Hosted engine deployment w/ two networks (one migration, one management).

Hello all, I'm slowly building a new ovirt over glusterfs cluster with 3 fairly beefy servers. Each of the nodes has the following network configuration: 3x1GbE: ILO, ovirtmgmt and SSH. 4x10GbE: Private and external VM network(s). 2x40GBE: GlusterFS and VM migration. Now, for some odd reasons, I rather keep the two 40GbE networks disconnected from my normal management network. My question is simple: I remember that I can somehow configure ovirt to use two different networks for for management / migration, but as far as I can see, I cannot configure the cluster to use a different network for migration purposes. 1. Am I missing something? 2. Can I somehow configure the hosted engine to have an IP in more than network (management and migration)? 3. More of a gluster question: As the 40GbE NICs and 1GbE NIC sitting on different switches, can I somehow configure gluster to fallback to the 1GbE NIC if the main 40GbE link fails? AFAIR bond doesn't support asymmetrical network device configuration. (And rightly so, in this case). Thanks, Gilboa

Hey! You can use one network for your storage management i.e. Gluster. If you are using the cockpit-ovirt UI you can simply add the two different FQDN for the different networks during the gluster deployment itself. Something similar to the screenshot below. As I have only one network, I mentioned the same network for both the inputs. But as you have two different networks you can mention the storage network in the first field and the management network in the second. Thanks Parth Dhanjal On Sun, Dec 13, 2020 at 11:59 PM Gilboa Davara <gilboad@gmail.com> wrote:
Hello all,
I'm slowly building a new ovirt over glusterfs cluster with 3 fairly beefy servers. Each of the nodes has the following network configuration: 3x1GbE: ILO, ovirtmgmt and SSH. 4x10GbE: Private and external VM network(s). 2x40GBE: GlusterFS and VM migration.
Now, for some odd reasons, I rather keep the two 40GbE networks disconnected from my normal management network. My question is simple: I remember that I can somehow configure ovirt to use two different networks for for management / migration, but as far as I can see, I cannot configure the cluster to use a different network for migration purposes.
1. Am I missing something? 2. Can I somehow configure the hosted engine to have an IP in more than network (management and migration)? 3. More of a gluster question: As the 40GbE NICs and 1GbE NIC sitting on different switches, can I somehow configure gluster to fallback to the 1GbE NIC if the main 40GbE link fails? AFAIR bond doesn't support asymmetrical network device configuration. (And rightly so, in this case).
Thanks, Gilboa _______________________________________________ Users mailing list -- users@ovirt.org To unsubscribe send an email to users-leave@ovirt.org Privacy Statement: https://www.ovirt.org/privacy-policy.html oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/ List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/QBCMH2J7HIKCQQ...

On Mon, Dec 14, 2020 at 7:00 AM Parth Dhanjal <dparth@redhat.com> wrote:
Hey! You can use one network for your storage management i.e. Gluster. If you are using the cockpit-ovirt UI you can simply add the two different FQDN for the different networks during the gluster deployment itself. Something similar to the screenshot below. As I have only one network, I mentioned the same network for both the inputs. But as you have two different networks you can mention the storage network in the first field and the management network in the second.
Thanks Parth Dhanjal
Parth, Thanks for taking the time to answer my question(s). I am aware that I can use different NICs for oVirt management and Gluster, we use it across the board in several oVirt 4.3 and 4.4 clusters. In this case, I'm looking to add an additional layer of protection by separating the inter-host / engine communication from external interference. Essentially I want to create the following configuration: 1. 10.0.1.0/24: oVirt hosted engine connection to the IT network (oVirt web management, SSH access to engine and hosts). 2. 10.0.2.0/24: oVirt hosted engine and hosts private network. (Used for vdsm inter host communication and VM migration). 3. 10.0.3.0/24: GlusterFS. Now granted, I can achieve the same by adding a NAT bridge and place the hosted engine web management on 10.0.2.0/24, but it's less than ideal. - Gilboa
On Sun, Dec 13, 2020 at 11:59 PM Gilboa Davara <gilboad@gmail.com> wrote:
Hello all,
I'm slowly building a new ovirt over glusterfs cluster with 3 fairly beefy servers. Each of the nodes has the following network configuration: 3x1GbE: ILO, ovirtmgmt and SSH. 4x10GbE: Private and external VM network(s). 2x40GBE: GlusterFS and VM migration.
Now, for some odd reasons, I rather keep the two 40GbE networks disconnected from my normal management network. My question is simple: I remember that I can somehow configure ovirt to use two different networks for for management / migration, but as far as I can see, I cannot configure the cluster to use a different network for migration purposes.
1. Am I missing something? 2. Can I somehow configure the hosted engine to have an IP in more than network (management and migration)? 3. More of a gluster question: As the 40GbE NICs and 1GbE NIC sitting on different switches, can I somehow configure gluster to fallback to the 1GbE NIC if the main 40GbE link fails? AFAIR bond doesn't support asymmetrical network device configuration. (And rightly so, in this case).
Thanks, Gilboa _______________________________________________ Users mailing list -- users@ovirt.org To unsubscribe send an email to users-leave@ovirt.org Privacy Statement: https://www.ovirt.org/privacy-policy.html oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/ List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/QBCMH2J7HIKCQQ...
participants (2)
-
Gilboa Davara
-
Parth Dhanjal