trion-development / docker-jenkins-docker-client Goto Github PK

Hello,

We based our CI environment on this project, deploying our jenkins in a GKE k8s cluster, and building our docker image artifacts from there in shell jenkins job that were using directly docker build and docker push.

The problem is that since a last upgrade [ https://kubernetes.io/blog/2020/12/02/dont-panic-kubernetes-and-docker/ ], the docker socket is not available on the node anymore.

Do you know if there is a workaround to keep our pipeline as close as possible from what was working a couple of weeks ago without a sudden need to re-design everything ?

I've seen people talking about kaniko and stuff, but we don't like the idea of having a build process that would not be reproducible in our local dev machine :/ Until now, it was exactly the exact same script that was used to build in a local dev env or in the CI env, and so it make it very smooth to modify and test them.

Hello, I'd like to ask if it would be possible to introduce release tags.

We are using your image in our Jenkins running in Kubernetes cluster. Recent docker upgrade to 19.03.9 brings API version upgrade as well 1.39 -> 1.40. We needed to upgrade all our worker nodes to latest docker version immediately.

It would be good to stick with some version and plan the upgrade more in advance :-). I would be happy to help with maintenance.

Use case:
Looking to use this image in my Kubernetes deployment component and run via kubectl rather than the docker cmd.

The Issue:
When passing the image in my Kubernetes cluster, I'm able to land on the /login page, however, because this is running on Kubernetes and not docker, a generated password is not present on the command prompt.

Potential solution:
Is there a way to pass a temporary default password via, the -e (environment)? With this approach login proceeds to the dashboard.

I'm trying your excellent distribution of Jenkins for training, and I have an issue with automated plugin install

At startup, I get

Configuration as Code enabled
WARN: install-plugins.sh has been removed, please switch to jenkins-plugin-cli
Configuration as Code: Installing default config
Starting /usr/local/bin/jenkins.sh as jenkins
Running from: /usr/share/jenkins/jenkins.war
webroot: EnvVars.masterEnvVars.get("JENKINS_HOME")

The second line is especially interesting.

Recent versions of Jenkins seems to provide a CLI able to install plugins, provided it knows which Jenkins instance to call (I guess).