Kubernetes

From Beginner to Expert

@kubernetesio @briandorsey @jeffmendoza

Advanced Concepts

@kubernetesio @briandorsey @jeffmendoza

Jobs

Run-to-completion, as opposed to run-forever

• Express parallelism vs. required completions
• Workflow: restart on failure
• Build/test: don’t restart on failure

​

Aggregates success/failure counts

​

Built for batch and big-data work

​

Status: GA in Kubernetes v1.2

...

@kubernetesio @briandorsey @jeffmendoza

Lab (Setup)

Cloud: advanced/README.md

​

Local: advanced/local.md

​

@kubernetesio @briandorsey @jeffmendoza

Pods

Pod

Lobsters

Log Grabber

Volume

Consumers

Log Aggregator

Can be used to group multiple containers & shared volumes

​

Containers within a pod are tightly coupled

​

Shared namespaces

• Containers in a pod share IP, port and IPC namespaces
• Containers in a pod talk to each other through localhost

@kubernetesio @briandorsey @jeffmendoza

Lab (Pod Patterns / Sidecar)

Cloud: advanced/README.md

​

Local: advanced/local.md

​

@kubernetesio @briandorsey @jeffmendoza

Generalized Labels (1.2)

​

​

​

Pod

​

frontend

Pod

​

frontend

env = qa

env = test

• env = prod
• tier != backend
• env = prod, tier !=backend

Expressions

• env in (test,qa)
• release notin (stable,beta)
• tier
• !tier

env = prod

env = prod

Dashboard

​

​

selector:

env = notin(prod)

@kubernetesio @briandorsey @jeffmendoza

ReplicaSets (1.2)

​

​

​

Replication Controller

​

Pod

​

frontend

Pod

​

frontend

app = demo

app = demo

app = demo

Similar to ReplicationController but supports generalized Selectors

selector:

matchLabels:

app: demo

matchExpressions:

- {key: color, operator: In, values: [blue,grey]}

color = blue

color = blue

color = grey

@kubernetesio @briandorsey @jeffmendoza

Deployments: Updates as a Service

Reliable mechanism for creating, updating and managing Pods

Deployment manages replica changes, including rolling updates and scaling

Edit Deployment configurations in place with kubectl edit or kubectl apply

Managed rollouts and rollbacks

Status: BETA in Kubernetes v1.2

...

@kubernetesio @briandorsey @jeffmendoza

Deployment Based Rolling Update

API

​

​

​

Pod

​

frontend

version = v1

version = v2

version= v1

type = BE

type = BE

version = v2

type = BE

@kubernetesio @briandorsey @jeffmendoza

Canary

Service

​

Label selectors:

version = 1.0

type = Frontend

​

​

​

Service

name = backend

​

Label selector:

type = BE

Replication Controller

​

Pod

​

frontend

version= v1

version = v1

type = BE

type = BE

Replication Controller

​

​

​

​

Pod

​

frontend

version = v2

type = BE

@kubernetesio @briandorsey @jeffmendoza

Lab (Canary / Update / Rollback)

Cloud: advanced/README.md

​

Local: advanced/local.md

​

@kubernetesio @briandorsey @jeffmendoza

Autoscale

​

$ kubectl autoscale rc frontend-rc-v1 --min=1 --max=20

@kubernetesio @briandorsey @jeffmendoza

Pod Horizontal Autoscaling

​

​

​

Replication Controller

​

Pod

​

frontend

name=locust

name=locust

​

​

​

Pod

​

frontend

name=locust

Pod

​

​

name=locust

Scale

CPU Target% = 50

Heapster

role=worker

role=worker

role=worker

role=worker

70% CPU

40% CPU

> 50% CPU

< 50% CPU

@kubernetesio @briandorsey @jeffmendoza

Lab (Auto Scale)

Cloud: advanced/README.md

​

Local: advanced/local.md

​

@kubernetesio @briandorsey @jeffmendoza

Graceful Termination

Goal: Give pods time to clean up

• finish in-flight operations
• log state
• flush to disk
• 30 seconds by default

​

Catch SIGTERM, cleanup, exit ASAP

​

Pod status “Terminating”

​

Declarative: ‘DELETE’ appears as an object field in the API

​

@kubernetesio @briandorsey @jeffmendoza