guestbook-go/

directory
v0.16.1 Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: May 1, 2015 License: Apache-2.0

README

GuestBook example

This example shows how to build a simple multi-tier web application using Kubernetes and Docker. It consists of a web frontend, a redis master for storage and a replicated set of redis slaves.

Step Zero: Prerequisites

This example assumes that you have forked the repository and turned up a Kubernetes cluster:

$ cd kubernetes
$ hack/dev-build-and-up.sh

Step One: Turn up the redis master.

Use the file examples/guestbook-go/redis-master-controller.json to create a replication controller which manages a single pod. The pod runs a redis key-value server in a container. Using a replication controller is the preferred way to launch long-running pods, even for 1 replica, so the pod will benefit from self-healing mechanism in kubernetes.

Create the redis master replication controller in your Kubernetes cluster using the kubectl CLI:

$ cluster/kubectl.sh create -f examples/guestbook-go/redis-master-controller.json

Once that's up you can list the replication controllers in the cluster:

$ cluster/kubectl.sh get rc
CONTROLLER                             CONTAINER(S)            IMAGE(S)                            SELECTOR                     REPLICAS
redis-master-controller                redis-master            gurpartap/redis                     name=redis,role=master       1

List pods in cluster to verify the master is running. You'll see a single redis master pod. It will also display the machine that the pod is running on once it gets placed (may take up to thirty seconds).

$ cluster/kubectl.sh get pods
POD                      IP                  CONTAINER(S)        IMAGE(S)            HOST                                                             LABELS                   STATUS
redis-master-pod-hh2gd   10.244.3.7          redis-master        gurpartap/redis     kubernetes-minion-4.c.lucid-walker-725.internal/104.154.49.134   name=redis,role=master   Running

If you ssh to that machine, you can run docker ps to see the actual pod:

me@workstation$ gcloud compute ssh --zone us-central1-b kubernetes-minion-4

me@kubernetes-minion-3:~$ sudo docker ps
CONTAINER ID        IMAGE                                  COMMAND                CREATED             STATUS
d5c458dabe50        gurpartap/redis:latest                 "/usr/local/bin/redi   5 minutes ago       Up 5 minutes

(Note that initial docker pull may take a few minutes, depending on network conditions.)

Step Two: Turn up the master service.

A Kubernetes 'service' is a named load balancer that proxies traffic to one or more containers. The services in a Kubernetes cluster are discoverable inside other containers via environment variables or DNS. Services find the containers to load balance based on pod labels.

The pod that you created in Step One has the label name=redis and role=master. The selector field of the service determines which pods will receive the traffic sent to the service. Use the file examples/guestbook-go/redis-master-service.json to create the service in the kubectl cli:

$ cluster/kubectl.sh create -f examples/guestbook-go/redis-master-service.json

$ cluster/kubectl.sh get services
NAME                    LABELS                                    SELECTOR                     IP                  PORT
redis-master            <none>                                    name=redis,role=master       10.0.186.234        6379

This will cause all new pods to see the redis master apparently running on $REDIS_MASTER_SERVICE_HOST at port 6379, or running on 'redis-master:6379'. Once created, the service proxy on each node is configured to set up a proxy on the specified port (in this case port 6379).

Step Three: Turn up the replicated slave pods.

Although the redis master is a single pod, the redis read slaves are a 'replicated' pod. In Kubernetes, a replication controller is responsible for managing multiple instances of a replicated pod.

Use the file examples/guestbook-go/redis-slave-controller.json to create the replication controller:

$ cluster/kubectl.sh create -f examples/guestbook-go/redis-slave-controller.json

$ cluster/kubectl.sh get rc
CONTROLLER                             CONTAINER(S)            IMAGE(S)                            SELECTOR                     REPLICAS
redis-master-controller                redis-master            gurpartap/redis                     name=redis,role=master       1
redis-slave-controller                 redis-slave             gurpartap/redis                     name=redis,role=slave        2

The redis slave configures itself by looking for the redis-master service name:port pair. In particular, the redis slave is started with the following command:

redis-server --slaveof redis-master 6379

Once that's up you can list the pods in the cluster, to verify that the master and slaves are running:

$ cluster/kubectl.sh get pods
POD                                          IP                  CONTAINER(S)            IMAGE(S)                            HOST                                                             LABELS                                   STATUS
redis-master-pod-hh2gd                       10.244.3.7          redis-master            gurpartap/redis                     kubernetes-minion-4.c.lucid-walker-725.internal/104.154.49.134   name=redis,role=master                   Running
redis-slave-controller-i7hvs                 10.244.2.7          redis-slave             gurpartap/redis                     kubernetes-minion-3.c.lucid-walker-725.internal/104.154.52.39    name=redis,role=slave                    Running
redis-slave-controller-nyxxv                 10.244.1.6          redis-slave             gurpartap/redis                     kubernetes-minion-2.c.lucid-walker-725.internal/130.211.144.5    name=redis,role=slave                    Running

You will see a single redis master pod and two redis slave pods.

Step Four: Create the redis slave service.

Just like the master, we want to have a service to proxy connections to the read slaves. In this case, in addition to discovery, the slave service provides transparent load balancing to clients. The service specification for the slaves is in examples/guestbook-go/redis-slave-service.json

This time the selector for the service is name=redis,role=slave, because that identifies the pods running redis slaves. It may also be helpful to set labels on your service itself--as we've done here--to make it easy to locate them later.

Now that you have created the service specification, create it in your cluster with the kubectl CLI:

$ cluster/kubectl.sh create -f examples/guestbook-go/redis-slave-service.json

$ cluster/kubectl.sh get services
NAME                    LABELS                                    SELECTOR                     IP                  PORT
redis-master            <none>                                    name=redis,role=master       10.0.186.234        6379
redis-slave             name=redis,role=slave                     name=redis,role=slave        10.0.22.180         6379

Step Five: Create the guestbook pod.

This is a simple Go net/http (negroni based) server that is configured to talk to either the slave or master services depending on whether the request is a read or a write. It exposes a simple JSON interface, and serves a jQuery-Ajax based UX. Like the redis read slaves it is a replicated service instantiated by a replication controller.

The pod is described in the file examples/guestbook-go/guestbook-controller.json. Using this file, you can turn up your guestbook with:

$ cluster/kubectl.sh create -f examples/guestbook-go/guestbook-controller.json

$ cluster/kubectl.sh get replicationControllers
CONTROLLER                             CONTAINER(S)            IMAGE(S)                            SELECTOR                     REPLICAS
guestbook-controller                   guestbook               kubernetes/guestbook:v2             name=guestbook               3
redis-master-controller                redis-master            gurpartap/redis                     name=redis,role=master       1
redis-slave-controller                 redis-slave             gurpartap/redis                     name=redis,role=slave        2

Once that's up (it may take ten to thirty seconds to create the pods) you can list the pods in the cluster, to verify that the master, slaves and guestbook frontends are running:

$ cluster/kubectl.sh get pods
POD                                          IP                  CONTAINER(S)            IMAGE(S)                            HOST                                                             LABELS                                   STATUS
guestbook-controller-182tv                   10.244.2.8          guestbook               kubernetes/guestbook:v2             kubernetes-minion-3.c.lucid-walker-725.internal/104.154.52.39    name=guestbook                           Running
guestbook-controller-jzjpe                   10.244.0.7          guestbook               kubernetes/guestbook:v2             kubernetes-minion-1.c.lucid-walker-725.internal/104.154.37.86    name=guestbook                           Running
guestbook-controller-zwk1b                   10.244.3.8          guestbook               kubernetes/guestbook:v2             kubernetes-minion-4.c.lucid-walker-725.internal/104.154.49.134   name=guestbook                           Running
redis-master-pod-hh2gd                       10.244.3.7          redis-master            gurpartap/redis                     kubernetes-minion-4.c.lucid-walker-725.internal/104.154.49.134   name=redis,role=master                   Running
redis-slave-controller-i7hvs                 10.244.2.7          redis-slave             gurpartap/redis                     kubernetes-minion-3.c.lucid-walker-725.internal/104.154.52.39    name=redis,role=slave                    Running
redis-slave-controller-nyxxv                 10.244.1.6          redis-slave             gurpartap/redis                     kubernetes-minion-2.c.lucid-walker-725.internal/130.211.144.5    name=redis,role=slave                    Running

You will see a single redis master pod, two redis slaves, and three guestbook pods.

Step Six: Create the guestbook service.

Just like the others, you want a service to group your guestbook pods. The service specification for the guestbook is in examples/guestbook-go/guestbook-service.json. There's a twist this time - because we want it to be externally visible, we set the createExternalLoadBalancer flag on the service.

$ cluster/kubectl.sh create -f examples/guestbook-go/guestbook-service.json

$ cluster/kubectl.sh get services
NAME                    LABELS                                    SELECTOR                     IP                  PORT
guestbook               <none>                                    name=guestbook               10.0.12.110         3000
redis-master            <none>                                    name=redis,role=master       10.0.186.234        6379
redis-slave             name=redis,role=slave                     name=redis,role=slave        10.0.22.180         6379

To play with the service itself, find the external IP of the load balancer:

$ cluster/kubectl.sh get services guestbook -o template --template='{{index . "publicIPs"}}'
current-context: "kubernetes-satnam_kubernetes"
Running: cluster/../cluster/gce/../../_output/dockerized/bin/linux/amd64/kubectl get services guestbook -o template --template={{index . "publicIPs"}}
[104.154.87.59]$

and then visit port 3000 of that IP address e.g. http://104.154.87.59:3000.

You may need to open the firewall for port 3000 using the console or the gcloud tool. The following command will allow traffic from any source to instances tagged kubernetes-minion:

$ gcloud compute firewall-rules create --allow=tcp:3000 --target-tags=kubernetes-minion kubernetes-minion-3000

If you are running Kubernetes locally, you can just visit http://localhost:3000 For details about limiting traffic to specific sources, see the GCE firewall documentation.

Step Seven: Cleanup

You should delete the service which will remove any associated resources that were created e.g. load balancers, forwarding rules and target pools. All the resources (replication controllers and service) can be deleted with a single command:

$ cluster/kubectl.sh delete -f examples/guestbook-go
current-context: "kubernetes-satnam_kubernetes"
Running: cluster/../cluster/gce/../../_output/dockerized/bin/linux/amd64/kubectl delete -f examples/guestbook-go
guestbook-controller
guestbook
redis-master-controller
redis-master
redis-slave-controller
redis-slave

However, the command will not delete the pods created by the replication controller. You can delete the pods manually. If you want to delete the pods together, you can use the commands below instead.

cluster/kubectl.sh stop -f examples/guestbook-go/guestbook-controller.json
cluster/kubectl.sh stop -f examples/guestbook-go/redis-slave-controller.json
cluster/kubectl.sh stop -f examples/guestbook-go/redis-master-controller.json
cluster/kubectl.sh delete -f examples/guestbook-go/guestbook-service.json
cluster/kubectl.sh delete -f examples/guestbook-go/redis-slave-service.json
cluster/kubectl.sh delete -f examples/guestbook-go/redis-master-service.json

To turn down a Kubernetes cluster:

$ cluster/kube-down.sh

Directories

Path Synopsis

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL