we have this situation in our network where whenever we have a power shutdown due to maintenance its takes the C6500 core CPU to reach 90% and takes almost an hour or more to settle down.
we have DHCP server configured on it for almost 3000 devices, so i suspect DHCP request broadcast flooding, the console hangs and nothing can be accessed.
Thus we started to shut all uplink ports and get them UP 2-3 ports at a time which takes 1-2 hrs to get network operational.
we have had a crash before, trying to decode it...
Client seems to be confused as for him its the SUP720 which is the best and should not get impacted.
thinking of dividing the dhcp services on both cores !!
any technical justification will be highly appreciated.