Docker production: Orchestration, security, and beyond

Description

Take your containers into production and operate Docker clusters like a pro. This workshop skips Docker basics and avoids the CI/CD tools needed to get code ready for production. Rather, it focuses squarely on the tools and techniques of daily building and operating container clusters using Docker.

Starting where previous Docker workshops leave off, Bret Fisher, Shawn Bower, and Tony Pujals dive into the new Swarm mode clustering (services), failover, blue-green deployments, monitoring, logging, troubleshooting, and security, covering the latest built-in features and common third-party tools as they walk you through installing them on your own five-node cloud Swarm cluster.

Outline:

Docker Swarm

• SSH into your five nodes and create a Swarm
• Spin up a sample multicontainer microservice app
• Use built-in commands to validate it’s working and logging; use troubleshooting techniques on the overlay networks and to jump into containers
• Deploy redundant containers and test high availability
• Use blue-green deployments to update apps without downtime, including rollback
• Deal with failed containers and failed cluster nodes

Common third-party tools for logging, monitoring, and troubleshooting

• Set up and test ELK (Elasticsearch, Logstash, and Kibana)
• Set up and test cAdvisor
• Set up and test Prometheus

Key features of production systems

• Docker security features—seccomp profile, custom cgroups, apparmor profiles, and namespace setup
• Persistent data and backups love containers too—options from host, bind-mounts, commit, etc., options from sidecar/util container, shared storage 9EFS/NFS), mesh storage (infinit), and REX-Ray

Other things you’ll learn along the way

• Troubleshooting and sysadmin tools you may have forgotten
• New tools for the container admin
• Staying up to date on Docker tool releases
• Command aliases for quicker day-to-day container management
• Shell scripts for common processes