cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Bookmark
|
Subscribe
|
9098
Views
5
Helpful
16
Replies

ip route vrf name 0.0.0.0 0.0.0.0 192.168.1.1 global in Nexus 5548

WILLIAM STEGMAN
Level 4
Level 4

Is there an equivalant command/feature in the Nexus platform to leak a global default into a VRF?  I see the ip route commands under the vrf context name config and can add a default route there, but is that the equivalent of the command above in the subject?  I'm trying to isolate another division's network in a VRF and still allow them Internet access without segementing the other resources into VRFs.

thank you

Bill

2 Accepted Solutions

Accepted Solutions

Lei Tian
Cisco Employee
Cisco Employee

Hi Bill,

NXOS doesnt have a global routing table like IOS. You need to use import/export to leak routes between VRF. For N5K, this is supported 6.0(2). Alternatively, you can use physical cable connects the ports in different VRF, but it will burn 2 ports.

http://www.cisco.com/en/US/docs/switches/datacenter/nexus5500/sw/release/notes/6_02/Nexus5500_Release_Notes_6_02.html#wp465598

HTH,

Lei Tian

View solution in original post

Got it. Here is what you need to do with static routes.

vrf context RED

ip route 0.0.0.0 0.0.0.0 e1/20 192.168.1.1

int e1/19

switch access vlan 255

interface vlan 255

ip add 192.168.1.1/24

int e1/20

no switchport

vrf member RED

ip add 192.168.1.2/24

ip route 192.168.64.0/24 vlan 255 192.168.1.2

ip route 192.168.65.0/24 vlan 255 192.168.1.2

HTH,

Lei Tian

View solution in original post

16 Replies 16

Lei Tian
Cisco Employee
Cisco Employee

Hi Bill,

NXOS doesnt have a global routing table like IOS. You need to use import/export to leak routes between VRF. For N5K, this is supported 6.0(2). Alternatively, you can use physical cable connects the ports in different VRF, but it will burn 2 ports.

http://www.cisco.com/en/US/docs/switches/datacenter/nexus5500/sw/release/notes/6_02/Nexus5500_Release_Notes_6_02.html#wp465598

HTH,

Lei Tian

Hi Lei,

can you give me an idea of how the physical cable would be setup?  I'm not prepared to create the additional vrfs and leak routes between them, but the cable option sounds interesting.  Would it be a matter of connecting one cable to say port eth 1/1 that is in the VRF and eth 1/2 that is in a vlan or l3 mode?  I see I would need to burn two

SFPs per switch too. 

thank you

Hi Bill,

Yes, you got the idea. Just connect one cable to two physical ports, one in VRF and the other port in a vlan or l3 mode in default VRF. You can even form routing protocol between the physical ports, so dynamically leak routes from VRF into default vrf.

HTH,

Lei Tian

Lei, I see I can only place a layer 3 interface in a VRF.  Would I need to assign the L3 interface an IP address in the VRF subnet, or is it good enough to just have the interface belong to the VRF with no IP address? 

thank you

Hi Bill,

Yes, you can only assign L3 interface into VRF. It is layer 3 concept, think each VRF as a seperarte routing table. If the physcial port is access port, you can put the SVI in a VRF.

HTH,

Lei Tian

Hi Lei.  I was planning on putting an SVI in the VRF, but how do I bridge that SVI with the physical port?  Could I simply put it in the same VLAN as the SVI? 

port 1/19 - vlan 255 (Global routing table that has Internet GW in that vlan)   <------->  port 1/20 - vlan 164  (SVI of that vlan in VRF)

thanks again

Hi Bill,

So e1/20 will be

switchport mode access

switch port access vlan 164

interface vlan164

vrf memeber RED

HTH,

Lei Tian

That's what I was imagining, but I now see another vlan, 165, that needs put into the same VRF.  So with that method I'd need to use another pair of interfaces in the same switch wouldn't I? 

thanks

No, you dont need to have another pair of interfaces. Just put the SVI for vlan 165, 164 into the same VRF, and form a routing neighbor between vlan 255 and vlan 164.

HTH,

Lei Tian

Lei, how could I form a neighbor relationship between an SVI in one vlan and an SVI in another VLAN?  They'd need to be on the same network wouldn't they?  And then in order to use the VRF for routing I'd need to filter the routes using a route map on the neighbor between the SVI in vlan 164 and vlan 255 to prevent learning those same routes outside the VRF should they exist there, right?  Sorry this question is dragging on, but I'm still unclear how to implement it. 

thanks

Unfortunately I don't have a lab to test this, but what if I were to create an L3 interface that is on the same subnet as the gateway in the global vlan and assign it to the VRF?

e 1/19 (vlan 255)  < ------------------------>   e 1/20 (VRF L3 192.168.1.2)

gateway 192.168.1.1 (vlan 255)                                                       

Would I then be able to form a routing adjacency with between the global routing table gateway and the VRF?  I would be assigning all 3 networks, 192.168.1.0, 192.168.64.0 and 192.168.65.0 to the same VRF.  But rather than use a routing protocol, could I just enter a default route under the VRF context to point to 192.168.1.1 and then from the global routing table of the Nexus enter an ip route 192.168.64.0/24 vlan 164 command? 

Can you tell me the subnets that are in the VRF, and the subnets that you want leak into default VRF?

HTH,

Lei Tian

I want to put 192.168.64.0/24 and 192.168.65.0/24 in the VRF.  The 192.168.1.0/24 is the vlan used between my core and Internet gateway.  I want to leak only the default 0.0.0.0/0 into that VRF in order to provide Internet access.  I then want to run OSPF in the VRF to communicate the WAN networks and routes inside the VRF along with a few static routes built under the vrf context to take care of all the non-Internet routing.

thanks Lei

Got it. Here is what you need to do with static routes.

vrf context RED

ip route 0.0.0.0 0.0.0.0 e1/20 192.168.1.1

int e1/19

switch access vlan 255

interface vlan 255

ip add 192.168.1.1/24

int e1/20

no switchport

vrf member RED

ip add 192.168.1.2/24

ip route 192.168.64.0/24 vlan 255 192.168.1.2

ip route 192.168.65.0/24 vlan 255 192.168.1.2

HTH,

Lei Tian