- kubectl
- minikube
- k8stail
- https://github.com/dtan4/k8stail/releases/download/v0.2.1/k8stail-v0.2.1-linux-amd64.zip
- https://storage.googleapis.com/kubernetes-release/release/v1.5.2/bin/linux/amd64/kubectl
- https://storage.googleapis.com/minikube/releases/v0.15.0/minikube-linux-amd64
minikube start --insecure-registry localhost:5000
eval $(minikube docker-env)
ᐅ kubectl get cs
NAME STATUS MESSAGE ERROR
scheduler Healthy ok
controller-manager Healthy ok
etcd-0 Healthy {"health": "true"}
ᐅ kubectl apply -f ./local-registry.yml
replicationcontroller "kube-registry-v0" created
service "kube-registry" created
pod "kube-registry-proxy" created
Looking at the previous output we can see we just deployed three things:
- a replicationcontroler (lets call them rc's)
- a service (like a loadbalancer)
- a pod (the actual container)
You can see the pods running in your minikube with:
k get pods
Note that we do NOT see one called kube-registry-proxy
Thats because its deployed them into a different namespace
namespaces are used to logically seperate deployed apps. If you deploy:
- app1 into namespace foo
- app2 into namespace bar
app1 and app2 will not be able to communicate with each other
There is also a special one called kube-system
This is a priv place.. essentially "ops" stuff will go here.
a quick grep namespace local-registry.yml will reveal that this got deployed into the kube-system namespace.
so lets add --namespace=kube-system to k get pods now:
ᐅ kubectl --namespace=kube-system get pods
NAME READY STATUS RESTARTS AGE
kube-addon-manager-minikube 1/1 Running 0 12m
kube-dns-v20-s51mx 3/3 Running 0 11m
kube-registry-proxy 1/1 Running 0 5m
kube-registry-v0-g16mf 1/1 Running 0 5m
kubernetes-dashboard-09s1j 1/1 Running 0 12m
kubernetes-dashboard-b4rzn 1/1 Running 1 88d
get a list of services in kube-system:
ᐅ kubectl get svc --namespace=kube-system
NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kube-dns 10.0.0.10 <none> 53/UDP,53/TCP 88d
kube-registry 10.0.0.122 <none> 5000/TCP 25m
kubernetes-dashboard 10.0.0.228 <nodes> 80:30000/TCP 88d
Lets see the the ports and details on the registry:
kubectl describe svc/kube-registry --namespace=kube-system
Name: kube-registry
Namespace: kube-system
Labels: k8s-app=kube-registry
kubernetes.io/name=KubeRegistry
Selector: k8s-app=kube-registry
Type: ClusterIP
IP: 10.0.0.122
Port: registry 5000/TCP
Endpoints: 172.17.0.7:5000
Session Affinity: None
No events.
We can "port-forward" :5000 using kubectl.. then you will have a docker registry running localy on :5000
POD=$(kubectl get pods --namespace kube-system -l k8s-app=kube-registry \
-o template --template '{{range .items}}{{.metadata.name}} {{.status.phase}}{{"\n"}}{{end}}' \
| grep Running | head -1 | cut -f1 -d' ')
kubectl port-forward --namespace kube-system $POD 5000:5000 &
verify its up:
curl localhost:5000
Handling connection for 5000
ᐅ cd sample_app_hello
ᐅ docker build .
Sending build context to Docker daemon 3.072 kB
Step 1 : FROM caarlos0/alpine-go
---> 3c5fe9414eed
Step 2 : WORKDIR /gopath/src/app
---> Using cache
---> 3e4821d29fa7
Step 3 : ADD . /gopath/src/app/
---> Using cache
---> cbd27bb09b2e
Step 4 : RUN go get -v app
---> Using cache
---> 8f70b96bac7f
Step 5 : ENTRYPOINT /gopath/bin/app
---> Using cache
---> 532abd0000aa
Successfully built 532abd0000aa
ᐅ docker tag 532abd0000aa localhost:5000/hello/1
ᐅ docker push localhost:5000/hello/1
The push refers to a repository [localhost:5000/hello/1]
bb810286635f: Mounted from hello/three
8184b846aefd: Mounted from hello/three
51a8a83b9792: Mounted from hello/three
5f70bf18a086: Mounted from hello/three
c954c0da97ee: Mounted from hello/three
745737c319fa: Mounted from hello/three
latest: digest: sha256:2052725c4f72a37fb3c4b43ae0e88b7a537f37041e2bdcdf7eb3271bdfbcb1b2 size: 2395
OK so now you should be able to deploy it!