Cisco C9300 and 3850 Series Stacking FAQ and Quick Start Guide

Quick Start: Stacking Cisco 3850 & C9300 Series Switches

  • Power off all switches before stacking. Stacking should always be done with all units powered down to avoid accidental reloads or merging issues.

  • Confirm that all switches are the same series (3850 or C9300 only), and ideally the same exact hardware model for best compatibility.

  • Verify all switches have the same IOS XE image and license level. Use show version and show license right-to-use to confirm. Use the auto-upgrade feature if needed.

  • Rack and cable all switches. For stacking, use only Cisco-certified StackWise cables (do not reuse 3750 or non-matching cables). For C9300L, fit the StackWise adapters and cable as per Cisco’s guide.

  • Cable the stacks in a ring:

    • Connect stack port 1 (P1) on switch 1 to stack port 2 (P2) on switch 2.

    • Continue this for all members, then connect the last switch’s P1 to the first switch’s P2, closing the ring.

  • For power redundancy (StackPower), connect up to 4 switches using StackPower cables. This combines all power supplies into a single virtual power resource.

  • Label every switch and stacking cable for future maintenance and troubleshooting.

  • Power on all switches at the same time (ideally within 1 minute) to ensure clean stack formation.

  • The master (active) switch is elected. To influence election, set priorities beforehand (15 = highest) with switch <number> priority <1-15> in config mode.

  • Console into the intended master switch and check stack status:

    • show switch

    • show switch stack-ports

    • show switch stack-ring speed

  • All switches should report “Ready” and all stack ports should show “OK.”

  • If any units show “Version Mismatch” or “Lic-Mismatch”, correct licensing or upgrade images to match the stack.

  • Configure the stack using the active/master switch; all ports become visible as one logical switch.

Best Practices and Recommendations

  • Always use ring topology, not chain. A full ring ensures redundancy—if one cable fails, the stack still operates.

  • Assign high priority to your most important/central/master switch and document all priorities.

  • Label and document all cable connections and stack member positions for future expansion or troubleshooting.

  • Run show switch to verify and document the master/standby/member roles after every stack change.

  • Keep a detailed diagram of physical and logical layouts, including member IDs, stack port assignments, and rack location.

  • Maintain spare stack cables and power modules to minimize downtime in the event of a failure.

  • For large deployments or frequent moves, test failover by rebooting the stack master and verifying seamless standby takeover.

  • Use only supported stack cable lengths (consult Cisco data sheets for maximum run). Overly long or damaged cables cause random drops and stack split-brain scenarios.

  • For C9300 (and 9300X/C9300L) series, be aware of StackWise-1T, -480, and -320 variants—do not mix incompatible stack types.

  • For mixed models within a series (such as 24- and 48-port 3850s), this is supported, but all must run the same IOS XE image/version.

Advanced Tips & Stack Expansion

  • Before adding new members, pre-stage and match IOS images. Power the new switch down, cable, rack, then power up.

  • Use the software auto-upgrade command on the active switch so new members are automatically upgraded on join.

  • When expanding, always use the correct StackWise adapter or cable kit for the exact model (especially 9300L!).

  • Use StackPower for energy redundancy for 3850 and supported 9300 models: power-pool “power sharing” can be set in “sharing" or “redundant” mode.

  • To remove a switch, copy the stack config, power off and safely remove cables. Always reverify stack health after removal using show switch.

Advanced Troubleshooting

  • If the stack does not form: check all cabling/logos are correct, all IOS images and licenses match, and that auto-upgrade is enabled if needed.

  • Use show switch stack-ports to verify all stack links display “OK.” If not, check physical connections and cable seating.

  • If you experience a split stack (split-brain/partition): correct physical connections, reboot all members, and verify with show switch.

  • For persistent failures or crash reload loops, break the stack, power ON switches one at a time, and upgrade/relicense individually before reforming stack.

  • Member switches that will not join may have boot variables set incorrectly. Use show boot and boot system commands to fix.

  • Always back up configuration before any stack changes!

FAQ—Cisco 3850 & 9300 Series Stacking

Can I mix 3850 and 9300 switches in a stack?

  • No. Only same-series switches are supported per stack. 3850s with 3850s, C9300s with C9300s.

Can I mix 9300 and 9300L/C9300X?

  • 9300X and regular 9300 can stack together at the StackWise-480 bandwidth, but C9300L only stacks with other 9300L models.

What is the maximum number of stacked switches?

  • Up to 9 for 3850, up to 8 for 9300/9300X/9300L.

Does adding/removing a switch cause downtime?

  • Removing a member reboots the stack. Adding a member should be done with the stack powered off, but auto-upgrade helps add new units while minimizing disruption.

How do I change the master switch?

  • Set the priority (switch <id> priority <1-15>) and reload the stack. Upon election, the highest-priority switch active at boot time becomes master.

Stack cable/port fails—does the stack go down?

  • In a ring topology, a single connection failure downgrades bandwidth but the stack remains operational (rings turn into chains). Redundancy is lost; always repair ASAP.

Can I run both data stack and power stack?

  • On 3850 and supported 9300, yes: Data stack for management/traffic, Power stack for shared power redundancy.

How do I check stack state/health?

  • show switch

  • show switch stack-ports

  • show version

  • show environment stack

  • SNMP/NMS for monitoring

Pro Tips for Seamless Network Operation

  • Always keep current, version-matched backups of configurations.

  • Use physical and logical documentation for every stack.

  • Perform regular, scheduled failover and stack health checks—especially after changes or firmware upgrades.

  • Keep extra stack cables, adapters, and power supplies on hand for emergencies.

Key CLI Reference

  • show switch

  • show switch stack-ports

  • show stack-power

  • switch <n> priority <1-15>

  • software auto-upgrade enable

  • show license right-to-use

  • show boot

Stacking Cisco switches the right way improves performance, reliability, management, and uptime for your entire LAN. For ongoing support, design, or upgrades, Reef Telecom’s team of Cisco experts are here to help.

Back to blog

Leave a comment

Please note, comments need to be approved before they are published.