How is Partitioning done in Hazelcast

Question

I am using Hazelcast v2.5. I have a few doubts related to partitioning in a cluster.

• How are the partitions identified ?

• When a m.get request is made how does Hazelcast identify in which partition the data resides? ( apart from the key )

• How is partitioning done when a new member joins the cluster?

• Can I get the entries in a partition if partition-ID is entered?

Answer 1

Hazelcast has a book that covers questions like yours. Partitions are inner technique for providing fault-tolerance. You can set partitions count (271 by default). As a client, you can work with more user-friendly objects, like Collections\Queues.. For maps, you can control backup-count and should hazelcast read from backups.

I guess you can manage partitions using PartitionManager