Hi, Can Kafka be configured (disable data distribution and replicas) in a HPC environment where the parallel file system and distributed storage perform the parallelism, replicas, and fault tolerant, the log and topic / partitions can be viewed by multiple brokers in the VMs without needing Kafka to distribute the data to brokers? Where such environment help or hinder kafka performance?
Also, will one topic with large number of partitions be better performance than large number of topics each with one partition in terms of resource consumer and scalability? Thank you. Kind regards, - j