ProSafe 7000 Managed Switch Software Administration Manual, Release 8.0.3
Managing Switch Stacks
17-15
v1.0, June 2010
4.
After a unit type is preconfigured for a specific unit number, attaching a unit with different unit type for
this unit number causes the switch to report an error. The
show switch
command indicates “config
mismatch” for the new unit and the ports on that unit don’t come up. To resolve this situation the
customer may change the unit number of the mismatched unit or delete the preconfigured unit type
using the
no member
<unit-id> command.
Upgrading Firmware
New code is downloaded via TFTP or xmodem to the management unit using the
copy
command. Once
code is successfully loaded on the management unit, it automatically propagates the code to the other units
in the stack. If some error occurs during code propagation to stack units then the
archive
command (in stack
configuration mode) may be issued to make another attempt to copy the software to the unit(s) that did not
get updated. Errors during code propagation to stack members could be caused by stack cable movement or
unit reconfiguration during the propagation phase. An error could also occur in the presence of excessive
network traffic (such as a broadcast event).
All units in the stack must run the same code version. Ports on stack units that don’t match the management
unit code version don’t come up and the
show switch
command shows a “code mismatch” error. To resolve
this situation the administrator may issue
archive
command. This command copies management unit’s
software to the other units with mismatched code version. Before issuing this command, be sure the code
running on the management unit is the desired code revision for all units in the stack. Once code is loaded to
all members of the stack, the units must be reset in order for the new code to start running.
Migration of Configuration With a Firmware Upgrade
In some cases, a configuration may not be carried forward in a code update. For updates where this issue is
to be expected, the following procedure should be followed:
1.
Save the current configuration by uploading it from the stack, using the copy command from the CLI.
2.
Load new code into the stack manager. Reboot the stack.
3.
Upon reboot, go into the boot menu and erase the configuration (“restore to factory defaults”)
4.
Continue with boot of operational code.
5.
Once the stack is up, download the saved configuration back to the master. This configuration should
then be automatically propagated to all members of the stack.
Code Mismatch
If a unit is added to a stack and it does not have the same version of code as that of the master, the following
should happen:
•
The “new” unit will boot up and become a “member” of the stack.