MX redundancy

[Unni]

I have a specific  customer requirement as below with 2 MX100 platforms and 2 ISP connections. Just wondering is this possible with Meraki.
The requirements are : In normal working conditions each MX should handle specific traffic (like MX1 for wifi and MX2 for video) and if there is a failure in any ISP or any MX, then all traffic should move to the active one and should come backup to original state the connection is restored.
If this is possible how can we do it.

Thanks

Unni

[MRCUR]

This is not possible with the MX as the redundancy (warm spare) is an active-passive design. All traffic is handled by the active MX unless there is a failover event and then all traffic is handled by the second MX until the primary MX comes back online. 

[AjitKumar]

Hi Unni,

This is my understanding.

MXes in HA works in Active-Passive mode. We need a single License for this functionality.
This means at a given time only 1 MX shall be in service. All the traffic needs to flow via the active MX.

The traffic flow may utilize both the Links (ISPs) using Load balancing and Internet Flow feature.
If you have SD-WAN you have further options to segregate traffic.

Failure of MX / ISPs shall swap the roles (Primary / Spare) of devices.

[PhilipDAth]

Ideally for the primary circuit get a fibre circuit with a /29 of address space.  Plug both MX's into this.  As a compromise you could get a circuit that comes with a single public IP address and an ISP router and plug both MX's into that.  This is more painful if you want to do inbound NAT - otherwise it is a price sensitive compromise.

You have the same options for the secondary circuit.  You also have the option of only plugging it into the primary MX.  If the primary MX fails the secondary MX will take over, and all traffic will only go out the primary WAN circuit.

Use a VLAN for WiFi, and another VLAN for everything else.  Then use flow preferences to route one VLAN out the primary circuit, and the other VLAN out the secondary circuit.  Everything will be able to failover if something dies.

https://documentation.meraki.com/MX/Firewall_and_Traffic_Shaping/MX_Load_Balancing_and_Flow_Preferen...