Stacking and Switch Virtualization
Stacking (e.g., Aruba VSF/backplane stacking) and switch virtualization (e.g., Cisco VSS, Aruba VSX) combine multiple physical switches into one logical device. Benefits: single management plane, simplified STP (one logical bridge), and multi-chassis link aggregation (MC-LAG) so servers dual-home without STP blocking. Know the difference between a true stack (shared control plane, single config) and an MLAG/VSX pair (independent control planes, synchronized state).
Traps: (1) Stack member IDs and priorities determine master election-misconfigured priority causes unexpected master. (2) Split-brain: if the stack links fail, fragments may both act as master; use a dedicated stack link and MAD (multi-active detection) or an out-of-band witness. (3) Firmware must be compatible across members. (4) MC-LAG requires a peer link (ISL) and consistent LACP keys. (5) A stack reboot may disrupt all members if the master fails without redundancy planning.
Rule of thumb: always configure an explicit master priority and enable split-brain detection; verify stack topology and member health after changes.
Worked example: A 4-member stack loses its stack link, splitting 2+2. Without MAD, both halves claim master and duplicate IPs/MACs appear. With MAD (e.g., a witness or LACP-based detection), one fragment disables its ports, preserving service.