These are oracle policies. Actually we have configured for both server (active and standby) and at a time policy is active for active server and for standby server policy is in deactivated state.
If there is a issue with active server, then standby server will be activated. And we will deactivated first policy and activated another one.
Because we dont have VIP/cluster in my environment.
OK - understood...
Who is initiating backup NBU or a crontab/task scheduler from client?
If backups are initiated from client side - I would relay only on one policy and in clients tab include both servers active and standby.... But if NBU is scheduling these than your approach seems to be OK..
Another way that may work would be to create a CNAME alias which points to the current active node. So as part of you failover process you update the DNS record. This may also help to reduce overhead when switching as clients may not require any changes to keep working, just reconnect. You would need to revie the TTL for the CNAME and host records to ensure changes propagate correctly.
Also within NetBackup to support this you may need to create client aliases as the RMAN stream will still come from the node. Have a look at appendix B of the NetBackup Oracle admin guide around how to support Oracle RAC which is kind of what you have without the cost.
And to answer your original question - yes having deactivated policies does have an impact on the master server as it creates more work to process. That said the imapct these days would be negligible on any current system, so I wouldn't be concerned.