Package org.apache.hadoop.hdfs.server.diskbalancer.connectors

package org.apache.hadoop.hdfs.server.diskbalancer.connectors

Connectors package is a set of logical connectors that connect to various data sources to read the hadoop cluster information. We currently have 3 connectors in this package. They are DBNameNodeConnector - This uses the connector from the original balancer package to connect to a real hadoop cluster. JsonNodeConnector - This connects to a file and reads the data about a cluster. We can generate a cluster json from a real cluster using the diskBalancer tool or hand-craft it. There are some sample Json files checked in under test/resources/diskBalancer directory. NullConnector - This is an in-memory connector that is useful in testing. we can crate dataNodes on the fly and attach to this connector and ask the diskBalancer Cluster to read data from this source.

Related Packages

Package	Description
org.apache.hadoop.hdfs.server.diskbalancer	Disk Balancer connects to a .DataNode and attempts to spread data across all volumes evenly.
org.apache.hadoop.hdfs.server.diskbalancer.command	Commands for disk balancer command line tool.
org.apache.hadoop.hdfs.server.diskbalancer.datamodel	Disk Balancer Data Model is the Data Model for the cluster that Disk Balancer is working against.
org.apache.hadoop.hdfs.server.diskbalancer.planner	Planner takes a DiskBalancerVolumeSet, threshold and computes a series of steps that lead to an even data distribution between volumes of this DiskBalancerVolumeSet.

Class	Description
org.apache.hadoop.hdfs.server.diskbalancer.connectors.ClusterConnector	ClusterConnector interface hides all specifics about how we communicate to the HDFS cluster.
org.apache.hadoop.hdfs.server.diskbalancer.connectors.ConnectorFactory	Connector factory creates appropriate connector based on the URL.
org.apache.hadoop.hdfs.server.diskbalancer.connectors.JsonNodeConnector	A connector that understands JSON data cluster models.