Services map a port on each cluster node to ports on one or more pods.
The mapping uses a selector
key:value pair in the service, and the
labels
property of pods. Any pods whose labels match the service selector
are made accessible through the service’s port.
For more information, see the Services Overview.
Services are created by passing a configuration file to the kubectl create
command:
$ kubectl create -f FILE
Where:
-f FILE
or --filename FILE
is a relative path to a
service configuration file in either JSON
or YAML format.A successful service create request returns the service name. You can use a sample file below to try a create request.
When creating a service, you must point to a service configuration file as the
value of the -f
flag. The configuration file can be formatted as
YAML or as JSON, and supports the following fields:
{
"kind": "Service",
"apiVersion": "v1",
"metadata": {
"name": string
},
"spec": {
"ports": [{
"port": int,
"targetPort": int
}],
"selector": {
string: string
},
"type": "LoadBalancer"
}
}
Required fields are:
kind
: Always Service
.apiVersion
: Currently v1
.metadata
: Contains:
name
: The name to give to this service.spec
: Contains:
ports
: The ports to map. port
is the service port to expose on the
cluster IP. targetPort
is the port to target on the pods that are part
of this service.selector
: The label key:value pair that defines the pods to
target.type
: Optional. If the type is LoadBalancer
, sets up a network load balancer
for your service. This provides an externally-accessible IP address that
sends traffic to the correct port on your cluster nodes.For the full service
schema see the
Kubernetes api reference.
The following service configuration files assume that you have a set of pods
that expose port 9376 and carry the label app=example
.
Both files create a new service named myapp
which resolves to TCP port 9376
on any pod with the app=example
label.
The difference in the files is in how the service is accessed. The first file does not create an external load balancer; the service can be accessed through port 8765 on any of the nodes’ IP addresses.
service-sample.json |
---|
|
service-sample.yaml |
---|
|
The second file uses
network load balancing to create a
single IP address that spreads traffic to all of the nodes in
your cluster. This option is specified with the
"type": "LoadBalancer"
property.
load-balancer-sample.json |
---|
|
load-balancer-sample.yaml |
---|
|
To access the service, a client connects to the external IP address, which forwards to port 8765 on a node in the cluster, which in turn accesses port 9376 on the pod. See the Service configuration file section of this doc for directions on finding the external IP address.
To list all services on a cluster, use the
kubectl get
command:
$ kubectl get services
A successful get request returns all services that exist on the specified cluster:
NAME LABELS SELECTOR IP PORT
myapp <none> app=MyApp 10.123.255.83 8765/TCP
To return information about a specific service, use the
kubectl describe
command:
$ kubectl describe service NAME
Details about the specific service are returned:
Name: myapp
Labels: <none>
Selector: app=MyApp
IP: 10.123.255.83
Port: <unnamed> 8765/TCP
NodePort: <unnamed> 31474/TCP
Endpoints: <none>
Session Affinity: None
No events.
To return information about a service when event information is not required,
substitute get
for describe
.
To delete a service, use the kubectl delete
command:
$ kubectl delete service NAME
A successful delete request returns the deleted service’s name.