# openSHift Running In Managed Podman

Kubernetes upstream has KIND, let me introduced SHRIMP, a way to quickly deploy a containerized a Microshift Kubernetes on any Linux system having a podman package.
[Microshift](https://next.redhat.com/project/microshift/) is an all in on binary of the OpenShift core components, targeted for edge deployments. The Microshift RPM and the cri-o container runtime are all that you need to run a single node OpenShift/Kubernetes cluster.
Alternatively, you can run Microshift inside an all-in-one container, allowing you to run Microshift on any kind of Linux host. This setup is perfect for small lab or developement environements using a minimal amount of resources and this is what were going to cover in this blog post.

Note that if you want the whole OpenShift Container Platform experience on your laptop, there is [OpenShift Local](https://developers.redhat.com/products/openshift-local/overview), which comes with the standard Cluster Operators in a VM and gives you the real Platform As A Service environment of OpenShift. A great product, OpenShift Local is nevertheless ressource hungry and will happily consumate four full cores and 8GB of RAM (and kicks the CPU fan of the laptop quite often)

Build the container image
-------------------------

~~~
$ git clone [e-mail]:openshift/microshift.git
$ cd microshift
$ make build-containerized-all-in-one-amd64
~~~

Now publish this container image in a container registry, or use the public image I built in https://quay.io/repository/rhn_support_ekasprzy/microshift-aio

Prepare podman to run Microshift in a container
-----------------------------------------------

We need to change the storage driver to **vfs**, so that we can start **overlayfs** inside the container:
~~~
cat /etc/containers/storage.conf
[storage]
driver="vfs"
~~~

We will create a specific bridge for this installation, so that the microshift installation is isolated from other containers:

~~~
sudo podman network create --subnet 10.87.87.0/24 --gateway 10.87.87.1 shrimpnet
~~~

Finally start the container:
~~~
sudo podman run --name shrimp --privileged \
--ip=10.87.87.2 --network=shrimpnet \
--publish 127.0.0.1:6443:6443 --log-level=info \
quay.io/rhn_support_ekasprzy/microshift-aio:4.10.0-0.microshift-2022-09-19-ubi-8.6
~~~

Verify that the MicroShift installation is running
--------------------------------------------------

Once the container is started, verify that we have the cri-o and microshit services running inside the shrimp:
~~~
sudo podman exec shrimp systemctl status cri-o
...
systemd[1]: Started Container Runtime Interface for OCI (CRI-O).
~~~

We can now extract the oc client and a kubeconfig to access the API server running inside the container:

~~~
mkdir .kube
sudo podman cp shrimp:/var/lib/microshift/resources/kubeadmin/kubeconfig $HOME/.kube/config
sudo chown $USER:$USER $HOME/.kube/config
sudo podman cp shrimp:/usr/local/bin/oc /usr/local/bin/oc
~~~

We now verify that we have a running API server
~~~
$ oc cluster-info
Kubernetes control plane is running at https://127.0.0.1:6443
~~~

and a running single node cluster
~~~
$ oc get nodes -o wide
NAME STATUS ROLES AGE VERSION INTERNAL-IP EXTERNAL-IP OS-IMAGE KERNEL-VERSION CONTAINER-RUNTIME
0e47b5635d8a Ready <none> 41h v1.21.0 10.87.87.2 <none> Red Hat Enterprise Linux 8.6 (Ootpa) 5.10.0-16-amd64 cri-o://1.23.3
~~~

as we see from the output above, the node internal IP matches the `--ip` parameter we gave to podman when creating the shrimp

Optionally create a systemd unit to run microshift on boot
----------------------------------------------------------

Finally we generate a systemd unit based on the running container, so we can autostart the shrimp on boot
~~~
sudo podman stop shrimp
sudo podman generate systemd shrimp | sudo tee -a /etc/systemd/system/shrimp.service
~~~

We can now use the systemd unit to run the shrimp
~~~
sudo systemctl enable --now shrimp
~~~
~~~
$ systemctl status shrimp
● shrimp.service - Podman container-46fc1b252bb0d0ba707a2210f06783cf28aa1d88faf2de1874b685087640a6fa.serv>
Loaded: loaded (/etc/systemd/system/shrimp.service; enabled; vendor preset: enabled)
Active: active (running) since Tue 2022-09-06 12:22:24 UTC; 31min ago
Docs: man:podman-generate-systemd(1)
Process: 255 ExecStart=/usr/bin/podman start 46fc1b252bb0d0ba707a2210f06783cf28aa1d88faf2de1874b6850>
Main PID: 518 (conmon)
Tasks: 2 (limit: 2339)
Memory: 14.9M
CPU: 209ms
CGroup: /system.slice/shrimp.service
└─518 /usr/bin/conmon --api-version 1 -c 46fc1b252bb0d0ba707a2210f06783cf28aa1d88faf2de187
~~~

Create a sample deployment and access it
----------------------------------------

~~~
$ oc new-project hello-nginx
$ oc create deployment hello-nginx --image=docker.io/library/nginx
$ oc create service nodeport hello-nginx --node-port=30003 --tcp=8080:80
$ w3m -dump http://10.87.87.2:30003

Welcome to nginx!
...
~~~