Skip to content

A virtualized lab environment for installing OpenShift 4.x baremetal disconnected using a Fedora CoreOS utility node.

Notifications You must be signed in to change notification settings

aerdei/ocp4-upi-virt-fcos

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

60 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

OpenShift 4.5 UPI disconnected on Libvirt with a Fedora CoreOS utility node

Diagram

Prerequisites

libvirt
virt-install
openvswitch
fcct
xz

Deploy utility node

Create the Libvirt network:

virsh net-define --file ocp4.xml && virsh net-start ocp4

Create an Open vSwitch bridge:

ovs-vsctl add-br ovsbr

Download Fedora CoreOS:

curl https://builds.coreos.fedoraproject.org/prod/streams/stable/builds/32.20200726.3.1/x86_64/fedora-coreos-32.20200726.3.1-qemu.x86_64.qcow2.xz -o /var/lib/libvirt/images/fedora-coreos-32.20200726.3.1-qemu.x86_64.qcow2.xz && xz --decompress /var/lib/libvirt/images/fedora-coreos-32.20200726.3.1-qemu.x86_64.qcow2.xz

Add your pull secret to .storage.files.$11 and your host's SSH key to passwd.users.$0.ssh_authorized_keys in utility.fcc.
Generate ignition config:

fcct -strict -pretty < utility.fcc > /var/lib/libvirt/images/utility.ign

Start Fedora CoreOS with the ignition config:

virt-install --connect qemu:///system --import --name utility.ocp.example.com --network network=ocp4,mac=12:34:56:00:00:53 --network bridge=ovsbr,mac=12:34:56:00:00:54,virtualport_type=openvswitch --ram 1024 --vcpus 1 --os-variant fedora29 --disk size=50,backing_store=/var/lib/libvirt/images/fedora-coreos-32.20200726.3.1-qemu.x86_64.qcow2,format=qcow2,bus=virtio --qemu-commandline="-fw_cfg name=opt/com.coreos/config,file=/var/lib/libvirt/images/utility.ign" --vnc --noautoconsole

When the utility node is up, you can start preparing the OpenShift installation.

Deploy OpenShift 4.5

Prepare the installer

Prepare RHCOS images for the install in the iso_rhcos directory.

wget https://mirror.openshift.com/pub/openshift-v4/dependencies/rhcos/4.5/latest/rhcos-installer.x86_64.iso https://mirror.openshift.com/pub/openshift-v4/dependencies/rhcos/4.5/latest/rhcos-installer-kernel-x86_64 https://mirror.openshift.com/pub/openshift-v4/dependencies/rhcos/4.5/latest/rhcos-installer-initramfs.x86_64.img -N -P ./iso_rhcos/

Download the extracted installer to the host:

curl http://192.168.122.53:8080/openshift-install -o openshift-install
chmod u+x ./openshift-install

Configure install-config.yaml in the cluster directory.
Create installer files and upload them to the utility VM:

./openshift-install create manifests --dir=cluster/
sed -i 's/mastersSchedulable: true/mastersSchedulable: false/g' cluster/manifests/cluster-scheduler-02-config.yml
./openshift-install create ignition-configs --dir=cluster/
rsync -Pavr --chmod=o+r ./cluster/* core@192.168.122.53:/opt/services/httpd/www/html/

Deploy nodes

Deploy bootstrap node:

bash deploy.sh bootstrap example.com

Note that you need to manually start all the nodes as they will not automatically reboot:

bash start.sh

When the bootstrap node is up, deploy the masters:

bash deploy.sh masters example.com

When the masters are up and the bootstrap process is complete, you are ready to destroy the bootstrap node and deploy the worker nodes:

virsh destroy bootstrap-0.ocp.example.com
virsh undefine bootstrap-0.ocp.example.com --remove-all-storage
bash deploy.sh workers example.com

Make sure to approve any CSRs that are generated while provisioning the workers:

export KUBECONFIG=$(pwd)/cluster/auth/kubeconfig
oc login -u system:admin
oc get csr -o go-template --template='{{range .items}}{{if  not .status}}{{printf "%s\n" .metadata.name}}{{end}}{{end}}' | xargs -i oc adm certificate approve {}

Cleanup

To destroy and undefine every virtual machine related to this cluster, run cleanup.sh.

Known issues

  • At startup, Fedora CoreOS will request IPs through DHCP on both of its interfaces. This should work on enp1s0, but a timeout must happen on enp2s0 for the boot process to continue.
  • RHCOS machines configured with virt-install currently don't reboot automatically. You need to invoke start.sh once they are in a shut down state.

About

A virtualized lab environment for installing OpenShift 4.x baremetal disconnected using a Fedora CoreOS utility node.

Topics

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages