Kafka stream app can't get its own local state store value - apache-kafka-streams

Question

I'm new comer to kafka and playing WordCount.java which materialized to state store named "counts-store"
created a topic streams-plaintext-input with 2 partitions then running 2 WordCount apps in same machine, say App1 and App2
the issue is why App1 processed the words but local state store return null while App2 is able to get count value from local state store. in other words, one App's local state store is in peer's state store.
here is the producer
kafka-console-producer.bat --bootstrap-server localhost:9092 --topic streams-plaintext-input
log line is added so I can know App1 or App2 process the word
KStream<String,String> words = source.flatMapValues(value -> {
System.out.println(Thread.currentThread().getName() + " text line splitted ["+value+"]");
return Arrays.asList(value.toLowerCase(Locale.getDefault()).split("\\W+"));
});
so far so good, App1 and App2 process new coming words alternately.
then add query state store code to the same class
for(int i=0; i<1000; i++){
TimeUnit.SECONDS.sleep(10);
String qstr = "new3";
ReadOnlyKeyValueStore<String, Long> keyValueStore = streamsClient.store(StoreQueryParameters.fromNameAndType(StateStoreName, QueryableStoreTypes.keyValueStore()));
System.out.println(new Date().getTime()+ " count for "+qstr+":" + keyValueStore.get(qstr));
}
now producer console enter new3, the interesting log I got:
App1:
streams-wordcount-2p-209b88a2-7f7e-4533-8478-709a01f9ae1e-StreamThread-1 text line splitted [new3]
1666998201324 count for new3:null
1666998211335 count for new3:null
App2 who Not process the word:
1666998204348 count for new3:1
1666998204348 count for new3:1
App1 processed new3 then its state store should has count value, why the count value appears in App2?
in other words, one APP may unable to get its own value from its own local state store, what the idea behind this? does it mean we have to make every state store to be queryable(Interactive Queries)? because it will affect programming model.
please advise, thank you very much!