org.apache.gobblin.util

Class ClustersNames

java.lang.Object

org.apache.gobblin.util.ClustersNames

public class ClustersNames
extends Object

Allows conversion of URLs identifying a Hadoop cluster (e.g. resource manager url or a job tracker URL) to a human-readable name.

The class will automatically load a resource named URL_TO_NAME_MAP_RESOURCE_NAME to get a default mapping. It expects this resource to be in the Java Properties file format. The name of the property is the cluster URL and the value is the human-readable name.

IMPORTANT: Don't forget to escape colons ":" in the file as those may be interpreted as name/value separators.

Field Summary

Fields

Modifier and Type	Field and Description
static String	URL_TO_NAME_MAP_RESOURCE_NAME

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	ClustersNames()

Method Summary

Modifier and Type	Method and Description
void	addClusterMapping(String clusterUrl, String clusterName)
void	addClusterMapping(URL clusterUrl, String clusterName)
String	getClusterName() Returns the cluster name on which the application is running.
String	getClusterName(org.apache.hadoop.conf.Configuration conf) Returns the cluster name on which the application is running.
String	getClusterName(String clusterUrl) Returns human-readable name of the cluster.
static ClustersNames	getInstance()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

URL_TO_NAME_MAP_RESOURCE_NAME

public static final String URL_TO_NAME_MAP_RESOURCE_NAME

See Also:

Constant Field Values

Constructor Detail

ClustersNames

protected ClustersNames()

Method Detail

getClusterName

public String getClusterName(String clusterUrl)

Returns human-readable name of the cluster. Method first checks config for exact cluster url match. If nothing is found, it will also check host:port and just hostname match. If it still could not find a match, hostname from the url will be returned. For incomplete or invalid urls, we'll return a name based on clusterUrl, that will have only alphanumeric characters, dashes, underscores and dots.

addClusterMapping

public void addClusterMapping(String clusterUrl,
                              String clusterName)

addClusterMapping

public void addClusterMapping(URL clusterUrl,
                              String clusterName)

getClusterName

public String getClusterName()

Returns the cluster name on which the application is running. Uses default hadoop Configuration to get the url of the resourceManager or jobtracker. The URL is then translated into a human readable cluster name using getClusterName(String)

See Also:

getClusterName(Configuration)

getClusterName

public String getClusterName(org.apache.hadoop.conf.Configuration conf)

Returns the cluster name on which the application is running. Uses Hadoop configuration passed in to get the url of the resourceManager or jobtracker. The URL is then translated into a human readable cluster name using getClusterName(String)

MapReduce mode Uses the value for "yarn.resourcemanager.address" from Configuration excluding the port number.

Standalone mode (outside of hadoop) Uses the Hostname of InetAddress.getLocalHost()

Use getClusterName(String) if you already have the cluster URL

Parameters:

conf - Hadoop configuration to use to get resourceManager or jobTracker URLs

See Also:

getClusterName()

getInstance

public static ClustersNames getInstance()