https://restic.net/)).
2. Copy the env example and set repository and password:
```bash
cp scripts/restic-backup.env.example scripts/restic-backup.env
# Edit scripts/restic-backup.env: set RESTIC_REPOSITORY and RESTIC_PASSWORD
```
Examples for `RESTIC_REPOSITORY`:
- Local: `RESTIC_REPOSITORY=/backup/restic`
- SFTP: `RESTIC_REPOSITORY=sftp:user@backup-host:/restic`
- S3: `RESTIC_REPOSITORY=s3:s3.amazonaws.com/bucket-name`
3. Initialize the repo (once):
```bash
export RESTIC_PASSWORD='your-password'
restic -r /backup/restic init # use your actual repo path
```
### Create a backup
From the repo root, run:
```bash
sudo ./scripts/restic-backup.sh
```
Stacks are stopped, then the backup runs, then they are started again. A failed backup still triggers the start so the homelab comes back up.
### Inspect backups
List snapshots (IDs and timestamps):
```bash
source scripts/restic-backup.env
sudo restic snapshots -r $RESTIC_REPOSITORY
```
List files in a snapshot (e.g. latest or by ID):
```bash
restic ls latest -r $RESTIC_REPOSITORY
restic ls -r $RESTIC_REPOSITORY <snapshot-id>
```
Browse a path inside a snapshot:
```bash
restic ls latest -r $RESTIC_REPOSITORY /srv/homelab/config
```
### See diffs between backups
Compare two snapshots (added, changed, removed files and content diff):
```bash
restic diff <older-snapshot-id> <newer-snapshot-id> -r $RESTIC_REPOSITORY
```
### Restore
Restore the latest snapshot into a directory (does not overwrite the repo; use a separate target dir):
```bash
restic restore latest --target /tmp/restore -r $RESTIC_REPOSITORY
```
Restore a specific snapshot or path:
```bash
restic restore <snapshot-id> --target /tmp/restore -r $RESTIC_REPOSITORY
restic restore latest --path /etc/caddy --target /tmp/restore -r $RESTIC_REPOSITORY
```
After restore, fix ownership on `data/` if needed (see [Monitoring permissions](#2-monitoring-permissions)).