Changing the way we Think and Talk about Computing Kubernetes

@tekgrrl #kubernetes #containercamp Mandy Waite Developer Advocate +MandyWaite @tekgrrl

Image by Connie Zhou

@tekgrrl #kubernetes #containercamp job hello_world = { runtime = { cell = 'ic' } // Cell (cluster) to run in binary = '.../hello_world_webserver' // Program to run args = { port = '%port%' } // Command line parameters requirements = { // Resource requirements ram = 100M disk = 100M cpu = 0.1 } replicas = 5 // Number of tasks } 10000 Developer View

@tekgrrl #kubernetes #containercamp Developer View

@tekgrrl #kubernetes #containercamp web browsers BorgMaster link shard UI shard BorgMaster link shard UI shard BorgMaster link shard UI shard BorgMaster link shard UI shard Scheduler borgcfg web browsers scheduler Borglet Borglet Borglet Borglet Config file BorgMaster link shard UI shard persistent store (Paxos) Binary What just happened? Cell Storage Developer View

Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Image by Connie Zhou Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world! Hello world!

#dockerjp #kubernetes @tekgrrl Developer View

#dockerjp #kubernetes @tekgrrl task-eviction rates and causes Failures

#dockerjp #kubernetes @tekgrrl Resource reclamation time limit: amount of resource requested usage: actual resource consumption Efficiency reservation: estimate of future usage potentially reusable resources

#dockerjp #kubernetes @tekgrrl one machine Efficiency Advanced bin- packing algorithms Experimental placement of production VM workload, July 2014

#dockerjp #kubernetes @tekgrrl Efficiency Used CPU Used CPU (in cores) Used Memory Used Memory Available Resources Stranded Resources

#dockerjp #kubernetes @tekgrrl The Machine is just a Resource Boundary Machines (Virtual and Bare Metal) have shapes: 2GB/Core, 4GB/Core, 8GB/Core, etc Workloads have shapes too: 1GB/Core, 3.5GB/Core, etc Machine Shapes Workload Shapes

#dockerjp #kubernetes @tekgrrl Computing Tetris

#dockerjp #kubernetes @tekgrrl Resource Stranding 5.5GB RAM Inaccessible CPU Fully Utilized

#dockerjp #kubernetes @tekgrrl Bin Packing 1 CPU Available 5.5GB RAM Available

#dockerjp #kubernetes @tekgrrl Bin Packing

#dockerjp #kubernetes @tekgrrl Images by Connie Zhou Observations: ● Efficiency comes from ○ Scavenging unused allocations ○ Effective Prioritization ○ Sharing resources ○ Overcommit ○ Smarter Scheduling ● The Datacenter is one big System ○ The machine is just a resource boundary ● Containers make everyone more productive http://kubernetes.io http://goo.gl/1C4nuo (Borg paper)

Kubernetes

@tekgrrl #kubernetes #containercamp Greek for “Helmsman”; also the root of the word “Governor” • Orchestrator for Docker containers • Supports multi-cloud environments • Inspired and informed by Google’s experiences and internal systems • Open source, written in Go Manage applications, not machines Kubernetes

@tekgrrl #kubernetes #containercamp web browsers y Kubelet Kubelet Kubelet Kubelet Kubernetes Master Replication Controller Scheduler API Server Kube-UI Container Registry kubectl Proxy web browsers Developer View (Kubernetes)

@tekgrrl #kubernetes #containercamp Service Label selectors: version = 1.0 type = Frontend Service Label selector: type = FE Replication Controller Pod Pod frontend Pod version= v1 version = v1 Replication Controller version = v1 #pods = 2 show: version = v2 type = FE type = FE VIP Canary Example Replication Controller Replication Controller version = v2 #pods = 1 show: version = v2 Pod frontend Pod version = v2 type = FE

@tekgrrl #kubernetes #containercamp Developer View (Replication Controller) spec: containers: - name: php-guestbook image: php-guestbook:containercamp resources: limits: memory: "128Mi" cpu: "500m" ports: - containerPort: 80 protocol: TCP replicas: 1 10000

@tekgrrl #kubernetes #containercamp php MySQL php php python memcached d Mapping to Kubernetes Client

@tekgrrl #kubernetes #containercamp Node3 Kubelet Proxy Pod Container Container Container Container Pod Container Container Container Container Node3 Kubelet Proxy Pod Container Container Container Container Pod Container Container Container Container Node1 Kubelet Proxy Pod Container Container Pod $ kubectl proxy --www=k8s-visualizer/ Visualizing Kubernetes Master APIs Scheduling REST (pods, services, controllers) AuthN Scheduler Replication Controller Container

Demo - Visualization

@tekgrrl #kubernetes #containercamp A Quick Guide to Nodes Node2 Kubelet Proxy disk = ssd type = FE Resources Labels Disks

@tekgrrl #kubernetes #containercamp Scheduling Capabilities Step 1 - Filtering Node Resources I need 3.5GB RAM and 2 Cores Node Resources Node Resources Node Resources podFitsResources

@tekgrrl #kubernetes #containercamp Scheduling Capabilities Step 1 - Filtering Node1 Kubelet Proxy Node2 Kubelet Proxy Node3 Kubelet Proxy Demo-mysql-disk I need demo- mysql-disk NoDiskConflict

@tekgrrl #kubernetes #containercamp PodFitsHost: Filter by NodeName PodSelectorMatches: Filter by Node Label (key and value) CheckNodeLabelPresence: Filter by Node Label Scheduling Capabilities Step 1 - Filtering

@tekgrrl #kubernetes #containercamp LeastRequestedPriority Rank by how much free space will be left on the node after the node is deployed [Default] CalculateNodeLabelPriority Prefer nodes with the specified label BalancedResourceAllocation CPU and Memory is balanced after the Pod is deployed [Default] CalculateSpreadPriority minimise number of Pods from the same service on the same node CalculateAntiAffinityPriority minimise number of Pods from the same service on nodes with the same value for a particular label Scheduling Capabilities Step 2 - Ranking

@tekgrrl #kubernetes #containercamp Kubernetes 1.0 as of mid July • Formerly announced at OSCON in July Open sourced in June, 2014 • won the BlackDuck “rookie of the year” award Google launched Google Container Engine (GKE) • hosted Kubernetes • https://cloud.google.com/container-engine/ Roadmap: • https://github.com/GoogleCloudPlatform/kubernetes/milestones Kubernetes Status

@tekgrrl #kubernetes #containercamp Google Container Engine (GA) Managed Kubernetes (Kubernetes v1) Manages Kubernetes master uptime Manages Updates Cluster Resize via Managed Instance Groups Centralised Logging Google Cloud VPN support

@tekgrrl #kubernetes #containercamp Kubernetes is Open Source We want your help! http://kubernetes.io https://github.com/GoogleCloudPlatform/kubernetes irc.freenode.net #google-containers @kubernetesio

@tekgrrl #kubernetes #containercamp Tweet questions to: @tekgrrl Questions

@tekgrrl #kubernetes #containercamp Example of Resource based scheduling Node1 (1 CPU) Kubelet Proxy Pod (0.6 CPU) Container Container Node2 (1 CPU) Kubelet Proxy Pod (0.6 CPU) Container Container Node3 (1 CPU) Kubelet Proxy Pod (0.6 CPU) Container Container