[ovirt-users] Reachability problems after putting ovirtmgmt and Gluster on different networks

[Stefano Stagnaro]

Hi, I have a question about an HostedEngine deployment. I followed the Jason's "oVirt 3.4, Glusterized" and it seems working. I've two host for virtualization (that also do the HA for the HostedEngine) and two other host for GlusterFS.

Now I would like to configure separate networks for the management and Gluster. On my Catalyst I've configured two VLANs with two different default gateways. My aim is to add the Gluster hosts with the IP of the "ovirtmgmt" and then add the storage domain pointing to the same host but using the IP of the "Gluster network".

Unfortunately, the HostedEngine has only one nic (on ovirtmgmt) and so I'm not able to reach the Gluster host with his "Gluster network" IP even if the L3-switch has all the networks configured. Reading this article http://brainscraps.wikia.com/wiki/Setup_Gateway_Routing_On_Multiple_Network_Interfaces I believe the problem is on the Gluster host that only has the ovirtmgmt gateway.

I've found out that someone has tried to add a second nic on the HostedEngine by editing the /usr/share/ovirt-hosted-engine-setup/templates/vm.conf.in "template" prior to start the deployment.

My question is: there is a better way to achieve this (I'm missing something for sure)? Is it worth having the ovirtmgmt separate to the storage? In production, I'll have five 1gbps ports per host (that could be bonded as preferred).

Thank you,
-- 
Stefano Stagnaro

Prisma Engineering S.r.l.
Via Petrocchi, 4
20127 Milano – Italy

Tel. 02 26113507 int 339
e-mail: stefanos at prisma-eng.com
skype: stefano.stagnaro