HazelcastAPI IPartitionService Arayüzü

Giriş

Şu satırı dahil ederiz

import com.hazelcast.internal.partition.IPartitionService ;

Kalıtım şöyle

CoreService

  IPartitionService

    InternalPartitionServiceImpl

Partition Arrangement

Açıklaması şöyle

Initial partition arrangement by default occurs lazily when the first partition operation hits the cluster. So once a partition op appears for execution, partition assignments will be split across the 2 members

constructor

InternalPartitionServiceImpl içindeki kod şöyle. İşte meşhur 271 sayısı ClusterProperty.PARTITION_COUNT sabitinden geliyor.

public InternalPartitionServiceImpl(Node node) {
  HazelcastProperties properties = node.getProperties();
  this.partitionCount = properties.getInteger(ClusterProperty.PARTITION_COUNT);
  ...
}

getPartitionCount metodu

Cluster içinde kaç tane partition olduğunu belirtir

isPartitionAssignmentDone metodu

Bu member için partition ataması yapılıp yapılmadığını belirtir

Örnek

Şöyle yaparız

if (!nodeEngine.getPartitionService().isPartitionAssignmentDone()
  && !hazelcastInstance.getCluster().getClusterState().isMigrationAllowed()) {
    // partitions have not been assigned yet and migrations are not allowed
    // so sql catalog can have no values -> skip check, because we anyway cannot
    // iterate over sqlCatalog values (operation will fail with exception, since
    // partition assignments do not exist and cannot be triggered).
    logger.info("Skipping data connection consistency check because"
                    + " initial partition assignment is not done yet.");
    return;
}