[ovirt-users] Self-Hosted engine 4.1.7 unavailable from the outside of the host

[Simone Tiraboschi]

On Mon, Dec 18, 2017 at 11:09 AM, Matteo Capuano <kapu.net at gmail.com> wrote:

>
>
> On Mon, Dec 18, 2017 at 8:55 AM, Simone Tiraboschi <stirabos at redhat.com>
> wrote:
>
>>
>>
>> On Sat, Dec 16, 2017 at 1:29 PM, Matteo Capuano <kapu.net at gmail.com>
>> wrote:
>>
>>> Hi everyone, my name’s Matteo and I’m a new oVirt user.
>>>
>>> I’m trying to install the gluster hyperconverged solution in a lab
>>> environment following the How-To wrote by Jason Brooks:
>>>
>>> https://ovirt.org/blog/2017/04/up-and-running-with-ovirt-4.1
>>> -and-gluster-storage/
>>>
>>> Sadly, I cannot understand how to configure the network on the
>>> self-hosted engine and I’m unable to make it available from outside the
>>> host.
>>>
>>> As written in the HowTo I created three hosts (ovirt-note 4.1.7) each
>>> one with two nics, one for gluster and one for management, the network is a
>>> static LAN with FQDN resolvable (also reverse) by a local DNS. Here the
>>> details:
>>>
>>> Gateway: 172.16.1.1
>>>
>>> DNS: 172.16.1.12
>>>
>>> Host1: 172.16.1.210 (management) – 172.16.2.210 (gluster)
>>>
>>> Host2: 172.16.1.220 (management) – 172.16.2.220 (gluster)
>>>
>>> Host3: 172.16.1.230 (management) – 172.16.2.230 (gluster)
>>>
>>> Engine: 172.16.1.200
>>>
>>> When installing the engine I choose to brigde the management’s nic
>>> (172.16.1.210) of host1 but, once the installation is completed, I'm unable
>>> to reach the engine from the LAN where the hosts are connected. The engine
>>> (172.16.1.200) can ping only host1 (172.16.1.210) and only host1 can ping
>>> the engine.
>>>
>>> As far as my network knowledge goes, to make the engine available from
>>> the outside of host1 I would need to use a third nic or to use some device
>>> to associate the ip and MAC address of the engine’s nic.
>>>
>> Ciao Matteo,
>> no, hosted-engine-setup should create a bridge for you.
>> No need to do custom configuration to expose your VMs.
>>
>> Are you trying on bare metal or on VMs with the engine VM as a nested VM?
>>
>
> Ciao Simone,
>
> thank you for your answer.
>
> My lab is a nested environmet. I installed ovirt on bare metal with the
> engine on another machine. On this setup i have 6 VMs:
>
> - 172.16.1.1  pfSense as firewall/gateway
> - 172.16.1.10  a windows 2016 used as network guest
> - 172.16.1.12  a nethserver installation as DNS server ( also on
> 172.16.2.12 )
> - the three ovirt-nodes as i described on my e-mail
>
> All the machines are pingable, only the engine inside host 1 is
> unreachable and can ping only host1.
>

OK, your issue is caused by vdsm-no-mac-spoofing filter on your L1 VMs.
Please create a custom vNic profile on your external oVirt engine setting
the "network filter" field to "no network filter".
Now you can edit your ovirt-node VMs setting the network profile of the
nics you are going to use for the management bridge to the profile with "no
network filter".
You have to shutdown and restart your ovirt-node VMs and then you could
retry the deployment.



>
>
> Cheers
>
> Matteo
>
>
>>
>>> I’ve looked around over the internet for a solution but every how-to
>>> I’ve found follows the same steps of Jason’s.  I’ve also already asked
>>> for help to Jason.
>>>
>>> Anyone could help me to solve this issue?
>>>
>>>
>>>
>>> Thank you
>>>
>>>
>>>
>>> Matteo
>>>
>>> _______________________________________________
>>> Users mailing list
>>> Users at ovirt.org
>>> http://lists.ovirt.org/mailman/listinfo/users
>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20171218/fd22d869/attachment.html>