Users Guide

Stacking 217
master. No changes or configuration are applied to the other stack members;
however, the dynamic protocols will try to reconverge as the topology could
change because of the failed unit.
If you remove a unit and plan to renumber the stack, issue a no member unit
command in Stack Configuration mode to delete the removed switch from
the configured stack member information.
How is the Firmware Updated on the Stack?
When adding a new switch to a stack, the Stack Firmware Synchronization
feature, if enabled, automatically synchronizes the firmware version with the
version running on the stack master per the configuration on the master
switch. The synchronization operation may result in either upgrade or
downgrade of firmware on the mismatched stack member. Use the boot auto-
copy-sw command to enable stack firmware synchronization (SFS).
Upgrading the firmware on a stack of switches is the same as upgrading the
firmware on a single switch. After downloading a new image by using the File
Download page or copy command, the downloaded image is distributed to all
the connected units of the stack. For more information about downloading
and installing images, see "Images and File Management" on page 503. When
copying firmware onto the switch in a stacked configuration, use the show sfs
and show version commands to check the status of stack firmware
synchronization prior to a reboot.
What is Stacking Standby?
The standby unit may be preconfigured or automatically selected. If the
current stack master fails, the standby unit becomes the stack master. If no
switch is preconfigured as the standby unit, the software automatically selects
a standby unit from among the existing stack units.
When the failed master resumes normal operation, it joins the stack as a
member (not as the master) if the new stack master has already been elected.
The stack master copies its running configuration to the standby unit
whenever it changes (subject to some restrictions to reduce overhead). This
enables the standby unit to take over the stack operation with minimal
interruption if the stack master becomes unavailable.