Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Real-World Kubernetes Deployments @ OSCON 2016
Search
Brandon Philips
May 16, 2016
Programming
1
630
Real-World Kubernetes Deployments @ OSCON 2016
Brandon Philips
May 16, 2016
Tweet
Share
More Decks by Brandon Philips
See All by Brandon Philips
Node.js Workflow with Minikube and Skaffold
philips
0
220
Manage the App on Kubernetes
philips
0
310
Production Backbone Monitoring Containerized Apps
philips
0
120
KubeCon EU 2017: Dancing on the Edge of a Volcano
philips
1
620
rkt - KubeCon EU keynote - 2017
philips
1
240
FOSDEM_Keynote_2017-_.pdf
philips
0
90
Tectonic Summit Day 2 Keynote
philips
0
310
Kubernetes: Simple to Manage Anywhere (self-hosted, Tectonic upgrade demo)
philips
0
280
KubeCon Keynote 2016- Distributed Systems Simplified on Kubernetes
philips
2
530
Other Decks in Programming
See All in Programming
型付き API リクエストを実現するいくつかの手法とその選択 / Typed API Request
euxn23
8
2.2k
Webの技術スタックで マルチプラットフォームアプリ開発を可能にするElixirDesktopの紹介
thehaigo
2
1k
RubyLSPのマルチバイト文字対応
notfounds
0
120
Why Jakarta EE Matters to Spring - and Vice Versa
ivargrimstad
0
1.1k
Kaigi on Rails 2024 〜運営の裏側〜
krpk1900
1
230
watsonx.ai Dojo #4 生成AIを使ったアプリ開発、応用編
oniak3ibm
PRO
1
140
役立つログに取り組もう
irof
28
9.6k
タクシーアプリ『GO』のリアルタイムデータ分析基盤における機械学習サービスの活用
mot_techtalk
4
1.4k
2024/11/8 関西Kaggler会 2024 #3 / Kaggle Kernel で Gemma 2 × vLLM を動かす。
kohecchi
5
930
『ドメイン駆動設計をはじめよう』のモデリングアプローチ
masuda220
PRO
8
540
AI時代におけるSRE、 あるいはエンジニアの生存戦略
pyama86
6
1.2k
Macとオーディオ再生 2024/11/02
yusukeito
0
370
Featured
See All Featured
Facilitating Awesome Meetings
lara
50
6.1k
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
29
2.3k
VelocityConf: Rendering Performance Case Studies
addyosmani
325
24k
Building Your Own Lightsaber
phodgson
103
6.1k
Being A Developer After 40
akosma
87
590k
Build The Right Thing And Hit Your Dates
maggiecrowley
33
2.4k
Optimizing for Happiness
mojombo
376
70k
Statistics for Hackers
jakevdp
796
220k
Designing on Purpose - Digital PM Summit 2013
jponch
115
7k
Making the Leap to Tech Lead
cromwellryan
133
8.9k
YesSQL, Process and Tooling at Scale
rocio
169
14k
GraphQLの誤解/rethinking-graphql
sonatard
67
10k
Transcript
Real World Kubernetes Deployments failure domains, upgrades, high-availability @coreoslinux @brandonphilips
Follow Along Instructions http://bit.ly/1XeUbMW Stickers Upfront Decorate your laptop, dog,
kid, phone.
Brandon Philips CTO, CoreOS github.com/philips
None
Build, Store and Distribute your Containers quay.io
Linux
Secure the Internet MISSION
Separate Apps from OS STRATEGY
Make Servers Consistent STRATEGY
Tolerate Machine Failures STRATEGY
Make Servers Easy to Upgrade STRATEGY
Simplify Application Upgrades STRATEGY
None
None
None
None
None
Application Packaging 1
Abstract away app from the OS OS App
None
None
Linux at Scale 2
Patches to the OS and kernel are hard Retest after
updates No automation SECURITY Dependency breakage Uptime risk APPLICATION
None
Auto-updating browsers fixed security We got HTML5 at the same
time
Clustering 3
Operations Paradise Easy scale out Painless app upgrades Tolerant of
machine failure
App Req/sec: 6,000 App Healthy: True
App Req/sec: 6,000 App Healthy: True
App Req/sec: 7,000 App Healthy: True
App Req/sec: 8,000 App Healthy: True
App Req/sec: 7,000 App Healthy: True
App Req/sec: 6,000 App Healthy: True
App Req/sec: 8,000 App Healthy: True
App Req/sec: 7,000 App Healthy: True
App Req/sec: 8,000 App Healthy: True
App Req/sec: 8,000 App Healthy: True
3 Application packaging Clustering Linux at scale
3 Application packaging Clustering Linux at scale
Follow Along Instructions https://github.com/philips/repositories 2016-OSCON-containers-at-scale
CoreOS+Kubernetes vagrant, aws, bare metal, etc coreos.com/kubernetes/docs/latest/
kubernetes architecture in practice
worker kubelet worker kubelet worker kubelet scheduler & API worker
kubelet w ku t worker kubelet
worker kubelet worker kubelet scheduler & API
worker & API works on 1 node too
kube-aws Initial Cluster Setup
worker kubelet worker kubelet controller scheduler, etcd & API
Demo Boot up a Cluster
Demo Run an App
Demo Understand the Network
Domains Let's Talk About Failure
Failure domains are regions or components of the infrastructure which
contain a potential for failure.
These regions can be physical or logical boundaries, and each
has its own risks and challenges to architect for.
Failure Feud - Machine Failure - Network/Disks/RAM/Processor/Power Supply - Rack
Failure - Network/Power - Data Center Failure - Network/Power/Fire/Semi-trucks - Internet Failure - Network/Political/Natural
Failure Analysis Kid Celebrating
None
Kid Hitting His Eye Failure Analysis - Failure is caused
by human error - Celebration continues; eye unnecessary - Kid has two eyes can continue seeing - Brain elects new eye automatically
Primary Datastore etcd operations
/etc distributed hence, the name...
a clustered key-value store GET and SET operations
a building block for higher order systems primitives for building
reliable distributed systems
Demo play.etcd.io
None
None
None
None
None
None
None
Failure Analysis etcd
worker kubelet worker kubelet scheduler & API
kube-aws high availability in cloud
scheduler & API EBS { ASG
etcd protects against - Machine Failure - Replication, automatic leader
election - Flakey Disk Failure - CRC checksums on WAL files - Network Failure - Timeouts and linearized state machine
etcd does not protect against - Denial of Service -
Future work on proxies - Lying etcd Peers - We do a ton of functional testing a hedge - Buggy or Broken Clients - Client deleting all keys requires restore from backup
Demo etcd restore backup
1 2 3 4 { Log
1 2 3 4 Entries
1 2 3 4 Indexes
Kubernetes Control API Service, Scheduler, Controller Manager
Failure Analysis Kubernetes
Demo etcd down for API server
worker kubelet worker kubelet scheduler & API
scheduler & API
Demo etcd restore for API server
scheduler & API
Demo node partition from API
worker kubelet worker kubelet scheduler & API
Demo node scaling up
worker kubelet worker kubelet scheduler & API worker kubelet
Demo node scheduled outage API
worker kubelet worker kubelet scheduler & API
Demo node unplanned outage
worker kubelet worker kubelet scheduler & API
Demo node downgrade/upgrade outage
worker kubelet worker kubelet scheduler & API
Future Work Upstream Kubernetes and Elsewhere
Upstream rktnetes Auth/OIDC Node self-signed TLS
Scaling 15x scheduler performance 30k pods on 1k nodes SIG-scale
Automatic Node Drain Locksmith Design Doc
None
Performance etcd3 /ZooKeeper snapshot disabled
Performance etcd3 /ZooKeeper snapshot disabled
Memory 10GB 2.4GB 0.8GB 512MB data - 2M 256B keys
Sounds good, but... Is anyone successful with all this in
prod?
Publically traded options exchange
Containers on CoreOS are powering ISE's high- throughput, low-latency financial
exchange Running in production Bare metal & AWS Billions of transactions a day 150 million req/sec
None
Secure the Internet MISSION
Separate Apps from OS STRATEGY
Make Servers Consistent STRATEGY
Tolerate Machine Failures STRATEGY
Make Servers Easy to Upgrade STRATEGY
Simplify Application Upgrades STRATEGY
None
Thank you! Brandon Philips @brandonphilips |
[email protected]
| coreos.com We’re
hiring in all departments! Email:
[email protected]
Positions: coreos.com/ careers