ospf on 3750 core switches

Answered Question
Dec 5th, 2010

Hi there,

We have two 3750 switches setup on local lan. Switch A is primary switch and has ospf setup on it and is a neighbour to a local MPLS Switch. Switch B has etherchannel to Switch A and default route out is switch A. hsrp on all interfaces has switch A as the primary.

Whats the best method to deploy ospf onto switch B? I've tried adding similar commands as switch A but it screws the routing and everything drops across the ospf link for switch A.

Thanks,

T.

I have this problem too.
0 votes
Correct Answer by Bharat Negi about 3 years 4 months ago

Hi Carl

To acheive full redundancy you must connect

     Switch A to MPLS switch with OSPF >>>>With this Switch A will have routing table from MPLS switch.

     Switch B to MPLS switch with OSPF >>>>With this Swtich B will have routing table from MPLS switch.

     Access switches uplinked to both switches (A&B) >>>With this Access switches have link level redundancy.

     Run HSRP between Switch A & B for all VLANs >>>With this device level redundancy will be acheived.

     Switch A connected to Switch B over a Trunk >>>With this traffic can move between switches.

All possible failure scenarios can be mitigated with this topology except one.  Explained below:

Consider Access switch link to Switch A fails. Now Swith B will become HSRP active.  At the same time Switch B to MPLS switch link fails.  Now Switch B don't have routing table to forward traffic.

Hence to avoid this situation you must enable OSPF between Switch A & B, so that they can also share their routing table.

    

Regards

Bharat

  • 1
  • 2
  • 3
  • 4
  • 5
Average Rating: 5 (1 ratings)
Reza Sharifi Sun, 12/05/2010 - 14:07

Carl,

Do you have the right image with Layer-3 capabilities installed in switch-b?

Why do you need to run OSPF on switch-b?

Assuming your users are connected to switch-b and switch-b connects to switch-a

If you turn on routing on switch-b, then you need to move all your route points, hsrp, etc.. to switch-b

Is that what you want to do?

HTH

Reza

carl-voicenet Sun, 12/05/2010 - 14:14

Hi there,

Yes the correct images are on both.

The reason for putting ospf on switch B is that if switch A dies then switch B will takeover as the primary switch and need the routes.

So are you saying that putting ospf on switch B when switchA is the primary for all vlans and the ospf ne then this will be an issue?

How is this normally done on core switches to enable redundancy?

Reza Sharifi Sun, 12/05/2010 - 14:25

Can you share a diagram of your network?

If switch a and b both are your core switches and you have other layet-2 switches that serve the users and connect to theses two, then you should be able to enable OSPF on switch-b and peer it with switch-a and run vrrp or hsrp for redundancy.

carl-voicenet Sun, 12/05/2010 - 14:39

Only have web access at the moment..

Layout:

3750A >> etherchannel >> 3750B

Each core switch has 4 x 3560 closet switches configured with trunk ports. Default gateway on all vlans is switchA via HSRP.

Reza Sharifi Sun, 12/05/2010 - 15:04

Are these 3560s dully uplinked (1 connnection to switch a and one connection to switch b)?

Reza Sharifi Sun, 12/05/2010 - 16:51

Then, you can't have a layer-3 between your 2 core switches, because if hsrp on switch-a fails, there in no way for switch-b to know about it and take over and become the hsrp master

HTH

Reza

carl-voicenet Sun, 12/05/2010 - 20:37

I dont follow that sorry..

If switch A dies then switchB takes over using HSRP..

What has layer3 got to do with it between switches?

Should i not use an ospf cost on the both switches and use a higher cost for switchB?

Cadet Alain Mon, 12/06/2010 - 01:09

Hi Carl,

Can you put a diagram of your topology indicating where you have HSRP and where you have OSPF.

The standby will take active role  indeed if he can communicate with him and so doesn't see hellos for the holdtime duration

but he must also be configured for preemption.

Your machines must be able to communicate at L2 with the new active router so ther must be a link between your access switches and your 2

distribution switches doing HSRP and there must be a link between your access switches.

I really don't understand the role of OSPF in your topology.

Regards.

carl-voicenet Tue, 12/07/2010 - 19:02

Access switches have uplink trunks to the 3750 core switches.

Core Switch A is the primary HSRP active device for all vlans.

Core Switch A has ospf currently running ok on it. Its neighbour is the MPLS Wan switch.

Core Switch B had HSRP on all interfaces and is the backup.

Core Switch B has some 3560 access switches attached as trunk ports.

I want to enable ospf on Core Switch B to take over the lan routing should core switch A fail.

Thinking of patching a second port from the MPLS switch into Core Switch B and then letting ospf choose the path to either switch.

Is it worth me adding a cost to the vlan for ospf routing. Making switch B the second route?

Attachment: 
Scott Cannon Tue, 12/07/2010 - 19:36

Hi Carl,

Assume CoreA dies. When CoreB doesnt receive hellos from CoreA (and the holddown expires) it will move itself into the active state. If CoreA has failed, and you do not have an uplink to CoreB traffic from your access layer wont reach it.

If you cant add an uplink from your access layer to CoreB then you dont have any redundancy should that link or its connected devices fail.

Does that make sense or have I misconstrued your query?

Rgds

Scott

Correct Answer
Bharat Negi Tue, 12/07/2010 - 22:09

Hi Carl

To acheive full redundancy you must connect

     Switch A to MPLS switch with OSPF >>>>With this Switch A will have routing table from MPLS switch.

     Switch B to MPLS switch with OSPF >>>>With this Swtich B will have routing table from MPLS switch.

     Access switches uplinked to both switches (A&B) >>>With this Access switches have link level redundancy.

     Run HSRP between Switch A & B for all VLANs >>>With this device level redundancy will be acheived.

     Switch A connected to Switch B over a Trunk >>>With this traffic can move between switches.

All possible failure scenarios can be mitigated with this topology except one.  Explained below:

Consider Access switch link to Switch A fails. Now Swith B will become HSRP active.  At the same time Switch B to MPLS switch link fails.  Now Switch B don't have routing table to forward traffic.

Hence to avoid this situation you must enable OSPF between Switch A & B, so that they can also share their routing table.

    

Regards

Bharat

Actions

Login or Register to take actions

This Discussion

Posted December 5, 2010 at 1:18 PM
Stats:
Replies:12 Avg. Rating:5
Views:1397 Votes:0
Shares:0
Tags: No tags.

Discussions Leaderboard

Rank Username Points
1 15,007
2 8,155
3 7,745
4 7,088
5 6,742
Rank Username Points
140
80
78
64
40