Users will visit https://che-eclipse-che.192.168.0.1.nip.io/#https://github.com/test/eclipse-che It has the devfile to create the workspace.
First user registration will happen via keycloak and then the workspace will be created. This means a new kubernetes namespace will also be created for the user.
In the devfile I have the environment variable:
env:
- value: /C/Users/Administrator/.m2:/home/user/.m2
name: CHE_WORKSPACE_VOLUME
But the volume is not available in the workspace. What am I doing wrong?
The way how you can use persistent storage in Che is described in https://www.eclipse.org/che/docs/che-7/end-user-guide/configuring-a-workspace-using-a-devfile/#adding-components-to-a-devfile_che
But I'm afraid it won't satisfy your need.
Are you asking for every user to use your local maven repository? If so, it's not possible.
You can only enable maven repository volume for each workspace, so after workspace restart files will be used from the previous time, but each workspace uses its own private space.
Related
I have a question to ask but ill explain my plan/requirement first
I have started on new company
I have been tasked to migrate a lot of microservices running on swarm to Kubernetes
there are about 50 microservices running now
right now we are using consul as key/value store for configuration files
due to a lot of mistakes in designing infrastructure, our swarm is not stable ( failing overlays and so on)
developers want to have sub-versioning on configuration as well but in a specific way
one project for all config files
they don't want to go through building stages
there are some applications that read live configurations (
changes occur regularly )
so I need to centralize the configuration and create a project for this task
I store Kubernetes manifests GitLab-ci files and app configurations there
when I include ci files in the target project I can't access config and Kube manifests ( submodule is not acceptable by developers )
I'm planning to use helm instead of kubectl for deployment
my biggest challenge is to provide the configuration live ( as the developer pushes it applies on cm )
am I on the right track?
any suggestion on how to achieve my goal?
I expect to be able to deploy projects and use multiple files and folders from other projects
create a ci file like this in your devops repo, this job should commit the config file to your repo when config changed.
commit-config-file-to-devops-repo:
script: "command to commit config file to your devops repo"
only:
refs:
- master
changes:
- path/some-config-file.json
- configs/*
change default ci file location to point to ci file in your devops repo
https://192.168.64.188/help/ci/pipelines/settings#custom-cicd-configuration-path
my/path/.my-custom-file.yml#mygroup/another-project
setup pipeline, apply config to k8s when file commited.
Personally I use argocd to sync helm chart to k8s, you can do it your way.
Read live configurations is normally not recommended, because changing config may cause error.
When using k8s, it is better to create configmap and inject config into environment variables
Then use rollout mechanism to restart the app.
Howeven, if you are using configmap volume
It will auto update config file when you change config
https://kubernetes.io/docs/tasks/configure-pod-container/configure-pod-configmap/#mounted-configmaps-are-updated-automatically
I'm trying to move from certificate based GitLab Kubernetes integration which got deprecated, to new agent based Kubernetes integration.
I use CI/CD workflow, created separate project for Gitlab Kubernetes Agents and registered them there.
The equation is how to restrict the usage of registered agents in other projects?
Previously when one uses certificate based approach, one can set target namespace in project setting, also one can set environment for the integrated cluster, to use it with protected environments.
Now Kubernetes context is just available in other projects under same group, and once you have access to CI\CD files you can do whatever you want, and deploy anywhere.
I have a microservice application in one repo that communicates with another service that's managed by another repo.
This is not an issue when deploying to cloud, however, when devving locally the other service needs to be deployed too.
I've read this documentation: https://skaffold.dev/docs/design/config/#remote-config-dependency and this seems like a clean solution, but I only want it to depend on the git skaffold config if deploying locally (i.e. current context is "minikube").
Is there a way to do this?
Profiles can be automatically activated based on criteria such as environment variables, kube-context names, and the Skaffold command being run.
Profiles are processed after resolving the config dependencies though. But you could have your remote config include a profile that is contingent on a kubeContext: minikube.
Another alternative is to have several skaffold.yamls: one for prod, one for dev.
I have created a K8 deployment for KIE WB and KIE Server.
For KIE WB I have created a Docker Image that configures the Post Commit hook so my repos are pushed to GitHub.
All of this works great.
My question revolves around restoring a repository upon pod creation. So when my pod starts up I would like to clone my GitHub repos then they're restored without having to manually import them.
I thought I could use the GitHub API and get a tarball and drop them into the .niogit/ path but that's not working (I see the file structure in git is different, it looks like just the java source and not the required files for KIE WB recognize that it's a repo).
I know I have to be doing something that has been done but I am not finding anything to get this going. I don't want to reinvent the wheel either :)
Any ideas?
You need to use initContainers with Kubernetes git-sync https://github.com/kubernetes/git-sync/
Please see Answers to Question How to clone a private git repository into a kubernetes pod using ssh keys in secrets?
So the Kubernetes git-sync initContainer will pull a git repository into a local directory which is shared with your Drools Container via emptyDir volume mount.
For More Information on initContainers please read https://kubernetes.io/docs/concepts/workloads/pods/init-containers/
If you have a web AND a worker role in an Azure solution, all the waiting for the publishing an update package, uploading to the cloud storage, waiting for the package to be deployed could be exhausting, an waste a lot of time.
How to upload/deploy only the worker or web role of an Microsoft Azure Solution, that contains both roles, and save both internet traffic and time?
There is no option to build a package for only one of the both roles, but if you have limited bandwidth or traffic, and want to save from the upload time (which can be quite a big portion if you have a lot of static content: Look here for an example), there is one option.
As maybe you know, the package generated from Visual Studio for deployment (the 'cspkg' file) is nothing more, than an archive file.
Suppose, you want to update the WORKER role only. The steps are:
Create the update package as normal
Open it with the best archive manager (7zfm)
Inside, besides the other files are 2 'cssx' files - one for each
role. Delete the unnecessary cssx file.
Upload to Azure Blob Storage (optional)
Update the instances from the Azure Management Portal using the
'local' or 'storage' source as normal
On the Role dropdown, select only the role you want to update
Press OK :)
Hope this helps.
It is a lot easier to just add two additional cloud projects to your solution. In one project, have it reference only your web role. In the other project, have it reference only your worker role.
You can keep the cloud project that references both roles and use that for local debugging but when it is time to deploy, right click the cloud project that references only role you wish to deploy and click "Publish"
You will end up maintaining configuration files for each cloud project but that sounds a lot easier than manually messing around with editing the package file each time.