I think the solution here is to add a yamlFile or whatever parameter for the Kubernetes agent type, and modify KubernetesDeclarativeAgentScript to check if that parameter is given, and then assuming it is and the scm object in the Pipeline context is non-null (i.e., this is either a multibranch Pipeline or a standalone Pipeline with "Pipeline from SCM"), it would use readTrusted to load the yaml and pass it on from there. Seems pretty simple.

Got a preliminary PR up at https://github.com/jenkinsci/kubernetes-plugin/pull/355

Merged in 1.10.0

Using yamlFile does not work with parameterized git branch builds.

I'm getting the following exception when the plugin tries to load the file:

hudson.plugins.git.GitException: Command "git fetch --tags --progress origin +refs/heads/${BRANCH}:refs/remotes/origin/${BRANCH} --prune" returned status code 128:
stdout: 
stderr: fatal: Couldn't find remote ref refs/heads/${BRANCH}

csanchez. See my previous comment regarding parameterized builds and declarative pipeline.

isn't that JENKINS-52623 ?
please do not reopen, and file a new issue instead if that's not the case

podTemplate(label: slave_label, cloud: "${cloud_name}", readFile : 'deploy.yml' ) {
 
}
 
i used deploy.yml and added all pod , namespace , name , image and resources etc.
but pod is not creating . its not even reading the yaml file. instead its going to docker hub internet. but my image is in my own artifactory . i checked kubernetes dashboard pod is not creating due jnlp was not created.
error is like pod bddrunner(my pod mentioned in yaml file) is offline
please help me to fix it.
 
yaml file:
apiVersion: v1
kind: Pod
metadata:
name: bdd-runner
namespace: edison-core
spec:
containers:

• name: 'jnlp'
  image: 'myartifactory/jenkins/jnlp-slave:3.10-1-alpine'
  command:
• cat
  tty: true
• name: 'custom-slave'
  image: 'my artifactorydocker-snapshot-abs/eis-test-data/absdata:latest'
  command:
• cat
  tty: true

Bulk closing resolved issues.