San topology

[rogerpeinetapp]

We currently have 2 clustered FAS3050 filers, and one Brocade switch. Now, we want to add one more switch because we run ouf of ports on the first switch. we don't want to use the 2nd switch as the redundancy, just want to use it for additional ports. what is the appropriate way to add it onto this environment?

I am NOT asking steps commands how this switch should be enabled or configured.

Thanks!

[chriskranz]

I would simply link the 2 FC switches. With all Brocade switches you can create an ISL. Take 2 or 4 ports and create an ISL (inter-switch link) between the 2 switches. Then spread all connections between the 2 switches. I'd definitely recommend having a port from each filer into each switch. You will use the ISL quite heavily, so don't be shy with the number of ports you assign to it. I'd say a minimum of 4 really.

[radek_kubka]

Hi and welcome to the Communities!

Can you explain a little bit though why you don't want to create two separate fabrics, thus having redundancy? This way you will have even more ports to play with, because none will be consumed by an ISL. Needless to say it will be way better from the resiliency point of view, too.

Regards,
Radek

[radek_kubka]

Hi,

Bear in mind you need Full Fabric license on Brocade to create an ISL - and their entry level switches come without this feature as a standard.

Cheers,

Radek

[chriskranz]

Having 2 separate fabrics would give you no greater redundancy. All systems would (should) have 2 HBA's, one into each switch of the same fabric. You lose 4 ports yes, but you would gain greater flexibility and still have redundancy. Having seperate fabrics limits the redundancy as you are forcing a switch failover if you simply lose a single port.

[radek_kubka]

Hi Chris,

Yes, OK, OK - I've made an assumption without knowing the actual state of facts: separate fabrics help *if* there are at least two FC ports on each host (I took this bit for granted ). But if this criteria is met, I will stand my ground: creating two fabrics is widely accepted best practice to achieve resiliency.

Cheers,

Radek