Core Ops Tutorial

Pre-requisites

Build etcdctl and fleetctl:

go get github.com/coreos/etcd/etcdctl
go get github.com/coreos/fleet/fleetctl

Build kubecfg:

git clone https://github.com/GoogleCloudPlatform/kubernetes
make
cp ./_output/local/bin/darwin/amd64/kubecfg ~/bin

Google Cloud SDK: https://cloud.google.com/sdk/

Provisioning CoreOS, etcd, fleet, and flannel

Control Node

gcloud compute instances create control \
--image-project coreos-cloud \
--image coreos-alpha-494-0-0-v20141108 \
--boot-disk-size 200GB \
--machine-type n1-standard-1 \
--can-ip-forward \
--scopes compute-rw \
--metadata-from-file user-data=control.yaml \
--zone us-central1-a

Setup Client SSH Tunnels

gcloud compute instances list

etcd

ssh -f -nNT -L 4001:127.0.0.1:4001 core@<control-external-ip>

kubernetes

ssh -f -nNT -L 8080:127.0.0.1:8080 core@<control-external-ip>

fleet

export FLEETCTL_TUNNEL="<control-external-ip>"

Update configs

sed -i -e 's/CONTROL-NODE-INTERNAL-IP/<control-internal-ip>/g' *.{json,yaml,service}

Cluster Configuration

flannel

etcdctl --no-sync set /coreos.com/network/config '{"Network":"10.244.0.0/16"}'

Worker Nodes

gcloud compute instances create node1 node2 node3 node4 node5 \
--image-project coreos-cloud \
--image coreos-alpha-494-0-0-v20141108 \
--boot-disk-size 200GB \
--machine-type n1-standard-1 \
--can-ip-forward \
--metadata-from-file user-data=node.yaml \
--zone us-central1-a

gcloud compute instances list

fleetctl list-machines

---

Cross container networking with flannel

View the subnet allocations in etcd

etcdctl --no-sync ls / --recursive

Communicate between two containers

Terminal 1

gcloud compute ssh core@node1

docker run -t -i busybox /bin/sh -c 'ifconfig eth0 && nc -l -p 80'

Terminal 2

gcloud compute ssh core@node2

Replace eth0-ip with the ip address from above.

docker run -t -i busybox /bin/sh -c 'nc <eth0-ip>:80'

---

Logging with logentries.com

etcdctl --no-sync set /logentries.com/token <token>

fleetctl start journal-2-logentries.service

fleetctl list-units

---

Sysadmin tools with Toolbox

Terminal 1

gcloud compute ssh core@node1

/usr/bin/toolbox

yum install tcpdump

ip addr show flannel0

tcpdump -i flannel0

Terminal 2

gcloud compute ssh core@node2

ping <node1-flannel0-ip>

---

Kubernetes

Installing Kubernetes with fleet

fleetctl start kube-kubelet.service 
fleetctl start kube-proxy.service
fleetctl start kube-apiserver.service
fleetctl start kube-controller-manager.service
fleetctl start kube-scheduler.service
fleetctl start kube-register.service

fleetctl list-units

kubecfg list minions

Deploying applications

Creating a replicationController

cat hello-stable-controller.json

kubecfg -c hello-stable-controller.json create replicationControllers

kubecfg list replicationControllers
kubecfg list pods

Horizontally scaling pods

Edit: hello-stable-controller.json

"replicas": 4

Update the replication controller

kubecfg -c hello-stable-controller.json update replicationControllers/helloStableController

kubecfg list pods

Creating and managing services

cat hello-service.json

kubecfg -c hello-service.json create services

gcloud compute firewall-rules create default-allow-hello --allow tcp:80

gcloud compute instances list

Rolling Updates

Send the Canary

cat hello-canary-controller.json 

kubecfg -c hello-canary-controller.json create replicationControllers

kubecfg list pods

gcloud compute instances list

while true; do curl http://<node-external-ip>; echo; sleep 2; done

Rolling Update

Terminal 1

while true; do curl http://<node-external-ip>; echo; sleep 2; done

Terminal 2

kubecfg --image "quay.io/kelseyhightower/hello:2.0.0" rollingupdate helloStableController

Terminal 3

watch kubecfg list pods