Ethernet Switches & Routers

Reply
Occasional Contributor
Posts: 8
Registered: ‎07-30-2010

Routing issues and design problems

We are experiencing some routing issues and I think I have identified a problem.

I have attached a visio of the way this network is built to help explain.

There are two sites, each with 2 core switches, rx16’s and rx8’s.

The core switches are connected between sites with 2 trunks each both running rstp over vlan 3002.

The core switches at each site are connected with all vlans tagged except vlan 3002.

As you can see the way the core switches are setup all of them are on the same subnet.  However because vlan 3002 is not tagged between the cores at each site the core switches can only ping the one that is directly connected (when I source the pings, from vlan 3002).

This doesn’t make any sense to me, is there any reason that this is set up this way?  Shouldn’t the core switches be on different subnets (for the between site connection’s).  What am I missing?

Also, why use 2 trunks between the sites, why not one big trunk? This would eliminate rstp and utilize all 4 ports.

Occasional Contributor
Posts: 8
Registered: ‎07-30-2010

Re: Routing issues and design problems

Sorry the image file didn't work, here is another copy.

Thanks,

Davidt

Super Contributor
Posts: 1,087
Registered: ‎12-13-2009

Re: Routing issues and design problems

Yes a 4 port LACP link would be the way to go here.

As for the design, I cannot say why it was done that way,

If you do need layer 2 extendard to both sites then routing is a much cleaner (and less problematic) approch.

Is one production and one side DR?  The design might of been done so that a server on the production side can fail and the servr in DR can take on its idenity e.g. IP address so no client changes are needed.

Occasional Contributor
Posts: 8
Registered: ‎07-30-2010

Re: Routing issues and design problems

Thanks for your help.

We don't need level 2 between the sites, the 3002 vlan isn't on any other ports except the ones in the diagram.  I would much rather just do level 3 routing between the sites (we are using ospf).

So my plan is this, please correct or suggest better ideas

Change the 2 trunks into a LACP 4 port one.

Change the vlan number and subnet on the secondary switches to something like... Vlan 3003, subnet: 10.99.2.8 /29

What do you think?

To answer you about the backup.

The switches at each site are running vrrp-extended on all vlans except 3002, between themselves.

For example, Switch 1 & 2 at site 1 have vrrp-extended between each other, for all vlans except 3002.

But there is no vrrp-extended between sites.

Thanks for your help,

Davidt

Join the Community

Get quick and easy access to valuable resource designed to help you manage your Brocade Network.

Download FREE NVMe eBook