ExtremeXOS Stacking Configuration Step by Step - 夜莺博客

ExtremeXOS Stacking Configuration Step by Step

ExtremeXOS stacking has one requirement that catches almost every first-time installer: stacking-support and stacking are two separate things. Enabling support makes a switch capable of joining a stack; enabling stacking actually builds it. A stack will not form if support is disabled on any member, and on older platforms such as the X440-G2, X590, X620, X690, and X870 it is disabled by default. This guide walks the full procedure — image consistency, stacking-support, port verification, easy setup, redundancy, and validation — in the order that avoids rebooting the stack twice.

Before You Configure Anything

Pre-checks that avoid a second reboot

Two pre-checks save most of the pain. First, every node must be running the same ExtremeXOS version and patch level, and must be booted from the same partition. Check each switch individually:

switch1 # show switch

Look at the Image Booted field for every node: if one switch booted from the secondary partition and the others from primary, the stack will either refuse to form or form and then fail. Second, physically cable the stacking ports per the supported topology (ring is preferred over daisy chain, because a single link failure does not split the stack).

Step 1 - Enable Stacking Support on Every Node

switch1 # enable stacking-support
switch1 # reboot

Run this on every switch that will be a member, then reboot. On new switches stacking support is enabled by default, but on X440-G2 and the X590/X620/X690/X870 families you must enable it explicitly. If a switch previously belonged to another stack, clear that state before proceeding:

switch1 # show stacking configuration
switch1 # disable stacking node-address 00:04:96:9c:e4:39
switch1 # reboot

The disable stacking node-address command takes the MAC of the node and clears its stacking configuration. Do this on every node that was part of a previous stack, or easy setup will inherit the old slot numbers and MAC.

Step 2 - Verify the Stacking Ports

Confirm that the cabled stacking ports are enabled and showing the expected link state before you try to build the stack:

switch1 # show ports 1,2
switch1 # show stacking stack-ports

A port in Link Down state means a cabling or technology problem, not a configuration problem: check the cable, and confirm that both ends use the same stacking technology (SummitStack-V, V80, V160 or V320). Mixing technologies on one link will never come up.

If you plan to stack over Ethernet ports rather than dedicated stacking ports, enable the alternate ports and — for the enhanced protocols — select the matching protocol:

switch1 # configure stacking protocol enhanced
switch1 # reboot

Enhanced protocol is mandatory for MPLS-capable stacks and for SummitStack-V80/V160/V320 on supported platforms.

Step 3 - Build the Stack with Easy Setup

Easy setup does five things at once on every node: enables stacking, configures a stack MAC address, assigns slot numbers (starting with slot 1 on the switch you run it from), configures the redundancy mode, and reboots the stack into the new configuration. Run it from the console of the switch that should be the master:

switch1 # enable stacking
You have not yet configured all required stacking parameters.
Would you like to perform an easy setup for stacking operation? (y/N) y

Answering y invokes configure stacking easy-setup. Slot 1 goes to the local node and becomes the master; other nodes take the next available slots and become backup and standby nodes. After the automatic reboot, verify the result:

switch1 # show stacking
switch1 # show stacking configuration

Reading the flags column

The Flags column is the part worth reading carefully. C means master-capable, A means active node in this topology, E means stacking is currently enabled, and M means the stack MAC in use matches the configured one. A node with lowercase letters only is configured but not active — usually a stacking link problem rather than a software one.

Step 4 - Set Redundancy and Priorities

Easy setup configures minimal redundancy. For production, prefer master-capable redundancy so that two nodes can take over as master, and set explicit priorities so the master election is deterministic rather than decided by MAC address:

switch1 # configure stacking redundancy master-capable
switch1 # configure stacking slot 1 priority 100
switch1 # configure stacking slot 2 priority 90

Slots 1 and 2 should be physically separate switches fed from different power sources if the stack is meant to survive a PSU or rack feed failure.

Step 5 - Validate and Maintain

switch1 # show stacking
switch1 # show stacking detail
switch1 # show stacking stack-ports
switch1 # show switch
switch1 # show version

show stacking detail confirms the topology (ring or daisy chain), the member slot numbers, the master, and the backup. If the topology reads as daisy chain when you cabled a ring, one of the two links in the ring is down — check show stacking stack-ports again and inspect the physical cabling.

Two ongoing rules for a healthy stack:

  • Upgrade all members from the same image, and plan for a stack-wide reboot when the active image changes.
  • Before adding or removing a member, verify that the new node has the same ExtremeXOS version and that stacking-support is enabled. A mismatched node left cabled into the stack causes continued topology instability.

Operational Notes

  • Ring topology over daisy chain — a ring survives one link failure, a daisy chain splits.
  • Clear inherited stacking state with disable stacking node-address before reusing hardware.
  • Set explicit slot priorities so the master election does not depend on MAC ordering.
  • Match stacking technology on both ends of every stacking link.
  • Keep the configuration consistent with the physical layout — a stack whose slot numbers no longer match the rack is a future outage.

Related reading: our ExtremeXOS password recovery guide, the stack split troubleshooting guide, and the switch stacking fundamentals article.

原文链接:https://documentation.extremenetworks.com/exos_22.6/GUID-8C7AE18B-B3A3-41B6-8884-99A8354FED94.shtml