Hi All,

Is there any update on this ??


On Thu, Jul 31, 2014 at 9:08 AM, Punit Dambiwal <hypunit@gmail.com> wrote:
Hi Antoni,

But all the bricks and gluster in the another subnet...I am using the Compute and gluster storage combined on the same host....

Ovirtmgmt :- 43.25.76.0/24
Gluster Storage :-  10.10.10.0/24

Thanks,
Punit 


On Wed, Jul 30, 2014 at 6:17 PM, Antoni Segura Puimedon <asegurap@redhat.com> wrote:


----- Original Message -----
> From: "Punit Dambiwal" <hypunit@gmail.com>
> To: "Antoni Segura Puimedon" <asegurap@redhat.com>
> Cc: "Sahina Bose" <sabose@redhat.com>, "Dan Kenigsberg" <danken@redhat.com>, barumuga@redhat.com, users@ovirt.org,
> "Itamar Heim" <iheim@redhat.com>, "Michal Skrivanek" <michal.skrivanek@redhat.com>, "Sven Kieske"
> <S.Kieske@mittwald.de>, "Frantisek Kobzik" <fkobzik@redhat.com>, ahadas@redhat.com
> Sent: Wednesday, July 30, 2014 12:11:30 PM
> Subject: Re: Network Question
>
> Hi Antoni,
>
> Please find the attached..for the network settings..

Shouldn't the bricks and all gluster be on the same subnet as storage?
Otherwise it is going to use the link scoped route of ovirtmgmt that is
on the same subnet as them.

>
>
> On Wed, Jul 30, 2014 at 5:42 PM, Antoni Segura Puimedon <asegurap@redhat.com
> > wrote:
>
> >
> >
> > ----- Original Message -----
> > > From: "Punit Dambiwal" <hypunit@gmail.com>
> > > To: "Antoni Segura Puimedon" <asegurap@redhat.com>
> > > Cc: "Sahina Bose" <sabose@redhat.com>, "Dan Kenigsberg" <
> > danken@redhat.com>, barumuga@redhat.com, users@ovirt.org,
> > > "Itamar Heim" <iheim@redhat.com>, "Michal Skrivanek" <
> > michal.skrivanek@redhat.com>, "Sven Kieske"
> > > <S.Kieske@mittwald.de>, "Frantisek Kobzik" <fkobzik@redhat.com>,
> > ahadas@redhat.com
> > > Sent: Wednesday, July 30, 2014 11:35:28 AM
> > > Subject: Re: Network Question
> > >
> > > Hi Antoni,
> > >
> > > Please find the attached screen shots...
> >
> > From what I see in the screen shots,
> >
> > The ip rules and routes say that communication to the
> > 43.25.76.0/24 should go via bond0.10 which is for the ovirtmgmt device.
> > That's why it happens.
> >
> > How did you setup the two networks? Are they on the same subnet?
> >
> > >
> > > Thanks,
> > > Punit
> > >
> > >
> > > On Wed, Jul 30, 2014 at 5:19 PM, Antoni Segura Puimedon <
> > asegurap@redhat.com
> > > > wrote:
> > >
> > > >
> > > >
> > > > ----- Original Message -----
> > > > > From: "Punit Dambiwal" <hypunit@gmail.com>
> > > > > To: "Sahina Bose" <sabose@redhat.com>
> > > > > Cc: "Dan Kenigsberg" <danken@redhat.com>, barumuga@redhat.com,
> > > > users@ovirt.org, "Itamar Heim" <iheim@redhat.com>,
> > > > > "Michal Skrivanek" <michal.skrivanek@redhat.com>, "Antoni Segura
> > > > Puimedon" <asegurap@redhat.com>, "Sven Kieske"
> > > > > <S.Kieske@mittwald.de>, "Frantisek Kobzik" <fkobzik@redhat.com>,
> > > > ahadas@redhat.com
> > > > > Sent: Wednesday, July 30, 2014 10:46:47 AM
> > > > > Subject: Re: Network Question
> > > > >
> > > > > Hi All,
> > > > >
> > > > > I am using the storage ip to mount the glusterfs to host...that
> > means the
> > > > > data communication should be run on this network but it's not....it's
> > > > > running on the ovirtmgmt network...
> > > >
> > > > Could you please attach the ip addresses of the gluster nodes, and the
> > > > result of doing:
> > > >
> > > > (on the vdsm host)
> > > >
> > > > ip rule show
> > > > ip route show table (add here each of the tables that show up in the
> > > > previous
> > > > command)
> > > >
> > > > >
> > > > > [image: Inline image 1]
> > > > >
> > > > >
> > > > >
> > > > > On Wed, Jul 30, 2014 at 4:26 PM, Sahina Bose <sabose@redhat.com>
> > wrote:
> > > > >
> > > > > >
> > > > > > On 07/30/2014 01:21 PM, Dan Kenigsberg wrote:
> > > > > >
> > > > > >> On Wed, Jul 30, 2014 at 01:04:48PM +0800, Punit Dambiwal wrote:
> > > > > >>
> > > > > >>> Hi,
> > > > > >>>
> > > > > >>> I am facing one strange problem with Ovirt network...i have the
> > below
> > > > > >>> network in my ovirt cluster :-
> > > > > >>>
> > > > > >>> eth0+eth1=bond0
> > > > > >>>
> > > > > >>> bond0.10= ovirtmgmt (43.250.X.X)
> > > > > >>> bond0.37= VMnetwork
> > > > > >>>
> > > > > >>> eth2+eth3=bond1 (Storage) (10.10.10.X)
> > > > > >>>
> > > > > >>> Glusterfs running on 4 HV's with each have 2 bricks...
> > > > > >>>
> > > > > >>> But still the glusterfs data transfer from ovirtmgmt network
> > instead
> > > > of
> > > > > >>> storage network ?? Is there any more setting required to change
> > it to
> > > > > >>> ovirtmgmt to storage ??
> > > > > >>>
> > > > > >> The title "Storage" means nothing to oVirt per se. The user needs
> > to
> > > > make
> > > > > >> sure
> > > > > >> that her NFS or iSCSI storage server uses the specific network.
> > > > > >> But unfortunately - and Bala is welcome to correct me if I'm
> > wrong -
> > > > there
> > > > > >> is no means to set the IP address which gluster uses via oVirt.
> > > > > >>
> > > > > >> Can it be done "underneath" oVirt?
> > > > > >>
> > > > > >
> > > > > > For glusterd communication - that is gluster peer to gluster peer -
> > > > > > there's a workaround to achieve this..Please see
> > > > http://lists.ovirt.org/
> > > > > > pipermail/users/2014-February/021213.html
> > > > > >
> > > > > > For VM to gluster host, the host IP used to mount the volume is
> > the one
> > > > > > used for communication by oVirt.
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>