I have a small site that contains only dual homed servers. Each server in this site is patched active/passive to two Catalyst 2960-48TT switches. There are 2 vlans in this site. There is currently one Catalyst 3560 which serves as a core switch doing the intervlan routing and also as a WAN router with 2 x 100Mb metro links from different providers to another site. I'm using EIGRP to look after my routing. The Gi interfaces are configured as trunks. This is diagram1.
I have A+B power to this site, but obviously this does not help the C3560. I had though of a static transfer switch or using an RPS, but if I introduce another C3560, I will get full redundancy to and in this site.
My question is, what is the best way to configure the extra C3560 switch?
I was thinking of setting it up as per diagram2, using the Gi ports as trunks and letting STP sort out the loops. I would then configure HSRP, or MHSRP on the LAN side of SW01 and SW02. My concern with this approach is that the core switches are not directly connected, would this be a potential issue?
I'd be grateful for any input, and I apologise for the poor quality of the diagrams!
...and thank you for all the replies to other posts on these forums, they really are a great resource.
a direct link between the two C3560 is strongly recommended you can build also with a port-channel of RJ-45 FE or GE.
I would deploy it.
Then using HSRP or GLBP you can serve the clients vlans users connected to the two 2960 switches.
The setup can work also as depicted in diagram 2.
Hope to help