An example Ubuntu virtual machine running in the Oracle Cloud Infrastructure (OCI) cloud.

For a similar example running in Azure see the rgl/azure-ubuntu-vm repository.

Usage (on a Ubuntu Desktop)

Install the tools:

./provision-tools.sh

Restart your shell:

exit

Review the main.tf file.

Login into oci:

# NB this saves the details at ~/.oci/config
# NB you really have to choose a region. without any arguments it will ask it.
# NB if you use another region, you must modify the vm_image_ocid variable
#    inside the main.tf file.
# NB this authentication session in only valid for one hour.
oci session authenticate --region eu-amsterdam-1 # NB save as the DEFAULT profile

Try using oci by listing the regions and users:

oci iam region list --auth security_token
oci iam user list --auth security_token

Initialize terraform:

make terraform-init

Launch the example:

make terraform-plan
make terraform-apply

At VM initialization time cloud-init will run the provision-app.sh script to launch the example application.

Wait for cloud-init to finish:

while ! ssh "ubuntu@$(terraform output -raw vm_ip_address)" cloud-init status --wait --long; do sleep 5; done

NB The cloud-init logs are at /var/log/cloud-init-output.log.

Test the app endpoint:

wget -qO- "http://$(terraform output -raw vm_ip_address)/test"

Connect to the VM serial console:

NB the console requires login, which means you must have previously set the ubuntu user password (by default it does not have a password; only ssh key login).

# NB the ssh command is alike:
#       ssh -o ProxyCommand='ssh -W %h:%p -p 443 ocid1.instanceconsoleconnection.oc1.eu-amsterdam-1.<id1>@instance-console.eu-amsterdam-1.oci.oraclecloud.com' ocid1.instance.oc1.eu-amsterdam-1.<id2>
bash -c "$(terraform output -raw vm_serial_console_ssh_command)"

You can also connect to the VNC console of the VM:

# NB the ssh command is alike:
#       ssh -o ProxyCommand='ssh -W %h:%p -p 443 ocid1.instanceconsoleconnection.oc1.eu-amsterdam-1.<id1>@instance-console.eu-amsterdam-1.oci.oraclecloud.com'-N -L localhost:5900:ocid1.instance.oc1.eu-amsterdam-1.<id2>:5900 ocid1.instance.oc1.eu-amsterdam-1.<id2>
bash -c "$(terraform output -raw vm_vnc_console_ssh_command)" & # start the tunnel in background.
vinagre localhost:5900 # open a VNC connection tru the local tunnel.

Connect to the VM and start a Debian LXC system container:

ssh "ubuntu@$(terraform output -raw vm_ip_address)" # enter the VM.
snap list lxd # show the lxd package version.
journalctl -u snap.lxd.daemon.service # show lxd logs.
lxc launch images:debian/11 debian # start the container.
lxc exec debian -- bash # enter the container.
lscpu
# NB if the container does not obtain an IP address from lxd managed dnsmasq
#    DHCP server, try to reboot the host. it seems lxd/docker iptables rules
#    are racing/conflicting with each other. it generally fubars when the
#    lxd iptables rules are after the docker ones.
#    NB we already workaround this by configuring cloud-init to reboot the
#       system, so the above problem should not occur anymore.
ping -c 3 debian.org
exit # exit the container.
lxc delete debian --force # destroy the container.
exit # exit the VM.

Destroy everything:

make terraform-destroy