Docker Volume

Why Volumes Are Needed

• Problem: Data inside a container is ephemeral—deleting the container deletes its data.
• Solution: Use Docker volumes to persist data across container lifecycles.

Storage Type	Description	Pros	Cons
Container internal	Data stored inside container	Simple	Deleted when container removed
Volume	Docker-managed directory on host	Persistent, shareable across containers	Managed by Docker
Bind mount	Direct host path mapping	Good for development	Tightly coupled to host

Basic Workflow

1. Create a volume:

docker volume create pg-data

2. Run a PostgreSQL container mounting the volume:

docker run -d \
  --name pg-demo \
  -e POSTGRES_USER=demo \
  -e POSTGRES_PASSWORD=demo1234 \
  -e POSTGRES_DB=labdb \
  -p 15432:5432 \
  -v pg-data:/var/lib/postgresql/data \
  postgres:15
 
# 윈도우 CMD에서는 아래 명령어를 사용하세요.
docker run -d ^
  --name pg-demo ^
  -e POSTGRES_USER=demo ^
  -e POSTGRES_PASSWORD=demo1234 ^
  -e POSTGRES_DB=labdb ^
  -p 15432:5432 ^
  -v pg-data:/var/lib/postgresql/data ^
  postgres:15

• -v pg-data:/var/lib/postgresql/data → Core option. Mounts the Docker volume pg-data to the container’s PostgreSQL data directory. This ensures data persists even if the container is deleted.
• -p 5432:5432 → Maps host port 5432 to container port 5432, allowing external clients (psql, DataGrip, etc.) to connect to the database.
• -e POSTGRES_USER/POSTGRES_PASSWORD/POSTGRES_DB → Sets the database user, password, and database name on first container startup. This initializes the database automatically.

3. Connect externally using psql or a GUI (DataGrip, DBeaver)

# Option 1: Local `psql` client (assumes installed on your machine):
psql -h localhost -p 15432 -U demo -d labdb # Password: demo1234
 
# Option 2: Using `psql` inside the container (no local installation required):
docker exec -it pg-demo psql -U demo -d labdb

4. Create Table & Insert Data

-- Create table
CREATE TABLE IF NOT EXISTS students (
  id SERIAL PRIMARY KEY,
  name TEXT NOT NULL,
  created_at TIMESTAMPTZ DEFAULT now()
);
 
-- Insert sample data
INSERT INTO students (name) VALUES ('Alice'), ('Bob');
 
-- Verify data
SELECT * FROM students ORDER BY id;

• SERIAL automatically increments the id.
• TIMESTAMPTZ DEFAULT now() adds the creation timestamp.
• Using IF NOT EXISTS avoids errors if the table already exists.

5. Stop and remove the container:

docker stop pg-demo && docker rm pg-demo

6. Re-run the container using the same volume & Confirm

# rerun
docker run -d \
  --name pg-demo \
  -e POSTGRES_USER=demo \
  -e POSTGRES_PASSWORD=demo1234 \
  -e POSTGRES_DB=labdb \
  -p 15432:5432 \
  -v pg-data:/var/lib/postgresql/data \
  postgres:15
 
# Confirm data remains
psql -h localhost -p 15432 -U demo -d labdb -c "SELECT * FROM students;

7. DataGrip

1. DataGrip → New → PostgreSQL
2. Host: localhost, Port: 15432
3. User: demo, Password: demo1234, Database: labdb
4. Test Connection → OK

• But ppl usually use DBeaver than DataGrip nowadays tho

Volume Backup & Restore

Backup

docker run --rm \
  -v pg-data:/var/lib/postgresql/data:ro \
  -v ~/dev/pg-volume-lab/backup:/backup \
  alpine sh -c 'tar czf /backup/pgdata_$(date +%Y%m%d_%H%M).tgz -C /var/lib/postgresql/data .'

Restore

docker run --rm \
  -v pg-data:/var/lib/postgresql/data \
  -v ~/dev/pg-volume-lab/backup/pgdata_YYYYMMDD_HHMM.tgz:/backup/backup.tgz:ro \
  alpine sh -c 'rm -rf /var/lib/postgresql/data/* && tar xzf /backup/backup.tgz -C /var/lib/postgresql/data --strip-components=1 && chown -R 999:999 /var/lib/postgresql/data'

• Then recreate the container using the same volume.

Practical Tips

• Version pinning: e.g., postgres:15.6 ensures reproducibility.
• Volume naming: Include service/environment, e.g., svc-db-prod-data.
• Backup strategy: Regular pg_dump or snapshots.
• Security: Store credentials in .env or Docker secrets.
• Cross-platform issues: Watch for permission problems on Linux/Mac/Windows.