Is it possible to have more than 2 Ethernet Interfaces in the HA cluster - I have the one for the Heartbeat and the one for the phones/system to use - is it possible to add another to each of the machines and have them use it in HA mode?
Level(3) wants to hand off SIP trunking directly to my box on a dedicated interface.
If I can’t add another physical, can I add a Virtual to each and have it switch properly?
Bump - Should I open a ticket instead?
I’m not sure I understand why you think a third interface is needed.
There’s certainly plenty of bandwidth available in the network. Have you given any thought to setting up an additional VLAN on your phone/system interface and using that? Many phones (my experience tends to run to Cisco, but others as well) can use VLANs to sequester traffic logically.
Even if you don’t want to use a VLAN, you can assign multiple addresses to one of your existing interfaces and set up actual networks that way.
Without some information on the topology you are looking for, it’s hard for anyone to give you good information. My gut says that two interfaces may already be one too many, so three would be “right out”.
On the other hand, you can set up as many interfaces as you’d care to. Linux can support over 8 interfaces out of the box, and that’s before you start getting creative with source code changes.
No - I am using the commercial FreePBX HA Module - it requires two physical interfaces.
This question is specifically for the Schmooze folks - I thought they would see it here, but apparently not.
I have already submitted a ticket to them - I know full well that Linux can support a LARGE number of physical and virtual interfaces - the twerps from Level(3) are wanting to hand-off a physical VPN feed to do SIP trunking directly to the box - if it was a simple box, no problem - add the interface.
But it is using Schmooze’s HA module - hence the question for them,
i have the same issue. the sip provider has an onsite connection and the current hardware only has two nic’s on it (it is not running under a vm but i may need to do that). the real question is whether there is a way that HA can use this dedicated connection
Bump.
Our specific issue is that the SIP provider wants to use an Internet-routable (not RFC1918) address on the PBX clusters so as to not have to worry about customer address overlap on their MPLS network. If we did this with a singe interface, that means it is likely we’d have to reconfigure a large ton 'o stuff if we switch SIP providers in the future. Our preference is to use RFC1918 addresses for everything internal and have a separate interface for dealing with a primary SIP provider, since they all seem to want to use routable addresses for MPLS.