Refer to the exhibit. A customer currently has a large EIGRP-based network with several remote sites attached. All remote sites connect to the two corporate data centers, depicted as 10.1.1.0 and 10.1

Configuring EIGRP stub on the data center routers constrains EIGRP query propagation so that queries do not bounce between data centers or flood the entire network, directly eliminating the stuck-in-active condition that occurs when remote site routers cannot reply to active queries within the active timer.

Performing summarization at the data centers and selectively leaking only necessary specific routes to remote sites reduces the number of prefixes remote site routers must maintain in their topology tables, lowering CPU and memory consumption while preserving optimal path selection to both data centers.

Summarization at remote sites toward the data centers does not reduce the routing table load on the remote routers themselves and does nothing to limit the EIGRP query scope that causes stuck-in-active events.

Setting the hello interval larger than the hold interval is a fatal misconfiguration - the hold timer must always exceed the hello interval, otherwise the hold timer expires before a new hello arrives, causing unnecessary adjacency resets.

Increasing only the hold interval may reduce false adjacency drops on flapping links but does not shrink the routing table or limit query propagation, so it does not reduce load on remote site routers or address stuck-in-active.