OSPF configure for ERS-5520

[KT]

all -

I never ever done or configured of OSPF protocol on ERS-5520 also I searched all of Avaya documents plus Avaya IP Routing Technical Guide and I followed based on the examples.

I have a stack with 2 Units that connect to OSPF provider and I assumed thier Cisco Routers.

also I have 2 VLANs which is TLAN and ELAN on the LAN side

also we connected to thier Cisco Router like:

Unit 1 port 1/23-24 to Cisco R1 and R2
Unit 2 port 2/23-24 to Cisco R1 and R2

IP address of port 1/23 is 172.31.15.2/30
IP address of port 1/24 is 172.31.15.9/30
IP address of port 2/23 is 172.31.15.6/30
IP address of port 2/24 is 172.31.15.13/30

Configure OSPF for stack 1

5520(config)#interface fast 1/23
5520(config-if)#brouter port 1/23 vlan 100 subnet 172.31.15.2/30 (why we need to create a VLAN for each interface that connect to cisco)
5520(config-if)#interface fast 1/24
5520(config-if)#brouter port 1/24 vlan 200 subnet 172.31.15.6/30
5520(config-if)#interface fast 2/23
5520(config-if)#brouter port 2/23 vlan 300 subnet 172.31.8.9/30
5520(config-if)#interface fast 2/24
5520(config-if)#brouter port 2/24 vlan 400 subnet 172.31.15.13/30
5520(config-if)#exit

5520(config)#ip routing
5520(config)#router ospf
5520(config-router)#router-id x.x.x.x (where do i get router ID? is it from OSPD provider?)
5520(config-router)#area x.x.x.x import noexternal
5520(config-router)#network 172.31.15.2 area x.x.x.x (where do i get area code?)
5520(config-router)#network 172.31.15.6 area x.x.x.x
5520(config-router)#network 172.131.15.9 area x.x.x.x
5520(config-router)#network 172.31.15.13 area x.x.x.x
5520(config-router)#network 192.168.0.2 area x.x.x.x
5520(config-router)#network 192.168.0.130 area x.x.x.x
5520(config-router)#exit
5520(config)#router ospf enable

Please feel free to comment these.

Thanks, KT

[Michael McNamara]

I'm not 100% sure if brouter ports are fully supported on the ERS 5500 series, you'll also need the Advanced License to enable OSPF on the ERS 5000 series.

Instead you might want to try just making the switch ports trunks (TagAll), then create the Layer 3 interface on the VLAN (config t; interface vlan 100; ip address x.x.x.x x.x.x.x; ospf enable) then add the VLANs to the ports you set as TagAll.

The OSPF router ID is a field used to identify this particular router, you should usually use either a loopback address or an IP address from on of the interfaces.

The OSPF area will be 0.0.0.0 for a simple OSPF network with a single backbone.

Good Luck!

[KT]

Mike -

1. we purchased a new Advance License for L3 Routing.
2. based on the Avaya IP Routing Technical Guide was configured by brouter with VLANs. you sais if it did not supported and doing this way.

5520#config t
5520(config)#int vlan 100
5520(config-if)#ip address 172.31.15.2 255.255.255.252
5520(config-if)#exit
5520(config)#ip ospf enable
5520(config)#vlan members add 100 1/23

I need to do the rest of VLANs 200, 300 and 400

The rest of my configure is the same right? my configure is appplied for stub ABR

here is my network diagram

Thanks a lot Mike

KT

[Uncle John]

OK, some best practices here.

1)  DO NOT USE BROUTER PORTS any more!  We used brouter ports because they did not participate in spanning tree.  Today we want to use a vlan and turn off spanning tree on those ports.  That provides the same capabilities that brouter ports provided, but now I can MLT so I get scaling.

2)  The OSPF ID MUST BE UNIQUE in your OSPF domain.  Best practice is to make it your management (CLIP) address.  That way when you look at the OSPF table you know immediately how to get to the associated router.

3)  For vlans that are not transit (ie. there are just PC's hanging off them, no other routers), then you want to enable those as "Passive" interfaces.  This reduces the CPU demand on the switch.

4)  You may also need to redistribute any other routing protocol (rip/bgp) into OSPF in order for it to learn those routes.

5)  Enable "ASBR" (Autonomous System Boundary Router) only on routers that connect to another AS.

6)  Keep the number of areas down to avoid Dikjstra's algorithm slowing down the CPU.

Cheers, Uncle John.

[Flintstone]

Hi Uncle John and welcome to the forum,
Thanks for sharing your knowledge etc 

CheerZ

[KT]

Uncle John -

Based on your remark of number 1. we never used brouter instead to config each VLANs or interfaces that connected to stub ABR.

5520#config t
5520(config)#int vlan 100
5520(config-if)#ip address 172.31.15.2 255.255.255.252
5520(config-if)#exit
5520(config)#ip ospf enable
5520(config)#vlan members add 100 1/23

like my case have totals 4 links connection.

Thanks, KT

[KT]

all -

i just re-arragned of my config:

Note: VLAN 10 and 20 for ELAN and TLAN for CS1K, IP phones , servers and so on....

R1(config)#int vlan 10
R1(config-if)#ip address 192.168.0.2 255.255.255.128
R1(config-if)#ip ospf network passive
R1(config-if)#int vlan 20
R1(config-if)#ip address 192.168.0.130 255.255.255.128
R1(config-if)#ip ospf network passive
R1(config-if)#exit

Configure OSPF interfaces to Cisco Router (OSPF cloud). W/O using brouter port per Uncle John
R1(config)#vlan create 100 type port
R1(config)#interface vlan 100
R1(config-if)#ip routing
R1(config-if)#ip address 172.31.15.2 255.255.255.252
R1(config-if)#ip ospf area 0.0.0.0 (don't know yet)
R1(config-if)#ip ospf enable
R1(config-if)#exit
R1(config)#vlan members add 100 1/23

R1(config)#vlan create 200 type port
R1(config)#interface vlan 200
R1(config)#ip routing
R1(config-if)#ip address 172.31.15.6 255.255.255.252
R1(config-if)#ip ospf area 0.0.0.0 (don't know yet)
R1(config-if)#ip ospf enable
R1(config-if)#exit
R1(config)#vlan members add 200 1/24

R1(config)#vlan create 300 port type
R1(config)#interface vlan 300
R1(config)#ip routing
R1(config-if)#ip address 172.31.15.9 255.255.255.252
R1(config-if)#ip ospf area 0.0.0.0 (don't know yet)
R1(config-if)#ip ospf enable
R1(config-if)#exit

R1(config)#vlan create 400 port type
R1(config)#interface vlan 400
R1(config)#ip routing
R1(config-if)#ip address 172.31.15.13 255.255.255.252
R1(config-if)#ip ospf area 0.0.0.0 (don't know yet)
R1(config-if)#ip ospf enable
R1(config-if)#exit

GLobal

R1(config)#ip routing
R1(config)#router ospf
R1(config-router)#router-id 192.168.0.132 (IP address of management)
R1(config-router)#area 0.0.0.0 (don't know yet)
R1(config-router)#exit
R1(config)#router ospf enable

Please feel free to add or comt.

Thanks, KT

[Michael McNamara]

Hi KT,

You had me ready to sign off until I saw the "SMLT 33" in your diagram... it's not jiving with the fact that you have 4 different IP networks (segments). 

Is your intent with the 4 links redundancy? If not you should simplify your design and remove 2 of them.

If you want 4 independent routes you'll need to use ECMP and set the max number of routes in the routing table to 2 (perhaps 4). Things can get really really interesting when you start to combine bonding/trunking technologies such as MLT/SMLT/LACP/vPC/PC (Layer 2 protocols) with Layer 3 designs.

Unfortunately I don't have time to get to detailed tonight, but hopefully that makes some sense.

Cheers!

[KT]

Mike -

I forgot to removed smlt out. also you brought good question about 4 links redundancy. I might ask the customer to remove 2 of them.

how abt my OSPF configure over all?

Thanks, KT

[Michael McNamara]

The basic OSPF configuration looks fine to me... interoperability with Cisco should work but might prove somewhat challenging.

Good Luck!

[KT]

you scared me Mike....

[KT]

All -

we tested ospf routing last night after 3rd party brought up these links. it was working five after we changed config.


1. LAN side was configured with passive (no broadcast) and the 3rd party want to see broadcast with this sub-net so we enable ospf this vlan.

2. we changed the uplinks ports to untaggall (i have no clue why he said no needed to tagg these ports)

Thanks your help... I love this forume

[Michael McNamara]

That's great to hear KT.

There are two big benefits to using a 802.1q trunk (tagAll). 1) You can bridge/switch multiple VLANs if necessary down the road without an outage or interruption, 2) With a 802.1q trunk you maintain the 802.1p CoS header which includes the Layer 2 QoS for any voice traffic.

Cheers!

[KT]

all -

Here is the 2nd site of ospf routing and I have a question on MLT between 2 building of fiber connection.

here are these sub-nets to advertise/broadcast to thier ospf interfaces

bldg 510:

10.46.126.0/25
10.46.126.128/25
172.31.9.8/30
172.31.9.20/30

bldg 520

10.46.127.0/25
10.46.127.128/25
172.31.9.12/30
172.31.9.16/30

Do i need to have a route of MLT1 b/w 2 building? or i just created layer 2 of MLT? i attached the network diagram also.

I'm appreciating of ur help...

Thanks, KT

[Michael McNamara]

Hi Kevin,

I'm not 100% clear if you want to stretch your VLANs between buildings or if you are just looking to provide connectivity between the VLANs on each switch.

If you just want to connect the VLANs then just build a new VLAN with a /30 network and make that the only VLAN on the MLT. Add IP interfaces for that VLAN, add routes and you should be done.

If you want to stretch your VLANs then you should just add the VLANs to the MLT and you've have the same VLANs on both stacks.

With respect to your diagram and the TLAN/ELAN, I find it usually "better" from a redundancy standpoint to cable one side of the CPU to one switch (say 1/1) and then the other side of the CPU to the other switch in the stack (say 2/1). If you have muliple Signaling Servers or VGMCs you can spread those out too. In case you happen to loose a single switch you won't loose all your connections and your core CPU should fail over to the standby (assuming you have a dual CPU system).

Cheers!