Redundancy on the fabric

[Aaron.S]

We currently have about 6 HP MSAs, 4 P-class enclosures, 2 c-class enclosures, and two Brocade Silkworm 3900s.

I'm trying to solve the issue of configuring it as so the Silkworm 3900s will be able to failover between themselfs. Or just simply route the data between themselfs to provide access to all devices regardless of which switch they are connected to.

[Víctor Cespón]

But... don't you already have every MSA and blade server connected to both switches? This way there are two independent fabrics, and two paths from each server to each storage device.

[Sivakumar MJ._1]

If you dont have your SAN Environment as per vcespon, Proceed in normalizing your SAN Environment.

Brocade Silkworm 3900 - 32 FC Ports (2Gig)

How many Ports are utilized now and how many free ports are avaialble in the Silkworm?

How many Paths you have now for each enclosures to the MSA's.

[Aaron.S]

Our MSAs are a bit old, so they do not have redundant Fiber. So far we each Silkworm is of the 32port variety, and the one we have been using has about 4 ports left on it. (soon to be filled)

Two weeks ago one of the Silkworms decided to arbitrarily set it's default zone to 'no access'. It proceeded to do this again the next day. This caused some major outages, and we were lucky it didn't kill our VMs.

I guess I'm basically looking for how to route between the two fiber switches? So that a server directly attached to Silkworm1 can still easily find the MSA which has been plugged into Silkworm2?

Best case scenario would then be only half of our MSA storage spontaneously disappears instead of it all.

Presently:
Each MSA has only one fibre connection to the switch
Some mass storage (SATA) devices have redundant fibre connections to the switch
Half of the servers have only one fibre connection to the switch
All of the 'c' and 'p' class enclosures have redundant fibre connections to the switch.