The minishift team could provide infrastructure to run CI against all provided add-ons.

The responsibilities would be

• minishift team to provide the infra
• addon owners to provide the tests

Hi,

After installing prometheus add-on, when trying to access it with displayed route, I was not able to access it with developer user. I could access with admin user. However, to do that I need to install/apply admin-user add-on.

It will be great to have a add-on that can help in deploying and setup Istio service mesh in minishift

The instructions follows:
https://github.com/snowdrop/istio-integration/blob/master/README.md

The end goal is we should be able to test these add-ons against specific OpenShift and Minishift versions. Contributors should be able to contribute a test file which would test the add-on. Also we should be able to test these add-ons for each Minishift release in an automated way.

This way we can maintain the quality of the repository and it will enable us to keep the add-ons in good shape.

From shell:
minishift addon apply kube-dashboard
-- Applying addon 'kube-dashboard':........Command "deploy" is deprecated, Use the rollout latest and rollout cancel commands instead.
.error: Deployment config "dashboard" waiting on manual update (use 'oc rollout latest dashboard')
Error applying the add-on: Error executing command 'oc rollout status dc/dashboard -n kube-dashboard'.

Minishift version: minishift v1.13.1+75352e5
System Version: macOS 10.13.2 (17C88)

Following worked:
-oc deploy dashboard -n kube-dashboard
+oc rollout latest dashboard -n kube-dashboard

Right now the command set that can be used in the addon file is limited to ssh, docker, oc etc., it will be great if we can add option to add extra commands like curl, wget, git etc.,

Che Minishift addon doesn't deploy Che properly.
I believe it is related to the fact that upstream Che switched Che 5 and Che 6 branches and now Che 6 is in master whereas Che Minishift addon expects it to be Che 5.
Here is the image declaration that is built by CI from master right now https://github.com/minishift/minishift-addons/blob/master/add-ons/che/templates/che-single-user.yml#L27.
Also, as @l0rd suggested we need a CI to fix a situation where Che addon is broken and devs don't know about that.

For example in cockpit/README.md there is a link to general readme, however this link is broken:

https://github.com/minishift/minishift-addons/blob/master/Readme.adoc#download-and-use-community-add-ons

Instead the README should be in upper case in the link to work:

https://github.com/minishift/minishift-addons/blob/master/README.adoc#download-and-use-community-add-ons

As per discussion https://github.com/minishift/minishift-addons/pull/69/files#r160385573 as there is no reason to keep the version at 3.7.0 cc @l0rd @jorgemoralespou

General-purpose web UI for Kubernetes clusters - https://github.com/kubernetes/dashboard

Those add ons that I have recently contributed lack READMEs. This issue is to track the addition of them.

Moved from minishift/minishift#1034

Lets create an addon so its easy to install fabric8 into minishift from the command line.

Currently the addon can't depend on another addon, it will be nice to add the add on dependency so that we can specify them via the addo scripts using include my-addon or when doing install i can add --dependencies='addon-a,addon-b' etc.,

It will be nice to have new add-on that helps in deploying RHOAR Launch, right now the steps are here https://appdev.openshift.io/docs/minishift-installation.html

It would be good to have a ready add-on template where contributor will just fill his details quickly.

It will reduce review comments like you need install section etc etc IMO.

I am thinking as

= {Name} Add-on
# Few lines of description for add-on

== Prepare Minishift [Optional]
# Maybe your add-on need Minishift to be run with extra configurations  

== Install add-on
# Steps to install add-on

== Apply {Name}
# Steps to apply your add-on

== Prerequisite steps for add-on [Optional]
# If your add-on required any steps to be performed before like Github integration etc.

== Remove {Name}
# Steps to remove your add-on

== Uninstall {Name} add-on [Optional]
# Steps to uninstall add-on

https://github.com/minishift/minishift/blob/master/Makefile#L235-L238

I'm getting compatibility issues between tiller V2.5 and the latest version of helm V2.7.

Moved from minishift/minishift#1000

As discussed in the devtools-program mailing list, it would be nice to have the Openshift template used in launch.openshift.io as part of the Minishift/CDK catalog

the current grafana plugin works with hawkular metrics - add one for prometheus

#79

Love the addon for Che. Really solves the problem with having to run unix scripts on windows to get things working. However, the replace stacks should really be put in it's own addon so that host system does not need to be unix based for Che to work.

https://github.com/minishift/minishift-addons/tree/master/add-ons/che#replace-che-stacks

With issues of broken links it would be great to have links checker in place which will run on every pull request.

Currently, we are have all existing add-on README's in markdown(.md) format.
However, our project README and contributing doc are in Asciidoc(.adoc).

Few PRs are there whose README are in Asciidoc - See Istio PR and in Markdown - See Ansible Service Broker PR.

We need to decide a consistent format here.

/cc @minishift/minishift-dev

Add-on provision http://cockpit-project.org/guide/latest/feature-kubernetes.html

Refer: minishift/minishift#1264

cc @l0rd @budhrg @gbraad

cc @l0rd

With that empty line minishift would throw the error

❯ minishift addon apply management-infra

-- Applying addon 'management-infra':Error applying the add-on: The variable(s)  are required by the add-on, but are not defined in the context

Istio add-on to be enhanced for supporting Initialisers, this includes the following steps:

1. Update the master-config.yaml to add Initializers plugin

(...)
admissionConfig:
  pluginConfig:
    GenericAdmissionWebhook:
      configuration:
        apiVersion: v1
        disable: false
        kind: DefaultAdmissionConfig
      location: ""
    Initializers:
      configuration:
        apiVersion: v1
        disable: false
        kind: DefaultAdmissionConfig
      location: ""
(...)

2. Enable Initializers API via apis/admissionregistration.k8s.io/v1alpha1=true ( this is by default with OpenShift), but good to check
3. Diable Istio for router and registry in default ns

oc annotate dc docker-registry sidecar.istio.io/inject='false' -n default
oc annotate dc router sidecar.istio.io/inject='false' -n default

4. Add cluster-admin role to istio-initializer-service-account in istio-system namespace to allow edit/watch Deployments

Hi @jorgemoralespou ,

I am getting some error on re-applying prometheus add-on after deleting it with provided command.

$ minishift addons apply prometheus --addon-env namespace=kube-system
-- Applying addon 'prometheus':.......
You have installed prometheus
To access prometheus go to https://prometheus-kube-system.192.168.42.141.nip.io

To delete:
    oc delete sa,clusterrolebinding,route,svc,secret,deployment,configmap -l app=prometheus -n kube-system --as=system:admin

$ oc delete sa,clusterrolebinding,route,svc,secret,deployment,configmap -l app=prometheus -n kube-system --as=system:admin
serviceaccount "prometheus" deleted
clusterrolebinding "prometheus-cluster-reader" deleted
service "prometheus" deleted
secret "prometheus-proxy" deleted
secret "prometheus-tls" deleted
deployment "prometheus" deleted
configmap "prometheus" deleted

$ minishift addons apply prometheus --addon-env namespace=kube-system
-- Applying addon 'prometheus':.    error: routes.route.openshift.io "prometheus" already exists
Error applying the add-on: Error executing command oc new-app -f prometheus.yaml -p NAMESPACE=#{namespace} -n #{namespace}.

Any idea?

The RAM size as mentioned in fabric8 usage guide is 5 GB and the command below

$ minishift start --memory=5000 --cpus=5 --disk-size=50g

uses 5000 which get printed as 4GB as per new pre-flight startup logs.

$ minishift start --memory 5000 --cpus=5 --disk-size=40g
-- Installing default add-ons ... OK
-- Checking if KVM driver is installed ... 
   Driver is available at /usr/local/bin/docker-machine-driver-kvm ... OK
-- Starting local OpenShift cluster using 'kvm' hypervisor ...
-- Minishift VM will be configured with ...
   Memory:    4 GB
   vCPUs :    5
   Disk size: 40 GB

We should tag/branch this repository when a minishift version happens so we can keep compatibility between the addons here provided and the minishift versions.

At some point in time we might want to provide 2 different versions of an addon, to support old minishift versions (CDK) and new minishift versions that provide new features (e.g. multiline descriptions in 1.5.0 that do not work in previous releases).

Add-on to provision Kubernetes package manager Helm - https://github.com/kubernetes/helm

Related to: minishift/minishift#1819

$ oc create --namespace openshift -f https://raw.githubusercontent.com/redhat-developer/s2i-dotnetcore/master/dotnet_imagestreams.json
$ oc create --namespace openshift -f https://raw.githubusercontent.com/redhat-developer/s2i-dotnetcore/master/templates/dotnet-example.json 
$ oc create --namespace openshift -f https://raw.githubusercontent.com/redhat-developer/s2i-dotnetcore/master/templates/dotnet-runtime-example.json 
$ oc create --namespace openshift -f https://raw.githubusercontent.com/redhat-developer/s2i-dotnetcore/master/templates/dotnet-pgsql-persistent.json

We are suggesting user to increase minishift memory to 5GB before enabling/applying che addon (in readme?).
It would be nice, if "minishift addon enable che" would check current minishift config and if it is below 5GB, it would print message like:

Che addon required at least 5GB RAM of minishift virtual machine. Your current config is at 2GB. Do you want to increase this to 5GB? [y/n]

Or at least print how it could be done.

See https://github.com/minishift/minishift-addons/tree/master/add-ons/workshop

There are other instances as well in different add-ons README.

With OpenShift 3.6.0

$ minishift status
Minishift:  Running
Profile:    os_360
OpenShift:  Running (openshift v3.6.0+c4dd4cf)
DiskUsage:  11% of 17.9G

$ minishift addons apply che
-- Applying addon 'che':
[CHE] Create Image Stream.
[CHE] Create the Che server Template.
[CHE] Creating mini-che project.
[CHE] Switching to mini-che....
[CHE] Deploying Che on minishift.
Please wait while the pods all startup!
You can watch in the OpenShift console via:
  minishift console
Then you should be able the open the Che dashboard here:
  http://che-mini-che.192.168.42.21.nip.io

With OpenShift 3.7.0

$ minishift status
Minishift:  Running
Profile:    os_370
OpenShift:  Running (openshift v3.7.0+7ed6862)
DiskUsage:  8% of 17.9G

$ minishift addon apply che
-- Applying addon 'che':
[CHE] Create Image Stream.
[CHE] Create the Che server Template.
[CHE] Creating mini-che project.Error from server (Forbidden): You may not request a new project via this API.
Error applying the add-on: Error executing command oc new-project mini-che --description="Eclipse Che on minishift" --as=developer.

When installing some addons we might have some pre-req like 4GB of RAM, 4 CPUS etc., it will be great if we can add option via metadata file or something so that when installing the addon the developer might know what the pre-req

the current addon works for 3.6 - but for 3.7 we also need the node exporter and alert manager

Prometheus is a monitoring framework for OpenShift. See https://github.com/openshift/origin/tree/master/examples/prometheus

Configures OpenShift to allow cross-origin requests from any address

We should add that a pull request commit message should have issue number e.g. Issue #<issue_number>

We have been following this in other minishift repositories and it helps to find out discussions , details with respect to a pull request very easily.

[1] https://github.com/minishift/minishift-addons/blob/master/CONTRIBUTING.adoc

https://github.com/minishift/minishift-addons/tree/master/add-ons/ansible-service-broker . Refer #76

• Update template to pull most recent template
• Pull templates from the upstream repo to ensure things stay up to
  date.
• Added an additional template that will permit the developer user
  correctly so that it can support apb cli functionality.
• Removed unnecessary templates/
• Added updated docs now that this README.md is considered the golden
  source for documenting running the broker and its tooling with minishift

Debezium is an opensource project for change data capture using a Kafka broker as a message bus. We would like to enable users to easily start and try Debezium

KubeVirt is a project to run Virtual Machines on Kubernetes and OpenShift.

KubeVirt can already be deployed on minikube, and a next step is to also get it running on OpenShift. it should then be provided as an add-on to minishift to make sure that it can be easily tested.

Things to be fixed

• Ability to run kubevirt on "arbitrary" clusters kubevirt/kubevirt#265
• Add macvtap kmod to the isos kubernetes/minikube#1845
• Add kvm modules
• Modify docker machine driver to do CPU passthrough
• code to deploy KubeVirt