On Wed, Jan 16, 2019 at 8:24 PM Gianluca Cecchi
<gianluca.cecchi@gmail.com> wrote:
>
> Just tested what in subject on a vm (so it has been a nested env) configured with 2 disks: 1x60Gb for OS and 1x100Gb for gluster.
> Used node ng from ovirt-node-ng-installer-4.3.0-2019011608.el7.iso and almost all the problems had with RC1 here:
> https://www.mail-archive.com/users@ovirt.org/msg52870.html
> have gone solved.
> The hosted engine deploy completed successfully and I'm able to access engine web admin and see the VM and its storage domain and connect to console.
>
> The only persisting problem is that I choose JBOD in my disk config (because I have only one disk) and the deploy gives error about missing diskcount parameter in ansible playbook, so I have to edit the gdeploy conf file from the gui and add the section
>
> [diskcount]
> 1
>
> after that all goes well.
> Good job guys!
>
> BTW: is it correct hat for data and vmstore I only have the volumes configured, while I have to manullay create storage domains?
I think the addition of storage domain may be missing? Gobinda?
>
> If I want to extend this initial configuration to 3 nodes, what would be the installation path to execute?
You can add the additional hosts via engine, and expand the volume.
The bricks are laid out differently - as there's no option to create
multiple bricks on engine UI. We did have an expand volume planned in
the gluster management dashboard in Cockpit, I don't recollect if it
made it to 4.3.
My only need at this time is to install the same iso on 2 additional hosts configured with the same hw: 1x60Gb disk for os and 1x100Gb for Gluster, and add them to the infra so that I will have the same volumes with the same size, but distributed on the three nodes, without interruption of service from current config.
One of the two additional nodes will work as arbiter
Thanks,
Gianluca