Scenario - migrate servers while maintaining their existing IP address from data centre 1 to data centre 2 with minimal downtime. Diagram attached.
I'm planning on using a Cisco ASR1001-X with AES license at DC1 and DC2 and configuring the routers with OTV to extend 10 x VLANs between the data centres. The join interface would connect directly to the WAN circuit NTU and the Internal Interface would connect to the switch and be configured as a service instance with 10 VLANs tagged using dot1q. The problem is that DC1 switch infrastructure is using Cisco Nexus 56xx configured with FabricPath. I can't find any information that suggests that i can patch the Cisco ASR router's internal interface directly into a FabricPath switchport or what the configuration would be.
Older OTV documentation refers to Nexus 7000 and OTV stating the following: "Because OTV encapsulation is done on M-series modules, OTV cannot read FabricPath packets. Because of this restriction, terminating FabricPath and reverting to Classical Ethernet where the OTV VDC resides is necessary."
Is this true for the Cisco ASR also? The only workaround i can think of is to install a cheap catalyst switch connected to the FabricPath domain and re-introduce spanning-tree at the edge but this seems backwards to me. Any help or suggestions appreciated? Thanks
Based on your diagram, it should work because the nexus switch is the demarcation point between the classical ethernet and fabricpath. Fbricpath frames should not hit to the ASR router.
As long as the ASR is configured properly to receive the dot1q tag from nexus and bridge it into the overlay interface, a layer 2 domain will be extended.
So it is possible to have switchports configured as routed, fabricpath and trunk/access in a fabricpath configuration? Do i need to add any spanning-tree pseudo or priority configuration?
|no ip address|
|service instance 1 ethernet|
|encapsulation dot1q 1|
|service instance 2 ethernet|
|encapsulation dot1q 2|
|service instance 3 ethernet|
|encapsulation dot1q 3|
|switchport mode trunk|
|switchport trunk allow vlan 1,2,3|
All the nexus switches in fabripatch domain are considered as a single giant switch and the STP root must belong to the fabricpath domain.
On the ASR routers (OTV AED), you would need to specify the vlans to be extended over the WAN and the site vlan for the communications between local OTV routers.
The configuration on IOS XE has different flavor when compared with NX-OS. It requires to bridge between internal interface and overlay interface.
Refer here for the configuration guide on IOS XE: