问题

I was trying to use io.fabric8 api to create a few resources in kubernetes using a pod-spec.yaml.

Config config = new ConfigBuilder()

.withNamespace("ag")

.withMasterUrl(K8_URL)

.build();

try (final KubernetesClient client = new DefaultKubernetesClient(config)) {

LOGGER.info("Master: " + client.getMasterUrl());

LOGGER.info("Loading File : " + args[0]);

Pod pod = client.pods().load(new FileInputStream(args[0])).get();

LOGGER.info("Pod created with name : " + pod.toString());

} catch (Exception e) {

LOGGER.error(e.getMessage(), e);

}

The above code works if the resource type is of POD. Similarly for other resource type it is working fine.

But if the yaml has multiple resource type like POD and service in the same file, how to use fabric8 Api ?

I was trying to use client.load(new FileInputStream(args[0])).createOrReplace(); but it is crashing with the below exception:

java.lang.NullPointerException

at java.net.URI$Parser.parse(URI.java:3042)

at java.net.URI.(URI.java:588)

at io.fabric8.kubernetes.client.utils.URLUtils.join(URLUtils.java:48)

at io.fabric8.kubernetes.client.dsl.base.BaseOperation.getMandatory(BaseOperation.java:208)

at io.fabric8.kubernetes.client.dsl.base.BaseOperation.get(BaseOperation.java:177)

at io.fabric8.kubernetes.client.handlers.PodHandler.reload(PodHandler.java:53)

at io.fabric8.kubernetes.client.handlers.PodHandler.reload(PodHandler.java:32)

at io.fabric8.kubernetes.client.dsl.internal.NamespaceVisitFromServerGetWatchDeleteRecreateWaitApplicableListImpl.createOrReplace(NamespaceVisitFromServerGetWatchDeleteRecreateWaitApplicableListImpl.java:202)

at io.fabric8.kubernetes.client.dsl.internal.NamespaceVisitFromServerGetWatchDeleteRecreateWaitApplicableListImpl.createOrReplace(NamespaceVisitFromServerGetWatchDeleteRecreateWaitApplicableListImpl.java:62)

at com.nokia.k8s.InterpreterLanuch.main(InterpreterLanuch.java:66)

Yaml file used

apiVersion: v1

kind: Pod

metadata:

generateName: zep-ag-pod

annotations:

kubernetes.io/psp: restricted

spark-app-name: Zeppelin-spark-shared-process

namespace: ag

labels:

app: zeppelin

int-app-selector: shell-123

spec:

containers:

- name: ag-csf-zep

image: bcmt-registry:5000/zep-spark2.2:9

imagePullPolicy: IfNotPresent

command: ["/bin/bash"]

args: ["-c","echo Hi && sleep 60 && echo Done"]

securityContext:

allowPrivilegeEscalation: false

capabilities:

drop:

- ALL

runAsNonRoot: true

securityContext:

fsGroup: 2000

runAsUser: 1510

serviceAccount: csfzeppelin

serviceAccountName: csfzeppelin

---

apiVersion: v1

kind: Service

metadata:

name: zeppelin-service

namespace: ag

labels:

app: zeppelin

spec:

type: NodePort

ports:

- name: zeppelin-service

port: 30099

protocol: TCP

targetPort: 8080

selector:

app: zeppelin

回答1:

You don't need to specify resource type whenever loading a file with multiple documents. You simply need to do:

// Load Yaml into Kubernetes resources

List result = client.load(new FileInputStream(args[0])).get();

// Apply Kubernetes Resources

client.resourceList(result).inNamespace(namespace).createOrReplace()

来源：https://stackoverflow.com/questions/53501540/kubectl-apply-f-spec-yaml-equivalent-in-fabric8-java-api