Initial commit: BTC Bot with dashboard, TA analysis, and 14 timeframes
This commit is contained in:
BTC Bot
251
SYNOLOGY_INSTALL.md
Normal file
251
SYNOLOGY_INSTALL.md
Normal file
@ -0,0 +1,251 @@
|
||||
# Installing BTC Bot on Synology NAS
|
||||
|
||||
Tested on Synology DS218+ (DSM 7.x, 2 GB RAM). The system runs three Docker
|
||||
containers: TimescaleDB, a data collector (WebSocket to Hyperliquid), and a
|
||||
FastAPI server with a web dashboard.
|
||||
|
||||
## Prerequisites
|
||||
|
||||
1. **Docker** -- install from Synology Package Center.
|
||||
2. **SSH access** -- enable in DSM: Control Panel > Terminal & SNMP > Enable SSH.
|
||||
3. **Git** (optional) -- install the Git Server package, or copy files manually.
|
||||
|
||||
## Step 1 -- Set NAS DNS
|
||||
|
||||
Docker on Synology often can't resolve hostnames during image builds. Fix this
|
||||
before anything else:
|
||||
|
||||
**DSM > Control Panel > Network > General tab > set DNS Server:**
|
||||
|
||||
| Field | Value |
|
||||
|-------|-------|
|
||||
| Preferred DNS | `8.8.8.8` |
|
||||
| Alternate DNS | `8.8.4.4` |
|
||||
|
||||
Click Apply.
|
||||
|
||||
## Step 2 -- Get the code onto the NAS
|
||||
|
||||
SSH into the NAS and clone/copy the project:
|
||||
|
||||
```bash
|
||||
ssh your_user@nas_ip
|
||||
cd /volume1/homes/your_user # or wherever you prefer
|
||||
git clone <repo_url> btc_bot
|
||||
cd btc_bot
|
||||
```
|
||||
|
||||
Or copy files via SMB/SCP to `/volume1/homes/your_user/btc_bot`.
|
||||
|
||||
## Step 3 -- Create directories
|
||||
|
||||
```bash
|
||||
sudo mkdir -p /volume1/btc_bot/{data,backups,logs,exports}
|
||||
```
|
||||
|
||||
These are mounted as volumes by the containers.
|
||||
|
||||
## Step 4 -- Configure environment
|
||||
|
||||
```bash
|
||||
cp .env.example .env
|
||||
nano .env # or vi
|
||||
```
|
||||
|
||||
At minimum, set a strong password for these two values:
|
||||
|
||||
```
|
||||
DB_PASSWORD=your_secure_password_here
|
||||
API_SECRET_KEY=your_secret_key_here
|
||||
```
|
||||
|
||||
The remaining values can stay at their defaults for data collection.
|
||||
|
||||
## Step 5 -- Build images
|
||||
|
||||
The standard `docker-compose build` may fail with DNS errors inside the build
|
||||
container. Use `docker build` with `--network host` to bypass this:
|
||||
|
||||
```bash
|
||||
cd ~/btc_bot
|
||||
|
||||
# Build both images (uses host network for DNS resolution)
|
||||
docker build --network host -f docker/Dockerfile.collector -t btc_collector .
|
||||
docker build --network host -f docker/Dockerfile.api -t btc_api .
|
||||
```
|
||||
|
||||
If you make code changes later, re-run these two commands to rebuild.
|
||||
|
||||
## Step 6 -- Start services
|
||||
|
||||
```bash
|
||||
cd ~/btc_bot/docker
|
||||
docker-compose up -d --no-build
|
||||
```
|
||||
|
||||
The `--no-build` flag tells compose to use the images you built in Step 5
|
||||
instead of trying to build through its own (potentially broken) network.
|
||||
|
||||
TimescaleDB will initialize the schema automatically from
|
||||
`docker/init-scripts/01-schema.sql` on first start.
|
||||
|
||||
## Step 7 -- Verify
|
||||
|
||||
```bash
|
||||
# Check all three containers are running
|
||||
docker-compose ps
|
||||
|
||||
# Watch collector logs (Ctrl+C to stop)
|
||||
docker-compose logs -f data_collector
|
||||
|
||||
# Check database health
|
||||
docker exec btc_timescale pg_isready -U btc_bot -d btc_data
|
||||
|
||||
# Run the health check script
|
||||
bash ~/btc_bot/scripts/health_check.sh
|
||||
```
|
||||
|
||||
The API dashboard is available at: **http://nas_ip:8000/dashboard**
|
||||
API docs (Swagger): **http://nas_ip:8000/docs**
|
||||
|
||||
## Setting up automated backups
|
||||
|
||||
The backup script dumps the database every run and keeps 30 days of history.
|
||||
|
||||
### Option A -- DSM Task Scheduler (recommended)
|
||||
|
||||
1. Open DSM > Control Panel > Task Scheduler
|
||||
2. Create > Scheduled Task > User-defined script
|
||||
3. Set schedule: every 6 hours
|
||||
4. User-defined script:
|
||||
```
|
||||
/bin/bash /volume1/homes/your_user/btc_bot/scripts/backup.sh
|
||||
```
|
||||
5. Run as: `root` (needed for docker exec)
|
||||
|
||||
### Option B -- Cron (via SSH)
|
||||
|
||||
```bash
|
||||
sudo crontab -e
|
||||
```
|
||||
|
||||
Add this line (runs every 6 hours):
|
||||
|
||||
```
|
||||
0 */6 * * * /bin/bash /volume1/homes/your_user/btc_bot/scripts/backup.sh >> /volume1/btc_bot/logs/backup.log 2>&1
|
||||
```
|
||||
|
||||
Backups are stored in `/volume1/btc_bot/backups/`.
|
||||
|
||||
## Setting up health monitoring
|
||||
|
||||
Add another scheduled task (every 5 minutes) to check if everything is running:
|
||||
|
||||
```
|
||||
/bin/bash /volume1/homes/your_user/btc_bot/scripts/health_check.sh
|
||||
```
|
||||
|
||||
## Common operations
|
||||
|
||||
### View logs
|
||||
|
||||
```bash
|
||||
cd ~/btc_bot/docker
|
||||
docker-compose logs -f data_collector # collector
|
||||
docker-compose logs -f api_server # API
|
||||
docker-compose logs -f timescaledb # database
|
||||
```
|
||||
|
||||
### Restart a single service
|
||||
|
||||
```bash
|
||||
cd ~/btc_bot/docker
|
||||
docker-compose restart data_collector
|
||||
```
|
||||
|
||||
### Stop everything
|
||||
|
||||
```bash
|
||||
cd ~/btc_bot/docker
|
||||
docker-compose down
|
||||
```
|
||||
|
||||
### Rebuild after code changes
|
||||
|
||||
```bash
|
||||
cd ~/btc_bot
|
||||
docker build --network host -f docker/Dockerfile.collector -t btc_collector .
|
||||
docker build --network host -f docker/Dockerfile.api -t btc_api .
|
||||
cd docker
|
||||
docker-compose up -d --no-build
|
||||
```
|
||||
|
||||
### Reset database (destroys all data)
|
||||
|
||||
Only do this if you need a fresh schema:
|
||||
|
||||
```bash
|
||||
cd ~/btc_bot/docker
|
||||
docker-compose down -v
|
||||
docker-compose up -d --no-build
|
||||
```
|
||||
|
||||
### Restore from backup
|
||||
|
||||
```bash
|
||||
gunzip /volume1/btc_bot/backups/btc_data_YYYYMMDD_HHMM.dump.gz
|
||||
docker exec -i btc_timescale pg_restore -U btc_bot -d btc_data --clean < /volume1/btc_bot/backups/btc_data_YYYYMMDD_HHMM.dump
|
||||
```
|
||||
|
||||
## Resource usage
|
||||
|
||||
The system is tuned for the DS218+ (2 GB RAM, dual-core):
|
||||
|
||||
| Container | Memory limit | Typical usage |
|
||||
|-----------|-------------|---------------|
|
||||
| TimescaleDB | 1.5 GB | ~400-800 MB |
|
||||
| Data Collector | 256 MB | ~50-100 MB |
|
||||
| API Server | 512 MB | ~50-100 MB |
|
||||
|
||||
PostgreSQL settings in `docker/timescaledb.conf` are optimized for this
|
||||
hardware. If running on a more powerful NAS, increase `shared_buffers` and
|
||||
`effective_cache_size` proportionally.
|
||||
|
||||
## Troubleshooting
|
||||
|
||||
### DNS errors during build
|
||||
|
||||
Use `--network host` when building (Step 5). If it still fails, check that
|
||||
your NAS can resolve DNS from the host:
|
||||
|
||||
```bash
|
||||
nslookup pypi.org
|
||||
```
|
||||
|
||||
If that fails, fix DNS in DSM (Step 1).
|
||||
|
||||
### Container keeps restarting
|
||||
|
||||
```bash
|
||||
docker logs btc_collector # check error output
|
||||
docker logs btc_api
|
||||
```
|
||||
|
||||
Common causes: wrong `DB_PASSWORD` in `.env`, database not ready yet (wait
|
||||
30 seconds after first start), or port 5432/8000 already in use.
|
||||
|
||||
### No data appearing
|
||||
|
||||
1. Check collector logs: `docker-compose logs data_collector`
|
||||
2. Verify WebSocket connectivity: the NAS needs outbound access to
|
||||
`wss://api.hyperliquid.xyz/ws` (port 443)
|
||||
3. Check database: `docker exec btc_timescale psql -U btc_bot -d btc_data -c "SELECT COUNT(*) FROM candles;"`
|
||||
|
||||
### Ports used
|
||||
|
||||
| Port | Service | Direction |
|
||||
|------|---------|-----------|
|
||||
| 5432 | PostgreSQL | Internal (+ exposed to host) |
|
||||
| 8000 | API/Dashboard | Inbound from LAN |
|
||||
| 443 | PyPI, Hyperliquid WS | Outbound |
|
||||
| 53 | DNS | Outbound |
|
||||
Reference in New Issue
Block a user