Under normal circumstances, it is not necessary to reboot the BPEs slots. After rebooting a controlling switch, BPE upstream port down events cause the BPE's software to bring down all extended ports. This makes the BPE slot appear as down to any adjacently attached devices, and traffic properly re-converges on any redundant paths. When the controlling bridge switch comes back up, the BPE comes back up without any intervention. Therefore, using the commands reboot or unconfigure switch {all | erase [all | nvram]} does not reboot the attached BPEs. To reboot attached BPEs, you must use the all option.
To reboot all attached BPEs, use the following command specifying the all option:
reboot {[time mon day year hour min sec] | cancel} {slot slot-number} | node-address node-address | stack-topology {as-standby} | all} | rolling}
To reboot a single BPE, use the preceding command, specifying the BPE's slot in slot-number.