Connectivity Between Application and Database

Connection strings

• Every Database generally has a
  • hostname (name/ip address)
  • username
  • password
  • database name (optional)
• Generally all languages use a connection string to establish connection
• When we run our application in containers in most of the cases the db details are passed as environmental variables and in somecases it will be file

Networking

• Docker supports creating networks of different types
  • bridge network (single host)
  • host network (single host)
  • macvoverlan (single host)
  • overlay (multi host)
• Default bridge network cannot resolve by names but uses only ip, custom bridge networks can resolve with names as well as ips

Running a simple 2 tier application

Python FastApi and Postgres

• Refer Here for the repo
• Database: postgres
  • Environmental variables
    • POSTGRES_USER=user
    • POSTGRES_PASSWORD=password
    • POSTGRES_DB=library
• Steps:
  • Create a bridge network with a predefined range of ip address 192.168.0.0/24
    
  • create a postgres container with name libdb with the environmental variables as shown above and attached to the network.
    
  • Build the image if you donot have it yet from the repo
  • Now run the container in the same network as db with name libapp1, and set DATABASE_URL environmental variable
    
    
• Repeat the above steps by attaching volume for libdb container
• Use this following dockerfile as reference, to build the docker image of library application

# Stage 1: Build Stage
FROM python:3.11-alpine AS builder

# Set environment variables to avoid Python writing .pyc files and to buffer output
ENV PYTHONDONTWRITEBYTECODE=1
ENV PYTHONUNBUFFERED=1

# Install necessary build dependencies
RUN apk update && apk add --no-cache gcc musl-dev libffi-dev openssl-dev postgresql-dev

# Create a directory for the application
WORKDIR /users

# Copy the requirements file to install dependencies
COPY requirements.txt .

# Install the application dependencies
RUN pip install --no-cache-dir -r requirements.txt

# Stage 2: Runtime Stage
FROM python:3.11-alpine

# Set environment variables to avoid Python writing .pyc files and to buffer output
ENV PYTHONDONTWRITEBYTECODE=1
ENV PYTHONUNBUFFERED=1

# Install runtime dependencies, including PostgreSQL client libraries
RUN apk add --no-cache libpq

# Create a directory for the application
WORKDIR /users

# Copy installed dependencies from the builder stage
COPY --from=builder /usr/local/lib/python3.11/site-packages /usr/local/lib/python3.11/site-packages
COPY --from=builder /usr/local/bin /usr/local/bin

# Copy the application source code to the working directory
COPY . .

# Ensure uvicorn is installed in case it wasn't properly copied or was missed
RUN pip install --no-cache-dir uvicorn

# Expose the port the app runs on
EXPOSE 8000

# Command to run the application
CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8000"]

Findout What a docker registry is and what is public registry and private registry

• Private registries are used to store images securely and access will be provided within your organization
• Popular private registries
  • Docker hub (Paid Plan)
  • Jfrog
  • Azure Container Registry
  • AWS Elastic Container Registry
• When we create a private registry, we create repository for each image
• Exercise:
  • Build library image usign python:3.11-alpine and push into
    • dockerhub (public)
    • acr
    • ecr