github
/
k3s
mirror of https://github.com/k3s-io/k3s
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Flower service and updated kubectl outputs in examples/clery-rabbitmq

pull/6/head
Marcin Wielgus 2015-07-08 11:01:55 +02:00
parent 03c27916a0
commit 2bf1e50b8a
4 changed files with 75 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

51
examples/celery-rabbitmq/README.md
View File

@ -172,7 +172,33 @@ The question now is, how do you see what's going on?
Flower is a web-based tool for monitoring and administrating Celery clusters. By connecting to the node that contains Celery, you can see the behaviour of all the workers and their tasks in real-time.
To bring up the frontend, run this command `$ kubectl create -f examples/celery-rabbitmq/flower-controller.yaml`. This controller is defined as so:
First, start the flower service with `$ kubectl create -f examples/celery-rabbitmq/flower-service.yaml`. The service is defined as below:
```yaml
apiVersion: v1
kind: Service
metadata:
labels:
name: flower
name: flower-service
spec:
ports:
- port: 5555
selector:
app: taskQueue
component: flower
type: LoadBalancer
```
It is marked as external (LoadBalanced). However on many platforms you will have to add an explicit firewall rule to open port 5555.
On GCE this can be done with:
```
$ gcloud compute firewall-rules create --allow=tcp:5555 --target-tags=kubernetes-minion kubernetes-minion-5555
```
Please remember to delete the rule after you are done with the example (on GCE: `$ gcloud compute firewall-rules delete kubernetes-minion-5555`)
To bring up the pods, run this command `$ kubectl create -f examples/celery-rabbitmq/flower-controller.yaml`. This controller is defined as so:
```yaml
apiVersion: v1
@ -213,17 +239,22 @@ Again, it uses the Kubernetes-provided environment variable to obtain the addres
Once all pods are up and running, running `kubectl get pods` will display something like this:
```
POD IP CONTAINER(S) IMAGE(S) HOST LABELS STATUS
celery-controller-h3x9k 10.246.1.11 celery endocode/celery-app-add 10.245.1.3/10.245.1.3 app=taskQueue,name=celery Running
flower-controller-cegta 10.246.2.17 flower endocode/flower 10.245.1.4/10.245.1.4 app=taskQueue,name=flower Running
kube-dns-fplln 10.246.1.3 etcd quay.io/coreos/etcd:latest 10.245.1.3/10.245.1.3 k8s-app=kube-dns,kubernetes.io/cluster-service=true Running
kube2sky kubernetes/kube2sky:1.0
skydns kubernetes/skydns:2014-12-23-001
rabbitmq-controller-pjzb3 10.246.2.16 rabbitmq library/rabbitmq 10.245.1.4/10.245.1.4 app=taskQueue,name=rabbitmq Running
NAME READY REASON RESTARTS AGE
celery-controller-wqkz1 1/1 Running 0 8m
flower-controller-7bglc 1/1 Running 0 7m
rabbitmq-controller-5eb2l 1/1 Running 0 13m
```
Now you know on which host Flower is running (in this case, 10.245.1.4), you can open your browser and enter the address (e.g. `http://10.245.1.4:5555`. If you click on the tab called "Tasks", you should see an ever-growing list of tasks called "celery_conf.add" which the run\_tasks.py script is dispatching.
`kubectl get service flower-service` will help you to get the external IP addresses of the flower service.
```
NAME LABELS SELECTOR IP(S) PORT(S)
flower-service name=flower app=taskQueue,component=flower 10.0.44.166 5555/TCP
162.222.181.180
```
Point your internet browser to the appropriate flower-service address, port 5555 (in our case http://162.222.181.180:5555).
If you click on the tab called "Tasks", you should see an ever-growing list of tasks called "celery_conf.add" which the run\_tasks.py script is dispatching.

13
examples/celery-rabbitmq/flower-service.yaml Normal file
View File

@ -0,0 +1,13 @@
apiVersion: v1
kind: Service
metadata:
labels:
name: flower
name: flower-service
spec:
ports:
- port: 5555
selector:
app: taskQueue
component: flower
type: LoadBalancer

1
examples/examples_test.go
View File

@ -220,6 +220,7 @@ func TestExampleObjectSchemas(t *testing.T) {
"../examples/celery-rabbitmq": {
"celery-controller": &api.ReplicationController{},
"flower-controller": &api.ReplicationController{},
"flower-service": &api.Service{},
"rabbitmq-controller": &api.ReplicationController{},
"rabbitmq-service": &api.Service{},
},

21
test/e2e/examples.go
View File

@ -127,6 +127,7 @@ var _ = Describe("Examples e2e", func() {
rabbitmqControllerYaml := mkpath("rabbitmq-controller.yaml")
celeryControllerYaml := mkpath("celery-controller.yaml")
flowerControllerYaml := mkpath("flower-controller.yaml")
flowerServiceYaml := mkpath("flower-service.yaml")
nsFlag := fmt.Sprintf("--namespace=%v", ns)
By("starting rabbitmq")
@ -144,14 +145,32 @@ var _ = Describe("Examples e2e", func() {
})
By("starting flower")
runKubectl("create", "-f", flowerServiceYaml, nsFlag)
runKubectl("create", "-f", flowerControllerYaml, nsFlag)
forEachPod(c, ns, "component", "flower", func(pod api.Pod) {
//TODO: Do a http request after a flower service is added to the example.
// Do nothing. just wait for it to be up and running.
})
content, err := makeHttpRequestToService(c, ns, "flower-service", "/")
Expect(err).NotTo(HaveOccurred())
if !strings.Contains(content, "<title>Celery Flower</title>") {
Failf("Flower HTTP request failed")
}
})
})
})
func makeHttpRequestToService(c *client.Client, ns, service, path string) (string, error) {
result, err := c.Get().
Prefix("proxy").
Namespace(ns).
Resource("services").
Name(service).
Suffix(path).
Do().
Raw()
return string(result), err
}
func forEachPod(c *client.Client, ns, selectorKey, selectorValue string, fn func(api.Pod)) {
var pods *api.PodList
var err error