Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Not A Defect
-
None
Description
When I try to override the jnlp container image I still always get the alpine image and do not get the specified image. It seems that event hough i specify a container with the name jnlp it just ignores it and uses the default alpine image.
agent {
kubernetes {
label "human-review-ui-pipeline-${env.BUILD_ID}"
defaultContainer 'jnlp'
yaml """
apiVersion: v1
kind: Pod
metadata:
{{ labels:}}
{{ pod-template: jenkins-slave-npm}}
spec:
{{ containers:}}
{{ - name: jnlp}}
image: "redhat-cop/jenkins-slave-image-mgmt"
"""
Attachments
Issue Links
- is duplicated by
-
-
- Resolved
-
Activity
I also have similar issue. Irrespective of yaml or ui override, the plugin is always pulling jnlp:alpine even though i have my custom jnlp named as jnlp
Jenkins: 2.150.3 and Kubernetes plugin: 1.14.3
Can we please reopen this ticket or explain what configuration will make my custom jnlp work?
yrsurya your pod is failing, you need to check why
Agent is not connected after 30 seconds
so it is probably using an old agent or another one you have with the same label
csanchez I am still having this issue. I don't have any pod templates defined in the `Kubernetes` configuration. I only define the pod / containers in the declarative pipelines. So there is no merging going on. But for some reason no matter what i put in for `image` for the `jnlp` container in the declarative pipeline, it gets ignored and the default alipine container gets used.
Same issue here with a declarative pipline.
Jenkins version: 2.190.1
Kubernetes plugin: 1.20.1
I've created a pod template under Cloud section on the ui. With the following options:
Pod Template::
Name: jenkins-builder
...
Container Template ::
Name: my-jnlp
Docker image: jenkins/jnlp-slave:latest
Working directory: /home/jenkins/agent
...
...
Workspace Volume: PVC
Claim name: jenkins-slave-claim
Then created this basic pipline:
pipeline {
agent {
kubernetes {
defaultContainer 'my-jnlp'
yaml """
apiVersion: v1
kind: Pod
metadata:
name: jenkins-builder
spec:
containers:
- name: busybox
image: busybox
command:
- cat
tty: true
"""
}
}
stages {
stage('start') {
steps{
container('busybox'){
sh "ls"
}
}
}
}
}
In the console I always get the default jnlp container:
apiVersion: "v1" kind: "Pod" metadata: annotations: buildUrl: "http://jenkins.default.svc.k8s.si.net:8080/job/test/20/" labels: jenkins: "slave" jenkins/test_20-s37fr: "true" name: "test-20-s37fr-xq14x-z8rq5" spec: containers: - command: - "cat" image: "busybox" name: "busybox" tty: true volumeMounts: - mountPath: "/home/jenkins/agent" name: "workspace-volume" readOnly: false - command: - "cat" image: "maven:3-alpine" name: "builder-new" tty: true volumeMounts: - mountPath: "/home/jenkins/agent" name: "workspace-volume" readOnly: false - env: - name: "JENKINS_SECRET" value: "********" - name: "JENKINS_TUNNEL" value: "jenkins-agent.default.svc.k8s.si.net:50000" - name: "JENKINS_AGENT_NAME" value: "test-20-s37fr-xq14x-z8rq5" - name: "JENKINS_NAME" value: "test-20-s37fr-xq14x-z8rq5" - name: "JENKINS_AGENT_WORKDIR" value: "/home/jenkins/agent" - name: "JENKINS_URL" value: "http://jenkins.default.svc.k8s.si.net:8080/" image: "jenkins/jnlp-slave:alpine" name: "jnlp" volumeMounts: - mountPath: "/home/jenkins/agent" name: "workspace-volume" readOnly: false nodeSelector: {} restartPolicy: "Never" volumes: - emptyDir: medium: "" name: "workspace-volume"
So it's not what I'd like to see. I cannot find out how I can use my Pod with the PVC.
I cannot workaround this with JENKINS-56375 unfortunately.
Robert Horvath
added a comment - - edited Same issue here with a declarative pipline.
Jenkins version: 2.190.1
Kubernetes plugin: 1.20.1
I've created a pod template under Cloud section on the ui. With the following options:
Pod Template::
Name: jenkins-builder
...
Container Template ::
Name: my-jnlp
Docker image: jenkins/jnlp-slave:latest
Working directory: /home/jenkins/agent
...
...
Workspace Volume: PVC
Claim name: jenkins-slave-claim
Then created this basic pipline:
pipeline {
agent {
kubernetes {
defaultContainer 'my-jnlp'
yaml """
apiVersion: v1
kind: Pod
metadata:
name: jenkins-builder
spec:
containers:
- name: busybox
image: busybox
command:
- cat
tty: true
"""
}
}
stages {
stage( 'start' ) {
steps{
container( 'busybox' ){
sh "ls"
}
}
}
}
}
In the console I always get the default jnlp container:
apiVersion: "v1"
kind: "Pod"
metadata:
annotations:
buildUrl: "http: //jenkins. default .svc.k8s.si.net:8080/job/test/20/"
labels:
jenkins: "slave"
jenkins/test_20-s37fr: " true "
name: "test-20-s37fr-xq14x-z8rq5"
spec:
containers:
- command:
- "cat"
image: "busybox"
name: "busybox"
tty: true
volumeMounts:
- mountPath: "/home/jenkins/agent"
name: "workspace-volume"
readOnly: false
- command:
- "cat"
image: "maven:3-alpine"
name: "builder- new "
tty: true
volumeMounts:
- mountPath: "/home/jenkins/agent"
name: "workspace-volume"
readOnly: false
- env:
- name: "JENKINS_SECRET"
value: "********"
- name: "JENKINS_TUNNEL"
value: "jenkins-agent. default .svc.k8s.si.net:50000"
- name: "JENKINS_AGENT_NAME"
value: "test-20-s37fr-xq14x-z8rq5"
- name: "JENKINS_NAME"
value: "test-20-s37fr-xq14x-z8rq5"
- name: "JENKINS_AGENT_WORKDIR"
value: "/home/jenkins/agent"
- name: "JENKINS_URL"
value: "http: //jenkins. default .svc.k8s.si.net:8080/"
image: "jenkins/jnlp-slave:alpine"
name: "jnlp"
volumeMounts:
- mountPath: "/home/jenkins/agent"
name: "workspace-volume"
readOnly: false
nodeSelector: {}
restartPolicy: "Never"
volumes:
- emptyDir:
medium: ""
name: "workspace-volume"
So it's not what I'd like to see. I cannot find out how I can use my Pod with the PVC.
I cannot workaround this with JENKINS-56375 unfortunately.
I tried with custom image didn't work so given jenkins-jnlp official image even didn't work
using this on scripted pipeline
node('eks-cluster') {
stage('Check jnlp') {
sh 'rm -rf *'