C-class blade enclosure & fibre failover

[Aaron.S]

I have a c-class enclosure with 3x bl460c units and 2x bl480c units running ESX.

Over the weekend the Brocade AE372A fibre unit in bay 4 failed and flopped over to using the one in bay 3.

Both Brocades are the same unit number, both connected to the same switch.

Unfortunately, this seemed to really mess thing up for the ESX nodes. Absolutely all of them lost any recognition of their datastores until after I reset bay3, so they could all connect through bay 4 once again. Then had to rescan all datastores for each node to resolve the issue.

I would like to troubleshoot this issue, but have no idea where to start as the symptoms dont appear to make sense.

-Aaron

[Uwe Zessin]

It is necessary to have an exact picture of the configuration.

> Both Brocades are the same unit number, both connected to the same switch

What does that mean ? "AE372A" is normally run as a Fibre Channel switch! Has it been set to Access Gateway mode?
What other components (this 3rd switch, what type of storage array) are involved?

Any zoning set up?
Proper Multipathing confirmed?

> Over the weekend the Brocade AE372A fibre unit in bay 4 failed and
> flopped over to using the one in bay 3.

There is no 'flopping' (did you mean 'failover'?) between those switches. The server blades have independent connections to the interconnect bays.