Hi,
"Question No1. What are all the problem will occurs due to " VPC SPLIT BRAIN" scenario."
It’s difficult to say exactly all the problems you could face in a vPC split brain scenario.
There could simply be issues with connectivity as without the vPC peer link, MAC addresses may not be synchronised correctly.
If you have devices that are only connected to one of the vPC peers i.e,, a device connected to an “orphan port”, then connectivity to these is likely to be affected. For example, assume you have a host H1 connected with a vPC to peers S1 and S2 sending traffic to a host only connected to one of the peers e.g., S2. If host H1 sent traffic on the port connected to S1 then traffic would normally be carried on the vPC peer link. In the case of the vPC peer-link failure there would be no path.
Intermittent connectivity is one issue, but a worse case would probably be a bridging loop leading to a broadcast storm. In a split brain scenario both peers have all their vPC member ports forwarding, and the vPC loop prevention does not stop a frame received on one vPC member being sent on another vPC member port. Assuming spanning tree is still enabled, this should break the loops, but loops may be seen before spanning tree is able to block te necessary links.
"Question No2 :What problems will occur while ""primary -operational secondary, Secondary-operational primary situation. Is it mandatory or Recommended to manually give back the role to primary ?"
There should be no issues in an environment where all devices are operating, but the configured vPC primary peer is the operational secondary and vice versa.
The one area where you would want to consider this is in respect of singly attached devices. If you have a large number of end systems that are singly attached i.e., with only one NIC, the recommendation is to attach these to the primary peer. In this case it is best if the configured primary is always the operational primary.
Regards
