Post on 15-Apr-2017
CoreOS in a Nutshell
Alex CrawfordSoftware Developer at CoreOS
alex.crawford@coreos.comgithub.com/crawford
What is CoreOS?
What is CoreOS?
What is CoreOS?
What is CoreOS?
The smartest way to run your container infrastructure.
tectonic.com @tectonic
QUAYSecure hosting for private container repositories
quay.io @quayio
Why build CoreOS?
you
you as a sw engineer
your
with Ada.Text_IO; procedure Hello_World is use Ada.Text_IO;begin Put_Line("Hello, world!");end;
#include <stdio.h> int main(){ printf("Hello, world!\n");}
package main import "fmt" func main() {
fmt.Println("Hello, world!")}
your containerimage
your/bin/java
/opt/app.jar
/lib/libc
your /bin/python/opt/app.py/lib/libc
your com.example.app
d474e8c57737625c
your d474e8c57737625c
Signed By: Alice
ops engineer
you as an ops engineer
your
your
com.example.webappx3
your
com.example.webappx3
your
???
com.example.webappx3
How do we do it?
reduce API contracts
minimal
kernelsystemdrktsshdocker
pythonjavanginxmysqlopenssl
app
dist
ro d
istro
dist
ro d
istro
dist
ro
dist
ro
pythonjavanginxmysqlopenssl
app d
istro
dist
ro d
istro
dist
ro d
istro
di
stro
kernelsystemdrktsshdocker
pythonopenssl-A app1
dist
ro d
istro
dist
ro d
istro
dist
ro
dist
ro java
openssl-B app2
javaopenssl-B app3
kernelsystemdrktsshdocker
pythonopenssl-A app1
dist
ro d
istro
dist
ro d
istro
dist
ro
dist
ro java
openssl-B app2
javaopenssl-B app3
CoreOS
CoreOS
container
dist
ro d
istro
dist
ro d
istro
dist
ro
dist
ro
container
container
OS operations
updates
OS operations
manual updates
automatic updates
automatic updates
atomic update with rollback
CoreOS Updates
machine configuration
OS operations
get into the cluster
machine config
[Service]ExecStart=/usr/bin/kubelet --api_servers=https://172.17.4.101 --register-node=true --hostname-override=172.17.4.201 --cluster_dns=10.3.0.10 --cluster_domain=cluster.local --tls-cert-file=worker.pem --tls-private-key-file=worker-key.pem
[Service]ExecStart=/usr/bin/kubelet --api_servers=https://172.17.4.101 --register-node=true --hostname-override=172.17.4.201 --cluster_dns=10.3.0.10 --cluster_domain=cluster.local --tls-cert-file=worker.pem --tls-private-key-file=worker-key.pem
[Service]ExecStart=/usr/bin/kubelet --api_servers=https://172.17.4.101 --register-node=true --hostname-override=172.17.4.201 --cluster_dns=10.3.0.10 --cluster_domain=cluster.local --tls-cert-file=worker.pem --tls-private-key-file=worker-key.pem
distributed configuration
cluster operations
etcd
/etcdistributed
AvailableLeader
Follower
AvailableLeader
Follower
AvailableLeader
Follower
UnavailableLeader
Follower
AvailableLeader
Follower
AvailableLeader
Follower
Temporarily Unavailable
Leader
Follower
Available
Leader
Follower
Unavailable
Leader
Follower
what should run
cluster operations
k8s/mesos/etc scheduler
scheduling
getting work to servers
scheduling
You
You
Scheduler API
You
Scheduler API
Scheduler
You
Scheduler API
Scheduler
Machine(s)
while true { todo = diff(desState, curState) schedule(todo)}
while true { todo = diff(desState, curState) schedule(todo)}
while true { todo = diff(desState, curState) schedule(todo)}
while true { todo = diff(desState, curState) schedule(todo)}
$ kubectl run example--image=quay.io/crawford/example--replicas=1
$ kubectl get podsPOD IPexample-97wt8 10.2.29.4
$ kubectl run example--image=quay.io/crawford/example--replicas=1
$ kubectl get podsPOD IPexample-97wt8 10.2.29.4
$ kubectl run example--image=quay.io/crawford/example--replicas=1
$ kubectl get podsPOD IPexample-97wt8 10.2.29.4
$ kubectl run example--image=quay.io/crawford/example--replicas=1
$ kubectl get podsPOD IPexample-97wt8 10.2.29.4
$ kubectl run example--image=quay.io/crawford/example--replicas=1
$ kubectl get podsPOD IPexample-97wt8 10.2.29.4
$ kubectl run example--image=quay.io/crawford/example--replicas=1
$ kubectl get podsPOD IPexample-97wt8 10.2.29.4
$ kubectl scale rc example--replicas=2
$ kubectl get podsPOD IPexample-97wt8 10.2.29.4example-f839d 10.2.29.8
$ kubectl scale rc example--replicas=2
$ kubectl get podsPOD IPexample-97wt8 10.2.29.4example-f839d 10.2.29.8
$ kubectl scale rc example--replicas=2
$ kubectl get podsPOD IPexample-97wt8 10.2.29.4example-f839d 10.2.29.8
$ kubectl scale rc example--replicas=2
$ kubectl get podsPOD IPexample-97wt8 10.2.29.4example-f839d 10.2.29.8
$ kubectl scale rc example--replicas=2
$ kubectl get podsPOD IPexample-97wt8 10.2.29.4example-f839d 10.2.29.8
$ kubectl scale rc example--replicas=2
$ kubectl get podsPOD IPexample-97wt8 10.2.29.4example-f839d 10.2.29.8
podenv=prodapp=web
podenv=prodapp=web
podenv=prodapp=web
rc web-prodselect(env=prod,app=web)count=3
podenv=prodapp=web
podenv=prodapp=web
podenv=prodapp=web
rc web-prodselect(env=prod,app=web)count=1
podenv=prodapp=web
rc web-prodselect(env=prod,app=web)count=1
podenv=prodapp=web
rc web-prodselect(env=prod,app=web)count=5
podenv=prodapp=web
podenv=prodapp=web
podenv=prodapp=web
podenv=prodapp=web
podenv=prodapp=web
rc web-prodselect(env=prod,app=web)count=5
where is it running
cluster operations
dns, LBs, k8s labels
services
flexible service discovery
k8s labels
podenv=devapp=web
podenv=testapp=web
podenv=prodapp=web
podenv=devapp=web
podenv=testapp=web
podenv=prodapp=web
service test.example.comselect(env=dev,app=web)
service beta.example.comselect(env=test,app=web)
ORselect(env=prod,app=web)
service example.comselect(env=prod,app=web)
podenv=testapp=web
podenv=prodapp=web
podenv=prodapp=web
podenv=devapp=web
podenv=testapp=web
podenv=prodapp=web
service test.example.comselect(env=dev,app=web)
service beta.example.comselect(env=test,app=web)
ORselect(env=prod,app=web)
service example.comselect(env=prod,app=web)
podapp=foo,version=1
service foo.cluster.localselect(app=foo)
podapp=foo,version=1
podapp=foo,version=2
service foo.cluster.localselect(app=foo)
podapp=foo,version=1
podapp=foo,version=2
service foo.cluster.localselect(app=foo)
podapp=foo,version=1
podapp=foo,version=2
service foo.cluster.localselect(app=foo)
podapp=foo,version=2
service foo.cluster.localselect(app=foo)
coreos.com/careers
work with us
Questions?