On 16 May 2018, at 09:42, Simone Tiraboschi <stirabos@redhat.com> wrote:



On Tue, May 15, 2018 at 6:31 PM,  <ovirt@fateknollogee.com> wrote:
Engine network config error

Following this blog post: https://www.ovirt.org/blog/2018/02/up-and-running-with-ovirt-4-2-and-gluster-storage/

I get an error saying the hosted engine setup is "trying" to use vibr0 (192.168.xxx.x) even though I have the bridge interface set to "eno1"

Regardless of whether the Edit Hosts File is checked or unchecked, it overwrites my engine IP entry from 10.50.235.x to 192.168.xxx.x

The same thing happens whether I set the engine IP to Static or DHCP (I don't have DNS, I'm using static entries in /etc/hosts).

Any ideas it "insists" on using "vibr0" instead of "eno1"?

Hi,
up to this point is absolutely fine: in the new node zero deployment flow, hosted-engine-setup bootstraps a local VM with an engine there to use that engine to configure the rest of the system (storage, network...).
That bootstrap VM runs over default natted libvirt network, that's why you see vibr0 and 192.168.xxx.x at that stage.

If you are facing any issue, it's definitively not there.

Yes, the blog post describes the 4.2.{0,1} behaviour while the local VM is a change introduced in 4.2.2.
I faced a similar error and in my case the VM creation failed because there was not enough space in /var/tmp. It wasn’t a problem before, because the engine VM was created directly on top of a gluster volume.

Andrea
-- 
Andrea Dell'Amico
http://adellam.sevenseas.org/