Traditional Deployment Issues

• Physical Servers: Running apps on physical servers led to resource allocation problems.
  • No defined resource boundaries caused one app to dominate resources, impacting others.
  • Solution: Run each app on a separate server, but this was unscalable, underutilized resources, and costly.

Virtualization

• Definition: Creating virtual versions of OS, servers, storage, or network resources without physical existence.
  • Uses software to simulate hardware, enabling multiple OS and apps on one server.
• Types:
  • Network Virtualization: Divides bandwidth into secure, assignable channels for improved speed, availability, and security.
  • Storage Virtualization: Pools multiple storage devices into a single, centrally managed device.
    • Benefits: Appears local, combines small volumes into one, enhances reliability/performance, supports all OS, offers high availability, disaster recovery.
  • Data Virtualization: Aggregates data from various sources into a single virtual view.
    • Benefits: Abstracts technical details (location, structure), connects multiple sources, federates data, delivers as needed.
  • Desktop Virtualization: Isolates desktop OS from endpoint, supports remote access, includes shared hosted desktops (e.g., Microsoft Remote Desktop Services).
  • Application Virtualization: Runs apps separately from devices, simplifying IT management (install, patch, update once).
  • Hardware Virtualization: Creates virtual machines (VMs) acting as real computers with their own OS.
    • Example: A Windows machine hosting a VM running Ubuntu Linux.

Docker Swarm

• Definition: A container orchestration engine for managing multiple Docker Engines across hosts.
  • Built using SwarmKit, secure by default, allows declaring apps as stacks of services.
• Swarm Structure:
  • Nodes: Instances of Docker Engine in the swarm.
    • Manager Node: Dispatches tasks, manages orchestration, and cluster state; elects a leader.
    • Worker Node: Executes tasks assigned by managers, reports task status via an agent.
  • Services: Define tasks to run on nodes, specify container image and commands.
  • Tasks: Atomic units carrying a container and commands, assigned to nodes, non-migratory.
• Features:
  • Integrated cluster management, decentralized design, declarative service model.
  • Supports scaling, desired state reconciliation, multi-host networking, service discovery, load balancing, secure by default, rolling updates.
• Swarm Setup:
  • Requires Docker Engine 1.12+.
  • Ports: TCP 2371 (cluster management), TCP/UDP 7946 (node communication), UDP 4789 (overlay network).
  • Commands:
    • Initialize swarm: docker swarm init --advertise-addr <MANAGER-IP>
    • Check swarm status: docker info
    • List nodes: docker node ls
    • Get worker join token: docker swarm join-token worker
    • Join swarm: docker swarm join --token <token>
• Overlay Network:
  • Built on host networks, enables secure container communication.
  • Created on swarm initialization: ingress (load balancing) and docker_gwbridge (connects overlay to daemon’s network).

Service Management

• Service: Defines tasks for nodes, supports declarative model, scaling, discovery, updates, load balancing, internal DNS.
• Commands:
  • Create service: docker service create --replicas <no> --name <name> -p <port> <image> <command>
  • List services: docker service ls
  • Inspect service: docker service inspect <service>
  • List nodes running service: docker service ps <service>
  • Scale service: docker service scale <service>=<scale>
  • Update service: docker service update --image <image> <service>
  • Delete service: docker service rm <service>

Key Points for Docker (from Previous Notes)

• Generic Commands:
  • List: docker <object> ls
  • Inspect: docker <object> inspect <name/id>
  • Remove: docker <object> rm <name/id>
  • Prune: docker <object> prune
• Images:
  • Pull: docker image pull <image-name>
  • Build custom image: docker image build -t <image-name> <context>
  • Push to Docker Hub: docker image push <username>/<image-name>
• Containers:
  • Run: docker container run -d -i -t --name <name> -p <host-port>:<container-port> -v <volume>:<path> <image>
  • Exec command: docker container exec <name/id> <command>
  • Terminal: docker container exec -it <name/id> bash
• Volumes:
  • Create: docker volume create <volume-name>
  • Remove: docker volume rm <volume-name>
  • Prune: docker volume prune

Notes for Revision