Lifecycle · Catalyst StackWise
Plan a StackWise upgrade cleanly — member roles, priorities, auto-upgrade handling, V-mismatch recovery, and a step-by-step cutover.
A Catalyst stack is one logical switch built from up to 8 physical members. Upgrading across a stack is a single operation — if you set it up right. If you don't, you get a split brain, a V-mismatch, or 8 hours of port flap while members silently TFTP-fetch 3 GB images over an already congested management VLAN.
Software auto-upgrade lets new stack members inherit the active image when added. Combined with install mode, this makes RMAs a plug-and-play affair. For cross-LTS upgrades, the whole stack reloads together — plan a ~10 minute outage.
StackWise-480 (9200/9300) and StackWise-1T (9300X) share a backplane that handles forwarding, control, and image consistency. Priorities determine who becomes Active. Set priority 15 on your preferred Active and 14 on the preferred Standby. If you don't set priorities, whoever boots first wins — which can cause port numbering to drift after a power cycle.
A stack member comes up on a different IOS-XE version than the Active. With auto-upgrade, the Active sync-copies its image to the new member. Without, the new member stays in V-mismatch (ports suspended) until you manually copy and reload.
Yes — it's the default on 17.x. A single install add file … activate commit handles all members.
Enable auto-upgrade, stage image on member 1, install add file … activate commit. All 4 members reload together after ~6 minutes, stack comes up on 17.12.
Replace failed member 2. Plug in, boot. Auto-upgrade syncs image from Active. Member joins as member 2 with inherited config. No manual IOS copy.