Similar zookeeper linux file system, the root "/" path starting called znode nodes, child nodes can have znode, each node may have data
Znode node types are: permanent, temporary, permanent and orderly, temporary and orderly. When a node is to create an orderly creation if it already exists, just behind the count increases, 2, 3 copies of similar windows file the same name and so on
There node operation: create create, get read, set modify, delete delete the current node, can not have child nodes (similar rm delete the folder), rmr delete the current node and child node, watcher monitor node
You can listen to the event node: The node to create, modify data nodes, delete nodes, to create a child node [delete listener will notify the parent node]
zookeeper will be persisted to zoo.cfg of dataDir directory
more than the nature of the zookeeper, zookeeper may decide to do:
cluster management, unified naming service, distributed configuration management, distributed message queues, distributed lock, distributed notification and coordination