Containerize Your Enterprise DevOpsDays Austin 2015 Zach Jordan, Account Manager Sandeep Parikh, Solutions Architect

www.starwars.com/may-the-4th

We work with developers and business to help them adopt Google Cloud We build architectural solutions and patterns demonstrating how to use Google’s cloud What We Do At Google

What We Actually Do At Google

What Does DevOps Do? “It worked on my machine!” DevOps is the practice of operations and development engineers participating together in the entire service lifecycle, from design through the development process to production support. Infrastructure and environments need to be repeatedly provisioned, orchestrated, and configured, but every environment is different.

You are the keepers of “state”

Containers, Containers, Containers!

From VMs to Containers

Google has been developing and using containers to manage our applications for over 10 years. Images by Connie Zhou

Why containers? • Performance • Repeatability • Isolation • Quality of service • Accounting • Visibility • Portability A fundamentally different way of managing applications Images by Connie Zhou

Everything at Google runs in containers: • Gmail, Web Search, Maps, ... • MapReduce, batch, ... • GFS, Colossus, ... • Even GCE itself: VMs in containers

• Gmail, Web Search, Maps, ... • MapReduce, batch, ... • GFS, Colossus, ... • Even GCE itself: VMs in containers We launch over 2 billion containers per week. Everything at Google runs in containers:

Google is betting big on containers as the future of cloud computing workloads We want to let everyone run applications and infrastructure the same way we do

Welcome to Kubernetes Run and manage a cluster of containers as a single system Orchestrate Docker containers by scheduling on to cluster nodes Ensure state and group into units for management and discovery Manage applications, not machines kubernetes.io

Kubernetes is... Fully open source and actively developed in the open Driven by partners, big and small, along with Google Available to be deployed everywhere, public and private kubernetes.io

Democratizing Deployments Kubernetes runs on many, many platforms Hybrid architectures use the same containers everywhere Public Cloud Kubernetes Private Cloud Kubernetes Bare Metal Kubernetes

https://github.com/GoogleCloudPlatform/kubernetes/tree/master/docs/getting-started-guides

Concepts Clusters Pods Controllers Services Labels Compute resources that run Kubernetes Master and Nodes Starting point for Kubernetes deployments Colocated containers with shared volumes Destination for tightly coupled components Smallest deployable unit Manage Pod lifecycle Replication ensures N Pods are always running Useful for web frontends or multiple workers Single, stable name, and address for Pods Provide proxy load- balanced connectivity to ephemeral Pods Organize and select groups of objects based on key-value pairs

Node Kubernetes Cluster Master etcd Scheduler API Server Manager ... kubelet Proxy Docker Node kubelet Proxy Docker Docker Pod Container Pod Container “zoom in and enhance”

Pods • Group of containers and volumes • Tightly coupled • Smallest unit of scheduling/placement in Kubernetes • Shared namespace • Share IP address and localhost • Share IPC • Ephemeral • Pods can die, they don’t come back • Need more? Replicate! Services • A group of Pods that work together • Defines Pod access policy • Provides a stable virtual IP and port • And a DNS name too • Virtual IP handled by Proxy • Watches and updates based on Pod lifecycle • Hides the networking complexity

Networking Pod IPs are routable Pods can reach each other without NAT, even across Nodes No complex brokering of port numbers

Kubernetes Networking Master Node Node Node Load Balancer Incoming requests for service:port

Setting Up a Cluster 1. Choose the infrastructure: GCE, AWS, Azure, Rackspace, on-premise 2. Choose the node OS: CoreOS, Atomic, RHEL, Debian, CentOS, Ubuntu 3. Provision machines: Boot VMs, install and run kube components 4. Configure networking: IP ranges for Pods, Services, SDN 5. Start cluster services: DNS, logging, monitoring 6. Manage nodes: kernel upgrades, OS updates, hardware failures Not the most fun but unavoidable to get a cluster up and running.

Google Container Engine Hosted Kubernetes with managed resources Run clusters on a bundle of Google Compute Engine resources: Instances, Disks, Networking, Load Balancer Built-in support for centralized logging and container health checking Private container registry to store and access your images at gcr.io

Using Kubernetes • Run Pods & Containers • Replication controllers • Services • Volumes • Secrets • A distinct set of problems from cluster setup and management • Accelerate development by focusing on the applications, not the cluster • Ops ensures processes for development, test, staging, production

Availability

What Else Is Coming? Network plugins Secrets Graceful termination Quota More volumes Downward API More platforms Performance Scalability High availability masters Scheduling Cluster federation Multi-cloud Easier setup

Move away from using tools to provision, orchestrate, configure environments Create Docker image build and test pipelines Leverage tools like Packer to create “golden images” The same containers can be used in development, test, staging, production The same deployment tools can be used regardless of infrastructure How Does Kubernetes Impact DevOps?

More and more examples with different patterns and components More documentation (always!) More in-depth and expanded tutorials and articles Complex, hybrid deployment examples How Can We Help?

What’s faster, booting 160 containers on GKE or making a latte?

Kubernetes kubernetes.io github.com/GoogleCloudPlatform/kubernetes Google Container Engine cloud.google.com/container-engine cloud.google.com/container-engine/docs/hello-wordpress cloud.google.com/container-engine/docs/guestbook Getting Started

Zach Jordan zachjordan@google.com Sandeep Parikh parikhs@google.com Contact Us