DevOps on Chen Kai Blog

Docker and Containers (5): Docker Compose — Multi-Container Applications

Tue, 20 Jun 2023 09:00:00 +0000

The previous articles taught you how to run containers with docker run, pass port mappings with -p, create networks with docker network create, and mount volumes with -v. Now imagine doing that for a web server, an API backend, a database, a cache, and a task queue — every time you start your development environment. Docker Compose replaces those 20+ commands with a single file and a single command: docker compose up.

Docker and Containers (3): Dockerfile Patterns — From Naive to Production

Sun, 18 Jun 2023 09:00:00 +0000

Most tutorials show you a 5-line Dockerfile and move on. When you deploy to production, you might find your image is 1.2 GB, builds take 8 minutes even for a one-line code change, and your security team flags vulnerabilities in packages you didn’t know were installed. Writing a good Dockerfile is a skill that pays off every time your CI pipeline runs.

Every Dockerfile Instruction#

Let’s go through every instruction you’ll use, with concrete examples.

Docker and Containers (1): Why Containers — The Problem VMs Didn't Solve

Fri, 16 Jun 2023 09:00:00 +0000

Every developer has heard the phrase “it works on my machine.” Virtual machines were supposed to fix that, and they did — at the cost of gigabytes of RAM, minutes of boot time, and an entire duplicate operating system per application. Containers asked a different question: what if we could isolate applications without duplicating the kernel?

The Actual Problem#

Consider deploying a Python web application. You need Python 3.11, specific pip packages, a particular version of libssl, and some system-level configuration. Your colleague’s app needs Python 3.9 and a conflicting libssl version. The staging server runs Ubuntu 20.04 while production runs Amazon Linux 2.

Cloud Computing (7): Cloud Operations and DevOps Practices

Fri, 26 May 2023 09:00:00 +0000

In 2017 GitLab lost six hours of database state. An engineer, exhausted, ran rm -rf on the wrong server during an incident. The backup procedures had silently been broken for months; nobody noticed because no one was restoring from backups. The lesson is not “be careful with rm”. The lesson is that operations is a system — tools, runbooks, monitoring, automation, and the rituals around them. When the system is healthy, no single tired engineer can take down production. When the system is rotten, every late-night fix is one keystroke from disaster.