From @therynamo:

"Another nicety would be being able to do something like kubectl apply -f, programatically. So instead of seeing, Error: deployment "insert_deployment_name_here" already exists, it would just overwrite with the new config. However, I may have yet again misread the docs."

The current code copy pastas resource names in several places (e.g., common.js and namespace.js). Keeping them in a single place (e.g., defining an enum in common.js) will make them less painful to update.

This may be an inquiry more than a request, but I'm trying to use this library to list/create a job in my cluster.

According to these docs, the correct api endpoint is:
GET /apis/batch/v1/namespaces/{namespace}/jobs/{name}.

When I use this library like so (assuming the namespace configured in the k8 constructor is my-namespace:
k8.namespace.jobs('my-job').get(a => console.log(a));

It makes a request to:

/api/v1/namespaces/my-namespace/jobs/my-job

How do I get the first part of that url be /apis/batch/v1/namespaces? Is there a setting I'm missing somewhere?

Thanks, and I appreciate all the work you've done on the library so far!

Is there a clean way to watch jobs?

The K8s URL is /apis/batch/v1/watch/namespaces/{namespace}/jobs/{name} (https://kubernetes.io/docs/api-reference/v1.6/#watch-47)

To get it to actually use that URL, I fiddled the version:

const batchWatch = new K8s.Batch({
  version: 'v1/watch',
  ...
});

const stream = batchWatch.ns.jobs('foo').get();

Is there supposed to be a better way of doing this?

Are there plans to support queries to cluster level resources queries

eg.
/api/v1/resourcequotas
/api/v1/replicationcontrollers

Is there a way to get resources over all namespaces? I couldn't see an obvious way, unless I'm missing something.

The Namespaces object supports a default resources argument, but no APIs leverage it. Change our existing APIs to pass the set of resources they support (e.g., Core). This will require adding some plumbing to ApiGroup to pass the array of resources to the Namespaces constructor.

kubernetes-client doesn't support ThirdPartyResource.

We should add support to the ApiGroup and add an example usage to README.md.

I'm getting following error

{
  "code": 404
}

this is my server.js file ( i have exported export NODE_TLS_REJECT_UNAUTHORIZED=0 )

const K8Api = require('kubernetes-client');
const k8 = new K8Api({
  url: 'https://10.2.2.2:8443',
  //version: 'v1beta1', 
  auth: {
    bearer: 'sulLF-Y-vn-wgfNcjXR5bb0JU7-E0G8LDKtKYyDGpQ0'
  }
});

function print(err, result) {
  console.log(JSON.stringify(err || result, null, 2));
  console.log(result);
}

k8.namespaces.replicationcontrollers.get('mongodb', print);
k8.namespaces.replicationcontrollers.get('mongodb-1', print);

But when i do

jjonagam-osx:k8s-api cjonagam$ kubectl get rc
NAME DESIRED CURRENT AGE
mongodb-1 3 3 4d

clusters, petsets, quota (not resourcequotas/'quota') and replicasets.

I am trying to list pods form specific deployment. On the docs it says it is supported on the rc(NOT deployment) but that does not work for me. I can list pods via kube.core.namespace.pods.get() only.

There should be a warning when users use a deprecated type. However there needs to be a 'type x version' matrix before this can be implemented.

From @jcrugzz on implementing this:
"Not a bad idea. The only tricky part is that deprecation depends on what k8s version the user is using. We should be showing a support matrix for types and k8s version after this merges"

Initial suggestion in #78

i would like to extend this k8s library to openshift which has additional objects like Route DeploymentConfig etc can you please guide me the steps that would be involved ?

Hi, for some reason I can create a deployment but I cannot load/list/update it since I will always get Error: the server could not find the requested resource. Other resources are just fine.

Any idea what might be wrong?

Sorry for opening a bug for a question - but was wondering how the watch support is implemented. Is is just polling or establishing a web socket connection?

Do the examples mean k9? Dropping the trailing s from k8s makes the abbreviation non-sensical.

Does this client have support for OIDC, similar to the kubernetes dashboard code (https://github.com/kubernetes/dashboard/blob/master/vendor/github.com/coreos/go-oidc/oidc/client.go)

I am using this javascript client and I am not sure if the instructions to authenticate using OIDC are clear in doc.

thanks, Rajesh

After looking at the k8s docs on deployment delete, it seems like passing orphanDependents: false should delete "everything" associated with the deployment.

Currently, I am trying to use the kubernetes-client api to accomplish full deletes of items associated with a deployment. For example, when using the following api:

const deployment = {
  {
  "apiVersion": "extensions/v1beta1",
  "kind": "Deployment",
  "metadata": {
    "name": "<insert_deployment_name_here>"
  },
  "spec": {
    "replicas": 2,
  ... // etc.
}

// Assume k8s is an instance of `kubernetes-api` v3.1
k8s.group(deployment).ns.deployments.delete({ body: deployment, orphanDependents: false }
// or instead using: 
k8s.group(deployment).ns.deployments.delete({ body: deployment, preservePods: false }

I would expect the above to delete an already existing deployment with the given deployment name, and all pods, and replicasets. However, the observed behavior is that the deployment does get deleted, though the pods and replicasets are orphaned (i.e. not deleted). If you run a kubectl delete deployment <deployment_name>, by default, all resources associated with the deployment will be deleted.

Perhaps you could let me know if I'm using the api wrong, or if kubernetes-client does not support this.

What I'm trying to avoid is having to have 3 callbacks, each deleting one of the above resources, in order to create a deployment with the same name.

Another nicety would be being able to do something like kubectl apply -f, programatically. So instead of seeing, Error: deployment "insert_deployment_name_here" already exists, it would just overwrite with the new config. However, I may have yet again misread the docs.

Any help is appreciated! Thanks for maintaining this project, its helped a lot and has allowed me to not have to write my own library to do exactly this. 👍 x 💯

When I do a k8.namespaces.replicationcontrollers.pod.get(print); Its a very long o/p..I am interested in knowing if the pod is in "Running" status or not..How can I fetch this information?
Please help!

The best practice for watching events is to use the new API endpoints (e.g., https://kubernetes.io/docs/api-reference/v1.7/#watch-64). See #123 for more motivation.

Guidance:

• Create a new Watch object that extends ContainerBaseObject
• Add a Watch object to each ApiGroup. See the how ApiGroup adds .namespace in constructor
• Add unit tests and integration tests

https://github.com/godaddy/kubernetes-client/blob/master/lib/pods.js

The deprecated warning has been around for 6 months. Remove it before publishing 4.0.0.

I am struggling with the difference between post and patch on different modules, so far...

I can post using the API (but not patch)...

const api = new k8s.Api({config})
api.group(manifest0).ns.kind(manifest0).patch({ body: manifest0 }

Error: 'the server does not allow this method on the requested resource'

I can patch (deployments only) but not post. Services as an example returns the same error on both post and patch...

const ext = new k8s.Extensions({config})
ext.namespaces.deployments(project_name) .patch({ body: body}, () => {response})

Error: 'the server does not allow this method on the requested resource'

E.g., POST /apis/extensions/v1beta1/namespaces/{namespace}/deployments/{name}/rollback

This is a special case of a more general problem we need to handle with other objects (e.g., proxy for Pods and Services).

It would be nice if this library also supported promises. Unfortunately, wrapping the request library for promises is fairly non-trivial. This makes it a bit unwieldy to promisify at runtime.

The request-promise library also supports a nodeify method for callbacks, allowing simulataneous support for callbacks and promises. However, this may make streaming more cumbersome.

Perhaps a provide your own request library approach?

Changelog / migration guide would be helpful.

I would recommend automated generation with https://github.com/conventional-changelog or https://github.com/semantic-release/semantic-release

The Python client does something similar by leverage code generated by swagger-codegen.

kubernetes-client could use swagger-js to generate code dynamically using swagger specs distributed with kubernetes-client or using swagger specs hosted on the kube-apiserver.

I'm struggling to get POST'ing YAML to the deployments API working.

extensions.ns('test').deployments.post({
    headers: { 'content-type': 'application/yaml' },
    body: `
    apiVersion: extensions/v1beta1
    kind: Deployment ... `
});

I see if I alter the library to log out what options are generated:

{ headers: { 'content-type': 'application/json' },
  path: [ '/apis/extensions/v1beta1/namespaces/test/deployments' ],
  body: 'apiVersion: extensions/v1beta1\nkind: Deployment\n ...' }

How can I get this library to send the application/yaml header? I see the API documention for k8s says it' accepts YAML as their CURL example is YAML based.

https://kubernetes.io/docs/api-reference/v1.7/#-strong-write-operations-strong--20

And click curl for the example

It is not clear from the README which version of the kubernetes API this supports. e.g 1.5, 1.6, 1.7

How does this module handle Authorization? Is it expected that the end-user will pass them in as headers?

I am having trouble running the example job from the k8s documentation using the client.

const Api = require('kubernetes-client');
var config = Api.config.fromKubeconfig();
config.apiVersion = 'batch/v1';
const ext = new Api.Extensions(config);

ext.namespaces.jobs.post({
    body: {
      apiVersion: 'batch/v1',
      kind: 'Job',
      metadata: {
        name: 'pi'
      },
      spec: {
        template: {
          spec: {
            restartPolicy: 'Never',
            containers: [{
              image: 'perl',
              command: ['perl', '-Mbignum=bpi', '-wle', 'print bpi(2000)'],
              name: 'pi'
            }]
          },
          metadata: {
            name: 'pi'
          }
        }
      }
    }
  },
  function(err, job) {
    if (err) return console.log('Failed to creata job: ' + err);
    console.log(job);
  });

When I run the script I get the following error:

Failed to creata job: Error: the server could not find the requested resource

But I can schedule the job without problem using kubectl

$ curl -O https://raw.githubusercontent.com/kubernetes/kubernetes.github.io/master/docs/concepts/workloads/controllers/job.yaml
$ kubectl create -f job.yaml
job "pi" created

Am I missing something or I a not configuring the extension correctly?

I used to be able to call the following:

k8.ns.pvc.matchLabels({ app: 'test' }).get((err, sRes) => {

but that now gives a 'resource not found' error. I now have to call the following instead:

k8.ns.pvc.get({ qs: { labelSelector: 'app=test' } }, (err, sRes) => {

I use watch in the following format:

kubernetes.api.ns.po.get({
      qs: {
        watch: true,
        labelSelector: PodManager.serviceNameLabel
      }
    }
})

After a while it stops listening to events.
Do you have an idea?

Thanks!

How to update and delete ConfigMap?

const api = new Api.Api({ url: process.env.K8S_HOST, namespace: kubeParam.namespace, ca: fs.readFileSync(KUBERNETES_CA_PATH,{'encoding':'utf-8'}), insecureSkipTlsVerify: true, auth: { bearer: fs.readFileSync(KUBERNETES_TOKEN_PATH,{'encoding':'utf-8'}), } }); api.core.configmap.delete({name: 'configmap-test'},print)

result
{ Error: the server could not find the requested resource }

Getting this error while connecting. how can i fix this?

{
  "code": "SELF_SIGNED_CERT_IN_CHAIN"
}

The requirements and expectations for potential contributors should be clear.

• We use https://github.com/conventional-changelog/standard-version
• Angular-style commit messages (for standard-version) https://github.com/angular/angular.js/blob/master/CONTRIBUTING.md

I wanted to programmatically get inside the container and check if a certain process is up or not..Is that possible?

When deleting a replication controller, the replica count of the pods is set to zero before the deletion. However, when deleting jobs and deployments, they job/deployment is deleted while leaving existing pods running.

Hi,

would you be open for adding a built-in in-cluster config?

Based on:
https://github.com/kubernetes/client-go/blob/124670e99da15091e13916f0ad4b2b2df2a39cd5/rest/config.go#L274
and
http://kubernetes.io/docs/user-guide/accessing-the-cluster/#accessing-the-api-from-a-pod

Something like:

const fs = require('fs')

const certPath = '/var/run/secrets/kubernetes.io/serviceaccount/ca.crt'
const tokenPath = '/var/run/secrets/kubernetes.io/serviceaccount/token'
const namespacePath = '/var/run/secrets/kubernetes.io/serviceaccount/namespace'

function getInClusterConfig () {
  const host = process.env.KUBERNETES_SERVICE_HOST
  const port = process.env.KUBERNETES_SERVICE_PORT
  let cert
  let namespace
  let bearer

  if (!host || !port) {
    throw TypeError('Unable to load in-cluster configuration, KUBERNETES_SERVICE_HOST and KUBERNETES_SERVICE_PORT must be defined')
  }

  try {
    cert = fs.readFileSync(certPath, 'utf8')
  } catch (err) {
    throw TypeError('Expected to load root CA config from ${certPath}, but got err', err)
  }

  try {
    bearer = fs.readFileSync(tokenPath, 'utf8')
  } catch (err) {
    throw TypeError('Expected to load token config from ${tokenPath}, but got err', err)
  }

  try {
    namespace = fs.readFileSync(namespacePath, 'utf8')
  } catch (err) {
    throw TypeError('Expected to load namespace config from ${namespacePath}, but got err', err)
  }

  return {
    url: `https://${host}:${port}`,
    cert,
    auth: { bearer },
    namespace
  }
}

The interface could be:

const K8Api = require('kubernetes-client');
const k8 = new K8Api.Core(K8Api.getInClusterConfig())

I'm happy to send a PR if you like the idea.
What do you think?

Is there a support to patch pod, I am looking to remove certain labels from a given pod.

I am trying to pass something like this in patch :
body: {op:"remove", path: "/metadata/labels/op"}

but it doesn't seem to be working.

Any advice??

Setting up RBAC is more involved (e.g., it's not just minikube --extra-config=apiserver.GenericServerRunOptions.AuthorizationMode=RBAC, you need a client cert with /O=system:masters and some setup to avoid breaking the other integration tests.

• Investigate the other setup required
• Make it easy to run that setup
• Ensure the other integration tests still work (another possibility is running two kube clusters, one for RBAC and one for every other integration test)

Is this supported? There is no mention in the "docs".

Hello! I'm trying to watch the events endpoint but the request it's not sending the watch=true. query string,

Here's how I'm doing it:

const JSONStream = require('json-stream');
jsonStream = new JSONStream();

const stream = k8.ns('myns').events.get({ qs: { watch: true } });
jsonStream.on('data', res => {
  // handling response
});

Hi,
Does the client support namespace creation?
Thanks,
Rafael

A set of typings for us Typescript users would be really cool.

How to update ConfigMap?

api.group(json).ns.kind(json).put({ body:JSON.parse(JSON.stringify(json)) }, print);

the result like so:
{ Error: the server does not allow this method on the requested resource}

When using the streaming example as documented in the readme I just get an error indicating that getStream is not a function.

I'm trying to get a services from kubernetes by passing kubeconfig to Api.config.fromKubeconfig() but I receive the following error:

Error
{ Error: read ECONNRESET
at exports._errnoException (util.js:1007:11)
at TLSWrap.onread (net.js:563:26) code: 'ECONNRESET', errno: 'ECONNRESET', syscall: 'read' }

Sample Code multible services

const Api = require('kubernetes-client');
const kubeconfig = Api.config.loadKubeconfig(
  expandHomeDir('~/.kube/kubeconfig')
);
const core = new Api.Core(Api.config.fromKubeconfig(kubeconfig));

function getKubeCoreService(){
  core.services.get(getServices);
}

function getServices(err, result){
  console.log(err);
}

How ever if I only request one service it executes without errors

Sample Code single service

const Api = require('kubernetes-client');
const kubeconfig = Api.config.loadKubeconfig(
  expandHomeDir('~/.kube/kubeconfig')
);
const core = new Api.Core(Api.config.fromKubeconfig(kubeconfig));

function getKubeCoreService(){
  core.services('servicename').get(getServices);
}

function getServices(err, result){
  console.log(err);
}

I have an existing deployment and I tried to add new label and update the deployment via:

GKE.prototype.update = function(json, successCallback, failureCallback) {
  var success = successCallback || console.log;
  var error = failureCallback || console.error;

  var type = this.getType(json);
  var exec = type === 'deployments' ? this.kube.extensions : this.kube.core;

  exec.namespaces[type](this.getName(json)).patch({ body: json }, function(err, result) {
    if (err) {
      error(err);
    } else {
      success(result);
    }
  });
};

after successful submission I have noticed that the deployment is the same, but a new replication controller was created with the new label.

This is unexpected behavior. How can I update the deployment properly?