Hi,

Can anyone share their experience on deploying hosted-engine with GlusterFS cluster?

I managed to setup a GlusterFS cluster and started to deploy hosted engine. At first stage I was beaten by firewall rules - the deployment process interrupted at GlusterFS config stage.

After fixing the rules I got engine up and running, but the host is non-operational state still. Logged in to Web UI -> and see 2 action points: "Gluster command failed on server" and Gluster status is disconnected for this server"

This is a bit strange, since the cluster was properly detected during deployment and deployment script was supposed to configure the cluster.

--== STORAGE CONFIGURATION ==--

Please specify the storage you would like to use (glusterfs, iscsi, fc, nfs3, nfs4)[nfs3]: glusterfs

[ INFO ] Please note that Replica 3 support is required for the shared storage.

Please specify the full shared storage connection path to use (example: host:/path): xxxx.xxxxx.xx:/oVirt

[ INFO ] GlusterFS replica 3 Volume detected

Do you want to configure this host and its cluster for gluster? (Yes, No) [No]: Yes

[ INFO ] GlusterFS replica 3 Volume detected

Any ideas how to fix this?

Thanks in advance!

Regards,
Artem

On Fri, Nov 3, 2017 at 2:28 PM, Martin Sivak <msivak@redhat.com> wrote:

Hi,

cockpit is enabled by default when you use ovirt-node. You will
probably have to install the necessary cockpit packages yourself on
pure CentOS - you will need cockpit and ovirt + gdeploy cockpit
plugins (sadly I do not recall the exact package names).

With regards to arbiter and the wizard.. I really do not know, but I
will alert my colleagues who might have more detailed knowledge of the
gluster part.

Denis, Sahina: can you please help me here?

Best regards

Martin Sivak

On Fri, Nov 3, 2017 at 11:29 AM, Artem Tambovskiy

<artem.tambovskiy@gmail.com> wrote:
> Thanks for an article, Martin!
> Any chance to configure a third cost to act as GlusterFS Arbitr only using
> this wizard?
>
> And stupid question - how to make this wizard up and running? I've
> everything installed and nothing is runnin on port 9090 :)
>
> Regards,
> Artem
>
> On Fri, Nov 3, 2017 at 12:49 PM, Martin Sivak <msivak@redhat.com> wrote:
>>
>> Hi,
>>
>> you should take a look at the hyper converged way of installing oVirt.
>> We have a cockpit wizard that does almost everything for you:
>>
>>
>> https://www.ovirt.org/documentation/gluster-hyperconverged/chap-Deploying_Hyperconverged/
>>
>> It uses three hosts and collocates the VMs together with Gluster storage.
>>
>> Best regards
>>
>> --
>> Martin Sivak
>> SLA /oVirt
>>
>> On Fri, Nov 3, 2017 at 8:39 AM, Artem Tambovskiy
>> <artem.tambovskiy@gmail.com> wrote:
>> > Thanks Eduardo!
>> >
>> > I think I can find a third server to build a glusterFS storage. So the
>> > first
>> > step will be to install a self-hosted engine on the new server and start
>> > building a glusterFS storage. IS there any easy way to migrate existing
>> > 5
>> > VM's running on the second bare-metal oVirt host, right? I found a
>> > little
>> > bit tricky moving oVirt backups between the hosts (at least I failed to
>> > replicate the existing VM's on the second server).
>> >
>> > Regards,
>> > Artem
>> >
>> >
>> >
>> > On Fri, Nov 3, 2017 at 10:24 AM, Eduardo Mayoral <emayoral@arsys.es>
>> > wrote:
>> >>
>> >> For HA you will need some kind of storage available to all the compute
>> >> nodes in the cluster. If you have no external storage and few nodes, I
>> >> think
>> >> your best option for storage is gluster , and the minimum number of
>> >> nodes
>> >> you will need for HA is 3 (the third gluster node can be metadata-only,
>> >> but
>> >> you still need that third node to give you quorum, avoid split-brains
>> >> and
>> >> have something that you can call "HA" with a straight face.
>> >>
>> >> Eduardo Mayoral Jimeno (emayoral@arsys.es)
>> >> Administrador de sistemas. Departamento de Plataformas. Arsys internet.
>> >> +34 941 620 145 ext. 5153
>> >>
>> >> On 03/11/17 08:10, Artem Tambovskiy wrote:
>> >>
>> >> Looking for a design advise on oVirt provisioning. I'm running a PoC
>> >> lab
>> >> on single bare-metal host (suddenly it was setup with just Local
>> >> Storage
>> >> domain) and
>> >> no I'd like to rebuild the setup by making a cluster of 2 physical
>> >> servers, no external storage array available. That are the options
>> >> here? is
>> >> there any options to build cheap HA cluster with just 2 servers?
>> >>
>> >> Thanks in advance!
>> >>
>> >> Artem
>> >>
>> >>
>> >> _______________________________________________
>> >> Users mailing list
>> >> Users@ovirt.org
>> >> http://lists.ovirt.org/mailman/listinfo/users
>> >>
>> >>
>> >
>> >
>> > _______________________________________________
>> > Users mailing list
>> > Users@ovirt.org
>> > http://lists.ovirt.org/mailman/listinfo/users
>> >
>
>