6248 Switch Redundancy and LACP (LAG) NICs

Your talking about trunk groups LAG, LACP, Etherchannel all do the same thing just with some minor differences.  If you plan on running trunk members to switch1, and switch2 then those switches must be in a stack.  I dont know the limits on creating trunk groups between stack members though so that is something you might want to look up.

An alternative though would be to use Adaptive Load Balancing.  All NICs in team are active but can be connected to any switch), or Adapter Fault Tolerance (One NIC in team is active, and one is standby) these work great, and dont require you to use a LAG/LACP team on the switch.

Hello,

I am trying to do something similar. I've got 4 servers with 4 nics each. I'm bounding them into pairs. I'll plug one port of a bonded nic group into one 6248 switch and the other nic into another 6248 switch. I will be using Adapter Fault Tolerance for the failover. Ultimately, I will be doing this with 4 x 6248s that are connected to each other via "Stacking Modules". Is this the best way to go about creating switch redundancy?  What happens if two switches fail and my bonded nics are plugged into both of them? I'm dead in the water right?

Thanks for any help you can give,

TF 

I would also be interested in a setup similar to this. In our case we only have 2 6248's. Is is possible to upgrade firmware on the secondary switch and upon it's completion move the links from the master to the secondary switch and upgrade the firmware on the (previously) master switch. We have an Equallogic array and are also using VMware - so this would be great for firmware upgrades across the board. Thanks for any input!