Not sure if I understand what you are asking here, but the need for a
gateway per network has emerged from the need to support other host
networks (not VM networks) beside the management one.
As an example, migration and storage networks can be defined, each passing
dedicated traffic (one for storage communication and another for VM
migration traffic), they may need to pass through different gateways.
So the management network can be accessed using gateway A, storage using B
and migration using C. A will usually be set on a host level as the host
default gateway, and the others will be set for the individual networks.
Otherwise, how would you expect storage to use a different router (than the
management one) in the network?
Thanks,
Edy.
On Thu, May 3, 2018 at 1:08 AM, Justin Zygmont <jzygmont(a)proofpoint.com>
wrote:
I don’t understand why you would want this unless the ovirtnode
itself was
actually the router, wouldn’t you want to only have an IP on the management
network, and leave the rest of the VLANS blank so they depend on the router
to route the traffic:
NIC1 -> ovirt-mgmt - gateway set
NIC2 -> VLAN3, VLAN4, etc…
https://www.ovirt.org/documentation/admin-guide/chap-Logical_Networks/
*Viewing or Editing the Gateway for a Logical Network*
Users can define the gateway, along with the IP address and subnet mask,
for a logical network. This is necessary when multiple networks exist on a
host and traffic should be routed through the specified network, rather
than the default gateway.
If multiple networks exist on a host and the gateways are not defined,
return traffic will be routed through the default gateway, which may not
reach the intended destination. This would result in users being unable to
ping the host.
oVirt handles multiple gateways automatically whenever an interface goes
up or down.
_______________________________________________
Users mailing list
Users(a)ovirt.org
http://lists.ovirt.org/mailman/listinfo/users