You can reboot the system, just in case the kernel was updated.
You can clean the existing system from the UI itself.
Upon selecting the option to deploy a Hyperconverged Infrastructure, it'll give you an option to clean the existing setup.
Also you need to ensure that ssh-key is available in the hosts file, so that ansible can execute roles on the server.
You can refer to this doc as well - https://ovirt.org/documentation/gluster-hyperconverged/chap-Deploying_Hyperconverged.html

On Fri, Oct 30, 2020 at 4:20 PM Gianluca Cecchi <gianluca.cecchi@gmail.com> wrote:

On Fri, Oct 30, 2020 at 11:43 AM Parth Dhanjal <dparth@redhat.com> wrote:
Hey!
It seems vdsm packages are missing.
Can you try installing vdsm-gluster and ovirt-engine-appliance packages?
In case you face repo issues, just install yum install https://resources.ovirt.org/pub/yum-repo/ovirt-release-master.rpm
Then try again.

Ok, thanks.
The appliance was already installed but vdsm-gluster was indeed missing.

So currently before deployment I executed on CentOS host (initially installed choosing "server" option):
yum update
yum install https://resources.ovirt.org/pub/yum-repo/ovirt-release44.rpm
yum update
yum install cockpit-ovirt-dashboard gluster-ansible-roles glusterfs-server vdsm-gluster
systemctl enable --now cockpit.socket
reboot

Is there a list of all what is needed for this test lab single host HCI with gluster? Anything else?

And what is the safe way to clean up before retrying deploy?

Thanks,
Gianluca