-
Type:
Bug
-
Resolution: Not A Defect
-
Priority:
Minor
-
Component/s: kubernetes-plugin
-
Environment:Jenkins 2.35
Kubernetes-Plugin 0.10
I've defined multiple pod-templates in the Kubernetes-Plugin with only 1 container (the pod template definition is attached).
When I run a new job from some pod-template (I tried with 2 different), the pod contains a non-defined container with an independent docker image and resources called jnlp.
Here is the result when I describe the pod:
Name: base-49928037d9757
Namespace: jenkins
Security Policy: hostmount-anyuid
Node: nodename
Start Time: Thu, 22 Dec 2016 09:40:05 +0100
Labels: jenkins=slave
jenkins/base=true
Status: Running
IP: 10.1.1.2
Controllers: <none>
Containers:
slave:
Container ID: docker://ab6c86f5c9900f4e1b0c19fe1dda21536d591048bcb4f124fe357d1cd2ccbd44
Image: myImage
Image ID: docker://sha256:66c32c2cc23742e0f860b4068127ab0365ce24fa21a9932b17c873bee0210725
Port:
Command:
/var/lib/jenkins/bin/jenkins-slave.sh
Args:
c161cc0981fa74e8ef64255824b21b61c23739824ba99a6aa843eafab85dadbd
base-49928037d9757
Limits:
cpu: 1
memory: 1Gi
Requests:
cpu: 1
memory: 1Gi
State: Running
Started: Thu, 22 Dec 2016 09:40:09 +0100
Ready: True
Restart Count: 0
Volume Mounts:
/opt/jenkins-slave from volume-0 (rw)
/var/run/secrets/kubernetes.io/serviceaccount from default-token-0xn6i (ro)
Environment Variables:
JENKINS_SECRET: c161cc0981fa74e8ef64255824b21b61c23739824ba99a6aa843eafab85dadbd
JENKINS_NAME: base-49928037d9757
JENKINS_LOCATION_URL: https://bcngpp1vsjenkins01.scytl.net/
JENKINS_URL: https://jenkins1.scytl.net
JENKINS_JNLP_URL: https://jenkins1.scytl.net/computer/base-49928037d9757/slave-agent.jnlp
HOME: /opt/jenkins-slave
jnlp:
Container ID: docker://52bd0d42aebf344c813d13f4436d0f78750a9c8f58bff62eefb76e4d6dfb951f
Image: jenkinsci/jnlp-slave:alpine
Image ID: docker://sha256:254fd665eaf0229f38295a9eac6c7f9bf32a2f450ecbcc8212f3e53b96dd339d
Port:
Args:
c161cc0981fa74e8ef64255824b21b61c23739824ba99a6aa843eafab85dadbd
base-49928037d9757
Limits:
cpu: 1
memory: 1Gi
Requests:
cpu: 1
memory: 1Gi
State: Running
Started: Thu, 22 Dec 2016 09:40:11 +0100
Ready: True
Restart Count: 0
Volume Mounts:
/home/jenkins from workspace-volume (rw)
/opt/jenkins-slave from volume-0 (rw)
/var/run/secrets/kubernetes.io/serviceaccount from default-token-0xn6i (ro)
Environment Variables:
JENKINS_SECRET: c161cc0981fa74e8ef64255824b21b61c23739824ba99a6aa843eafab85dadbd
JENKINS_NAME: base-49928037d9757
JENKINS_LOCATION_URL: jenkinsUrl
JENKINS_URL: jenkinsUrl
JENKINS_JNLP_URL: jenkinsUrl/computer/base-49928037d9757/slave-agent.jnlp
HOME: /home/jenkins
Conditions:
Type Status
Initialized True
Ready True
PodScheduled True
Volumes:
volume-0:
Type: NFS (an NFS mount that lasts the lifetime of a pod)
Server: ipNfs
Path: /opt/jenkins-shared
ReadOnly: false
workspace-volume:
Type: EmptyDir (a temporary directory that shares a pod's lifetime)
Medium:
default-token-0xn6i:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-0xn6i
QoS Tier: Guaranteed
The entrypoint of the slave container (which I've defined) is using a bash based on csanchez bash script, and is connecting using Jenkins Remoting layer.
How can you explain this new container called jnlp? It is consuming so much resources per each pod and decreasing my environment.
I tried a downgrade of the Kubernetes-Plugin to 0.9 and it isn't happening.
