We are currently upgrading and replacing hardware accross 5 large sites and upgrading configurations accross the board to support features that have never been used here in the past(portfast, bpduguard, snooping, port security, etc). I'm trying to deterimine final configuration needs prior to production implementation.
I'm leaning more toward using flex links for managing our trunk uplinks between our core and our access layer, and have been able to verify in a test environment that they'll work the way I anticipate. Access layer switches will still use portfast and BPDU guard to prevent users from introducing unknown LAN equipment into the environment.
The question I have is how would the core be configured with regards to flex links? I've seen various articles on how to configure flex links, and how to use them to load balance VLANs accross multiple uplinks, but I'm not sure how the core would be configured.
Access Layer: Three 3560 48s
The 3560s will have 2 port Etherchannels between them, with the top switch trunked to the core via a 2 port Etherchannel, and the bottom switch trunked to the core via a 2 port Etherchannel. Etherchannel One on each switch will be the 'active' flexlink with Etherchannel Two the backup. VLANs will be load-balanced accross the links.
That said, with the core, since there will be multiple fiber uplinks into the core from multiple groups of access switches, how are flex links handled there? Should I leave STP running on the core? Can I just configure multiple flex links on the core (one for each group of access switches)? Is that even a recommended configuration? Is there something I'm not considering?