minishift/minishift

Run OpenShift 3.x locally | This project does not see active developement and maintenance.

Minishift

• Welcome to Minishift!
• Getting started
• Documentation
• Community

Welcome to Minishift!

Minishift is a tool that helps you run OpenShift locally by running a single-node OpenShift cluster inside a VM. You can try out OpenShift or develop with it, day-to-day, on your local host.

Note	Minishift runs OpenShift 3.x clusters. Due to different installation methods, OpenShift 4.x clusters are not supported. To run OpenShift 4.x locally, use CodeReady Containers.

Minishift uses libmachine for provisioning VMs, and OpenShift Origin for running the cluster. The code base is forked from the Minikube project.

---

Getting started

To download the latest binary and review release notes, see the Minishift releases page.

Minishift requires a hypervisor to start the virtual machine on which the OpenShift cluster is provisioned. Make sure that the hypervisor of your choice is installed and enabled on your system before you start Minishift.

For detailed installation instructions for Minishift and the required dependencies, see the Getting started documentation.

Documentation

Minishift documentation is published as a part of the OpenShift Origin documentation library. Check out the latest official Minishift documentation for information about getting started, using, and contributing to Minishift:

• Getting started with Minishift

• Using Minishift

• Developing and contributing to Minishift

• Command Reference

In addition, you can review the release notes and project roadmap here on GitHub:

• Release Notes

• Roadmap

Community

Contributions, questions, and comments are all welcomed and encouraged!

You can reach the Minishift community by:

• Signing up to our mailing list

• Joining the #minishift channel on Freenode IRC

For information about contributing, building, and releasing Minishift, as well as guidelines for writing documentation, see the Contributing to Minishift topics.

If you want to contribute, make sure to follow the contribution guidelines when you open issues or submit pull requests.

minishift's Issues

[0.3.0] Stuck on Windows 10 and VirtualBox

minishift 0.3.0
Windows 10 64-bit
VirtualBox 5.0.26

Stuck at:

>minishift-windows-amd64.exe start --cpus=2 --memory=4096 --vm-driver=virtualbox
Starting local OpenShift cluster...

Initially it did not work but showed that openshift start had been attempted in the VM. minishift start finished and said to connect to http://192.168.99.100:8433. That redirected to a 10.x.x.x address and didn't work.

After I disabled firewall and anti-virus (Kaspersky), it got to the state I'm reporting in this issue where minishift start --cpus=2 --memory=4096 --vm-driver=virtualbox does not finish (in 5-10min at least).

Add update command

It would be good with an update command as not many users would know that its the status command that can tell if there is a newer version and do an auto update.

Implement HTTP proxy support

Setup is on Windows 7, VirtualBox 5.1.4, and minishift version: 0.6.0. Both http_proxy and https_proxy are set and point to my cntlm installation. Here is the output from minishift

$ minishift start --cpus=2 --memory=4096 --vm-driver=virtualbox 
Starting local OpenShift cluster...
E0914 10:58:12.980074   13152 start.go:89] Error starting host: Error creating. Error creating machine: Error in driver during machine creation: open /.minishift/cache/iso/boot2docker.iso: The system cannot find the path specified.. Retrying.
E0914 10:58:14.438220   13152 start.go:89] Error starting host: Error getting state for host: machine does not exist. Retrying.
E0914 10:58:15.868363   13152 start.go:89] Error starting host: Error getting state for host: machine does not exist. Retrying.
E0914 10:58:15.868363   13152 start.go:95] Error starting host:  Error creating. Error creating machine: Error in driver during machine creation: open /.minishift/cache/iso/boot2docker.iso: The system cannot find the path specified.
Error getting state for host: machine does not exist
Error getting state for host: machine does not exist

What help can I provide to help troubleshoot this issue?

Metrics via cadvisor broken

Requires upgrade to OpenShift 1.3.0-alpha.3 once it's released.

Rename dashboard command to console

And update log message.

Document the more value of minishift

Hi,

I think it would be valuable to have a document which explains what 'minishift' great features is and what are the limitations it has as oppose to OpenShift.

Few of the features and limitations are listed in README#features and README#known-issues respectively.

Thanks

Service url is not reachable in minishift setup

Hi,

I am trying to deploy one of the Karaf rest example from the quickstarts into my local minishift environment. I deployed it and service up and running succesfully.
However i can not access to service url, here is the outputs

Pod:
[eiltpeh@localhost blueprint]$ oc get pods
NAME READY STATUS RESTARTS AGE
exposecontroller-1-au4i4 1/1 Running 1 23h
fabric8-docker-registry-1-6kpgp 1/1 Running 2 23h
fabric8-forge-1-xmdbx 1/1 Running 1 23h
fabric8-q1anx 1/1 Running 2 23h
gogs-1-p1gjr 1/1 Running 1 23h
jenkins-1-xdkhq 1/1 Running 1 23h
karaf-camel-rest-sql-1-ldwbe 1/1 Running 0 12m
mysql-4-9vps0 1/1 Running 2 23h
nexus-1-aoiyd 1/1 Running 9 23h
router-1-dcv6f 2/2 Running 4 23h

Routes:
[eiltpeh@localhost blueprint]$ oc get routes
NAME HOST/PORT PATH SERVICE
karaf-camel-rest-sql karaf-camel-rest-sql-default.192.168.99.100.xip.io karaf-camel-rest-sql
mysql mysql.xip.io mysql
router router.xip.io router

[eiltpeh@localhost blueprint]$ curl http://karaf-camel-rest-sql-default.192.168.99.100.xip.io

503 Service Unavailable

No server is available to handle this request.

[eiltpeh@localhost blueprint]$ curl http://karaf-camel-rest-sql-default.192.168.99.100.xip.io/camel-rest-sql/books

503 Service Unavailable

No server is available to handle this request.

Minishift service IP:
[eiltpeh@localhost blueprint]$ minishift service karaf-camel-rest-sql --url
http://192.168.99.100:32449
[eiltpeh@localhost blueprint]$ curl http://192.168.99.100:32449
curl: (7) Failed connect to 192.168.99.100:32449; Connection refused

Kubernetes Port:
[eiltpeh@localhost blueprint]$ kubectl get service karaf-camel-rest-sql --output='jsonpath="{.spec.ports[0].NodePort}"'
"32449"[eiltpeh@localhost blueprint]$

I disabled firewall, ipv6 yet it is still not working
Could you please help ?

Thanks

Project name collision

There is another minishift project in github: https://github.com/arachnidlabs/minishift

minishift dashboard does not target new port 8443

With minishift version 0.2.1 I have console at: https://192.168.64.2:8443 but when I run minishift dashboard it opens https://192.168.64.2 (No port 8443)

list services and how they can be accessed

Right now if we do minishift service foo it will try open the service for me in a browser (or whatever app)... would love to be able to just minishift service --list and show all the services and how to access them

minishift service <service-name> invalid url/404

when doing minishift service takes to the right url and then a redirection happens to a vagrant.f8 urls resulting in HTTP 404 and when adding hosts entries in /etc/hosts throws the the following error in browser

{"error":"invalid_request","error_description":"The request is missing a required parameter, includes an invalid parameter value, includes a parameter more than once, or is otherwise malformed.","state":"http://192.168.64.23:31809/"}

Windows, 'Error creating new host: dial tcp: missing address' error from minishift start

Finally had a chance to give this a go, using the latest release:

PS C:\Users\gashcrumb> minishift start --memory=8192
Starting local OpenShift cluster...
E0921 09:11:57.932205   16704 start.go:89] Error starting host: Error creating new host: dial tcp: missing address. Retrying.
E0921 09:11:58.072763   16704 start.go:89] Error starting host: Error creating new host: dial tcp: missing address. Retrying.
E0921 09:11:58.210321   16704 start.go:89] Error starting host: Error creating new host: dial tcp: missing address. Retrying.
E0921 09:11:58.211320   16704 start.go:95] Error starting host:  Error creating new host: dial tcp: missing address
Error creating new host: dial tcp: missing address
Error creating new host: dial tcp: missing address

Am using hyperv, also minikube works fine for me on this machine.

Upgrade to openshift 1.3.0 GA

minishift service foo - Should support service context-path annotations

A service can be annotated in kubernetes to include a prefix path for the service, such as /hello in this example: https://github.com/rhuss/fabric8-maven-plugin/blob/master/samples/external-resources/src/main/fabric8/service.yml

In my example then I type minshift service hello and it opens

http://192.168.64.5:32623/

But with the annotated the context-path should include /hello eg so its

http://192.168.64.5:32623/hello

In my own project I manually declare a service so I can specify that context-path annotation but also show how to chose port numbers for the service

I have a src/main/fabric8/service.yml file with some kubenetes metadata:

metadata:
  annotations:
    api.service.kubernetes.io/path: /hello
  name: "hello"
spec:
  ports:
  - port: 8181
    protocol: "TCP"
    targetPort: 8080
  type: LoadBalancer

Allow the subdomain to be configured and default it to an xip.io default OOTB

When you create a Route on MiniShift with an empty Host it gets defaulted to: $(name)-$(namespace).router.default.svc.cluster.local which doesn't work on the host and there's no way to change it.

It'd be nice if minishift start had a subdomain parameter which defaulted to $(minishift ip).xip.io so that if the Router was used with MiniShift that they work.

On our old Vagrant VM we configured the subdomain using a master yaml file change here:
https://github.com/fabric8io/fabric8-installer/blob/master/vagrant/openshift/Vagrantfile#L100-L101

I had a peek into the MiniShift VM and wasn't sure where the master yaml config file is; am guessing we're not using one right now?

Suggestion : Expose minishift cmds as REST endpoints

This could be interesting that we wrap the minishift commands around Rest endpoints (http://thenewstack.io/make-a-restful-json-api-go/). That could help us like also with the oc client to develop tools required for the developer

Does it support Ingress?

Minikube has added support for Ingress recently in addition to NodePort to allow exposing apps to the network. Does Minishift support that? and if not, what are the plans for it?

Starting minishift again after stopping it stuck on "Starting local cluster..." message

Hi,

After performing minishift stop -> minishift start -> minishift stop -> minishift start, it looks like the command got stuck. Command not ending.

$ minishift status
Stopped

$ minishift start
Starting local OpenShift cluster...
...# not ending

minishift logs - https://gist.github.com/budhrg/a389fa1cb5f2e53fb826e7049173f05d.

Continuously getting following kind of errors in logs:

[1.999827422s] [1.999824135s] Etcd node listed ...

Any support to get cluster metrics? If yes, how?

Hi,

Is there any support for cluster metrics? Is it configurable?

Something similar to this https://docs.openshift.com/enterprise/3.1/install_config/cluster_metrics.html.

What can we do with option get-openshift-versions?

The command minishift get-openshift-versions lists all the OpenShift versions which are available :

 - v1.3.0-rc1
- v1.3.0-alpha.3
- v1.2.1
- v1.3.0-alpha.2
- v1.3.0-alpha.1
...

But what can we do ? Is it possible to select the version that we want to deploy when we create a minishift instance ?

Cannot open Fabric8 console.

I have installed minishift and I can access the different application consoles as per the service command with the exception of the fabric8 console. I deployed wiht the following options:

gofabric8 deploy -y --domain=$(192.168.122.31).xip.io --api-server=$(192.168.122.31)

I get a what looks like a redirect and then a blank page.

Question: Why the ROADMAP same as that of minikube?

Hi,

I found that the roadmap of minishift is exactly same as minkube.
Isn't it the roadmap of minishift should/will be different that minikube?

Although, I understand minishift is being build on top of minikube but the goal is obviously different IMO which might make the roadmap different too.

Let me know if I am missing anything here.

Thanks.

lets set a high default DeploymentConfig timeout?

see this issue:
openshift/origin#10531

since fabric8 moved to DeploymentConfig on OpenShift, I get failures where the DeploymentConfig deployer pod times out and stops trying to deploy - even though the docker image hasn't been downloaded yet.

If there is a way to do so - it'd be nice for minishift to set the global DeploymentConfig timeout value (the docs seem to hint there is one - not sure if its configurable though ;). Then we could give folks laptops plenty of time to try download docker images.

Should we have 'minishift install-cli' to download, make it executable 'oc' binary?

Hi,

As a part of better user experience, can we have 'oc' binary installation as part of minishift CLI command?
It can be:

minishift install-oc (or any name)
# instruction to export binary path
# or eval command for $(minishift install-oc)

Right now, user has to find the 'oc' binary by its own.

Add new subcommand to change openshift config

Ala openshift ex config patch.. & restart openshift after changes applied.

0.3.2 fails to start on OSX

This was working, but now not - I'm getting:
Error starting host: Error starting stopped host: Machine didn't return an IP after 120 seconds, aborting

Error doing docker ps when client is newer

I have docker for mac beta installed, and when I do docker ps I get the following error:

$eval $(minishift docker-env)
$docker ps
Error response from daemon: client is newer than server (client API version: 1.24, server API version: 1.23)

This is the docker version output:

Client:
 Version:      1.12.0-rc2
 API version:  1.24
 Go version:   go1.6.2
 Git commit:   906eacd
 Built:        Fri Jun 17 20:35:33 2016
 OS/Arch:      darwin/amd64
 Experimental: true
Error response from daemon: client is newer than server (client API version: 1.24, server API version: 1.23)

Ability to configure templates and other resources

The CDK provides a list of Kubernetes applications (= templates) that we can use to create some pods (EAP, Ruby, Php, NodeJs, ...).

I would like to suggest that we have an option to deploy --add-templates such templates within minishift :

Here is a list of such templates :

DBs:
https://github.com/openshift/origin/tree/master/examples/db-templates

Sample apps/Quickstarts:
https://github.com/openshift/origin/tree/master/examples/quickstarts (note, the files in this dir are drawn from upstream repos as described here: https://github.com/openshift/origin/blob/master/examples/quickstarts/README.md

Jenkins:
https://github.com/openshift/origin/tree/master/examples/jenkins

Middleware image templates:
https://github.com/jboss-openshift/application-templates

xPaas:
https://github.com/openshift/openshift-ansible/tree/master/roles/openshift_examples/files/examples/v1.3

minishift start --deploy-registry --deploy-router on existing instance failed

Using minishift 0.4.1, i have performed the following commands on existing instance that don't have router or registry installed:

minishift stop
minishift start --deploy-registry --deploy-router
Starting local OpenShift cluster...
E0825 11:21:14.608549   17105 start.go:104] Error starting cluster:  Something went wrong running an SSH command!
command : 
cd /var/lib/minishift;
sudo /usr/local/bin/openshift admin policy add-scc-to-user hostnetwork -z router --config=openshift.local.config/master/admin.kubeconfig;
sudo /usr/local/bin/openshift admin router --service-account=router --config=openshift.local.config/master/admin.kubeconfig

err     : exit status 1
output  : info: password for stats user admin has been set to JRRKM8qdZv
--> Creating router router ...
    error: serviceaccounts "router" already exists
    clusterrolebinding "router-router-role" created
    deploymentconfig "router" created
    service "router" created
--> Failed

Even if the command failed, when i have started the OpenShift cluster with minishift start only,
the router & registry has been deployed correctly:

$ oc get pods 
NAME                            READY     STATUS    RESTARTS   AGE
docker-registry-1-zcscb         1/1       Running   0          6h
fabric8-docker-registry-f7md5   1/1       Running   6          8d
fabric8-forge-nr9nm             1/1       Running   6          8d
fabric8-re8ps                   1/1       Running   6          8d
gogs-xo7uk                      1/1       Running   7          8d
jenkins-ss9yg                   1/1       Running   5          8d
nexus-0iqsx                     1/1       Running   15         8d
router-1-43h6y                  1/1       Running   0          6h

not able to set `eval $(minishift docker-env)`

Hi,

I'm trying to set the docker environment variables to my local docker client by issuing:

eval $(minishift docker-env)

However I'm seeing that the execution goes in to background mode.

I'm using the latest docker for mac
I'm only seeing this behavior after I updated to the latest version of minishift.

Here is the output for minishift docker-env command:

zhassan:~ zhassan$ minishift docker-env
export DOCKER_TLS_VERIFY="1"
export DOCKER_HOST="tcp://192.168.64.11:2376"
export DOCKER_CERT_PATH="/Users/zhassan/.minishift/certs"
# Run this command to configure your shell:
# eval $(minishift docker-env)

Volume mounted under /var/cache & permission denied

This error is reported using this config to mount a volume

kind: PersistentVolume
apiVersion: v1
metadata:
  name: pv0001
  labels:
    type: local
spec:
  capacity:
    storage: 10Gi
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: "/tmp/data01"


kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: myclaim-1
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 3Gi

kind: Pod
apiVersion: v1
metadata:
  name: mypod
  labels:
    name: frontendhttp
spec:
  containers:
    - name: myfrontend
      image: nginx
      ports:
        - containerPort: 80
          name: "http-server"
      volumeMounts:
      - mountPath: "/usr/share/nginx/html"
        name: mypd
  volumes:
    - name: mypd
      persistentVolumeClaim:
       claimName: myclaim-1      

using minishift 0.7.1 & xhyve docker-machine 0.3.2

2016/09/23 15:58:10 [warn] 1#1: the "user" directive makes sense only if the master process runs with super-user privileges, ignored in /etc/nginx/nginx.conf:2
nginx: [warn] the "user" directive makes sense only if the master process runs with super-user privileges, ignored in /etc/nginx/nginx.conf:2
2016/09/23 15:58:10 [emerg] 1#1: mkdir() "/var/cache/nginx/client_temp" failed (13: Permission denied)
nginx: [emerg] mkdir() "/var/cache/nginx/client_temp" failed (13: Permission denied)

issue after update to 0.6.0

Like usual after minishift start, i have accepted the update to 0.6.0, and i have received the following error:

$ ./minishift start
There is a newer version of minishift available (v0.6.0). Do you want to
automatically update now [y/N]? y
Updating to version 0.6.0
E0908 20:07:23.880468    4614 update.go:188] Failed to exec updated binary: no such file or directory
[aboucham@aboucham minishift]$ ./minishift start
Starting local OpenShift cluster...
oc is now configured to use the cluster.
Run this command to use the cluster: 
oc login --username=admin --password=admin

Note that, after 2 nd attempt, it works.

In case of expired login then maybe make it easy to login from minishift

I left my laptop the night over and now when I get back to minishift to open the web console with

davsclaus:/Users/davsclaus/Documents/workspace/client/$ minishift service fabric8
Error getting fabric8 service: the server has asked for the client to provide credentials (get services fabric8)
Check that minishift is running and that you have specified the correct namespace (-n flag).
davsclaus:/Users/davsclaus/Documents/workspace/client/$ minishift status
Running

If I do an oc login and then afterwards the minishift service command works for opening the web console.

I wonder if minishift can report something about expired login, or even allow to login?

routes & DNS not working on OS X

nice work on minishift; it worked OOTB! I ran gofabric8 and it got everything up and running; the only downside is I can't access any of the routes from my host. I'm guessing its that we don't have a minishift equivalent of landrush. I could hack around it with /etc/hosts but its a bit of a shame. e.g. I can open the console via

minishift service fabric8

but the OAuth doesn't work as its redirecting to vagrant.f8 which doesn't resolve. I guess gofabric8 needs to remove all Routes if running on minishift so that we don't try to use DNS names and then tries to use the ip address & nodePort for all links etc

Minishift running in Docker container

Just a crazy idea: Would it be possible to let minishift run within a Docker container ? Or is this completely off since this doesn't require to setup an external Docker daemon (maybe there are already other solutions for this, too).

The benefits would be:

• Extended functionality like adding a registry by setting env vars, starting gofabric8 etc. could be easily provided by the Docker image or a sub-image.
• Can be easily used in integration tests e.g. by fabric8-maven-plugin
• Even easier installation when a Docker daemon is already present.

As I said, this could be completely off-focus of minishift, at their could be already alternatives which provides this in a better way.

Add support for creating resources on start

it'd be awesome if we could easily do a minishift start which would automatically install fabric8 as well; using a command line flag maybe?

We could maybe get gofabric8 to generate a big blob of YAML we could use to install everything or something to minimise the extra code required inside minishift

'--iso-url' works with other iso other than boot2docker.iso?

Hi,

Does the option --iso-url works with other type of iso other than boot2docker.iso?

I tried with rancheros.iso but looks like its not working.

➜  minishift git:(master) ✗ minishift start --memory=4000 --iso-url='https://github.com/rancher/os/releases/download/v0.6.1/rancheros.iso' --vm-driver=virtualbox
Starting local OpenShift cluster...
Downloading ISO
 41.00 MB / 41.00 MB [=============================================================================================================================] 100.00% 0
E0921 00:33:58.260267   22248 start.go:100] Error updating cluster:  Process exited with: 1. Reason was:  ()

Running Specific Version of OpenShift with minishift

is it possible to control the version of openshift one can use with minishift ? I see there is 'get-openshift-versions` but I dont see any option to set the required version from the list.

oc login changes the current context from minishift

The current context is changed following an oc login. Steps to recreate:

After minishift starts the current context is minishift

➜  minishift start
Starting local OpenShift cluster...
oc is now configured to use the cluster.
Run this command to use the cluster:
oc login --username=admin --password=admin
➜  kubectl config current-context
minishift

login

➜ oc login --username=admin --password=admin
Login successful.

You have access to the following projects and can switch between them with 'oc project <projectname>':

  * default
    kube-system
    openshift
    openshift-infra

Using project "default".

check the current context again

➜ kubectl config current-context
default/192-168-64-89:8443/admin

minishift start to annotate the Node with the output of `minishift ip` so that the external IP can be found

this is a copy of this issue in minikube kubernetes/minikube#544; which will probably get fixed by minikube but figured i'd put it here to be sure it gets done

As a workaround we can add the annotation if its missing in gofabric8 too

Tell current version in auto update

So the user know what he/she current runs and what the new version is. For example if there is patch releases then it can be assumed safter to update from 1.0.2 to 1.0.3 than from 1.0.2 to 1.2.0 etc.

Is a new boot2docker.iso required for each minishift update

The first thing that happens after every minishift update is that a new boot2docker is is pulled down. Is this .iso required to be versioned with minishift, or could just reference a concrete boot2docker version and maybe be split into 2 repositories?

minishift version displays minikube

When run this command:

$minishift version

Got this output:

minikube version: 0.2.1

create a new release to use the latest greatest origin release

currently thats v1.3.0-alpha.3 at the time of writing - minishift uses v1.3.0-alpha.1 right now
https://github.com/openshift/origin/releases

No installation instructions are provided

The readme.md file refers to installation instructions part of the releases section which are not there

Additional point:

Is this file part of a release (minishift-darwin-amd64) a go binary, .... ?

Specify the IP of the cluster

It would be good to specify the IP address of the cluster rather than having one assigned from the pool and then having to query it to find out what it is.

brew install minishift

It would be super cool to have a brew install command for minishift which could handle updates etc. as well.
And yes I know it's super easy to download and install a binary but devs are lazy and aren't always aware when there are updates - brew does a great job of handling this.

docker daemon need to start with insecure registry

If we want to work with openshift as developer then inside minishift docker daemon should be running with --insecure-registry 172.30.0.0/16.

[prkumar@dhcp193-119 minishift]$ ./out/minishift docker-env
export DOCKER_HOST=tcp://192.168.42.110:2376
export DOCKER_CERT_PATH=/home/prkumar/.minishift/certs
export DOCKER_TLS_VERIFY=1
# Run this command to configure your shell: 
# eval $(minishift docker-env)

[prkumar@dhcp193-119 minishift]$ eval $(./out/minishift docker-env)

[prkumar@dhcp193-119 minishift]$ docker images
REPOSITORY          TAG                 IMAGE ID            CREATED             SIZE

[prkumar@dhcp193-119 minishift]$ ./oc cluster up
-- Checking Docker client ... OK
-- Checking for existing OpenShift container ... OK
-- Checking for openshift/origin:latest image ... 
   Pulling image openshift/origin:latest
   Pulled 1/4 layers, 26% complete
   Pulled 2/4 layers, 70% complete
   Pulled 3/4 layers, 81% complete
   Pulled 3/4 layers, 95% complete
   Pulled 4/4 layers, 100% complete
   Extracting
   Image pull complete
-- Checking Docker daemon configuration ... FAIL
   Error: did not detect an --insecure-registry argument on the Docker daemon
   Solution:

     Ensure that the Docker daemon is running with the following argument:
        --insecure-registry 172.30.0.0/16

Windows?

Windows support would be pretty awesome...

lets create an openshift registry OOTB so that ImageStreams and S2I can be used

the default way to build things on OpenShift is via ImageStreams which requires the openshift registry; so we should probably default to creating that (but make it so folks could opt out maybe?)