Synchronizing data between partition replication in Kafka, copying data from the leader of the partition to the follower requires a thread (replicationFetcherThread) follower (a follower is equivalent to a consumer) to actively pull messages from the leader in batches, which greatly improves throughput.
On the Fusioninsight Manager interface, when an alarm of insufficient disk capacity of Kafka is received, and the cause of the alarm has been eliminated from the hard disk hardware failure, the system administrator needs to consider expanding the capacity to solve this problem.
A typical Katka cluster contains several Producers, several Brokers, several Consumers and a ZooKeeper cluster.
What is the module used to manage the active and standby states of the Loader Server process in Loader?