org.apache.flume.node

Class PropertiesFileConfigurationProvider

java.lang.Object

org.apache.flume.node.AbstractConfigurationProvider

org.apache.flume.node.PropertiesFileConfigurationProvider

All Implemented Interfaces:

ConfigurationProvider

Direct Known Subclasses:

PollingPropertiesFileConfigurationProvider

public class PropertiesFileConfigurationProvider
extends AbstractConfigurationProvider

A configuration provider that uses properties file for specifying configuration. The configuration files follow the Java properties file syntax rules specified at Properties.load(java.io.Reader). Every configuration value specified in the properties file is prefixed by an Agent Name which helps isolate an individual agent's namespace.

Valid configuration files must observe the following rules for every agent namespace.

• For every <agent name> there must be three lists specified that include <agent name>.sources, <agent name>.sinks, and <agent name>.channels. Each of these lists must contain a space separated list of names corresponding to that particular entity.
• For each source named in <agent name>.sources, there must be a non-empty type attribute specified from the valid set of source types. For example: <agent name>.sources.<source name>.type = event
• For each source named in <agent name>.sources, there must be a space-separated list of channel names that the source will associate with during runtime. Each of these names must be contained in the channels list specified by <agent name>.channels. For example: <agent name>.sources.<source name>.channels = <channel-1 name> <channel-2 name>
• For each source named in the <agent name>.sources, there must be a runner namespace of configuration that configures the associated source runner. For example: <agent name>.sources.<source name>.runner.type = avro. This namespace can also be used to configure other configuration of the source runner as needed. For example: <agent name>.sources.<source name>.runner.port = 10101
• For each source named in <sources>.sources there can be an optional selector.type specified that identifies the type of channel selector associated with the source. If not specified, the default replicating channel selector is used.
• For each channel named in the <agent name>.channels, there must be a non-empty type attribute specified from the valid set of channel types. For example: <agent name>.channels.<channel name>.type = mem
• For each sink named in the <agent name>.sinks, there must be a non-empty type attribute specified from the valid set of sink types. For example: <agent name>.sinks.<sink name>.type = hdfs
• For each sink named in the <agent name>.sinks, there must be a non-empty single-valued channel name specified as the value of the channel attribute. This value must be contained in the channels list specified by <agent name>.channels. For example: <agent name>.sinks.<sink name>.channel = <channel name>
• For each sink named in the <agent name>.sinks, there must be a runner namespace of configuration that configures the associated sink runner. For example: <agent name>.sinks.<sink name>.runner.type = polling. This namespace can also be used to configure other configuration of the sink runner as needed. For example: <agent name>.sinks.<sink name>.runner.polling.interval = 60
• A fourth optional list <agent name>.sinkgroups may be added to each agent, consisting of unique space separated names for groups
• Each sinkgroup must specify sinks, containing a list of all sinks belonging to it. These cannot be shared by multiple groups. Further, one can set a processor and behavioral parameters to determine how sink selection is made via <agent name>.sinkgroups.< group name<.processor. For further detail refer to inividual processor documentation
• Sinks not assigned to a group will be assigned to default single sink groups.

Apart from the above required configuration values, each source, sink or channel can have its own set of arbitrary configuration as required by the implementation. Each of these configuration values are expressed by fully namespace qualified configuration keys. For example, the configuration property called capacity for a channel called ch1 for the agent named host1 with value 1000 will be expressed as: host1.channels.ch1.capacity = 1000.

Any information contained in the configuration file other than what pertains to the configured agents, sources, sinks and channels via the explicitly enumerated list of sources, sinks and channels per agent name are ignored by this provider. Moreover, if any of the required configuration values are not present in the configuration file for the configured entities, that entity and anything that depends upon it is considered invalid and consequently not configured. For example, if a channel is missing its type attribute, it is considered misconfigured. Also, any sources or sinks that depend upon this channel are also considered misconfigured and not initialized.

Example configuration file:

 #
 # Flume Configuration
 # This file contains configuration for one Agent identified as host1.
 #

 host1.sources = avroSource thriftSource
 host1.channels = jdbcChannel
 host1.sinks = hdfsSink

 # avroSource configuration
 host1.sources.avroSource.type = org.apache.flume.source.AvroSource
 host1.sources.avroSource.runner.type = avro
 host1.sources.avroSource.runner.port = 11001
 host1.sources.avroSource.channels = jdbcChannel
 host1.sources.avroSource.selector.type = replicating

 # thriftSource configuration
 host1.sources.thriftSource.type = org.apache.flume.source.ThriftSource
 host1.sources.thriftSource.runner.type = thrift
 host1.sources.thriftSource.runner.port = 12001
 host1.sources.thriftSource.channels = jdbcChannel

 # jdbcChannel configuration
 host1.channels.jdbcChannel.type = jdbc
 host1.channels.jdbcChannel.jdbc.driver = com.mysql.jdbc.Driver
 host1.channels.jdbcChannel.jdbc.connect.url = http://localhost/flumedb
 host1.channels.jdbcChannel.jdbc.username = flume
 host1.channels.jdbcChannel.jdbc.password = flume

 # hdfsSink configuration
 host1.sinks.hdfsSink.type = hdfs
 host1.sinks.hdfsSink.hdfs.path = hdfs://localhost/
 host1.sinks.hdfsSink.batchsize = 1000
 host1.sinks.hdfsSink.runner.type = polling
 host1.sinks.hdfsSink.runner.polling.interval = 60
 

See Also:

Properties.load(java.io.Reader)

Constructor Summary

Constructors

Constructor and Description
PropertiesFileConfigurationProvider(String agentName, File file)

Method Summary

Methods

Modifier and Type	Method and Description
FlumeConfiguration	getFlumeConfiguration()

Methods inherited from class org.apache.flume.node.AbstractConfigurationProvider

getAgentName, getConfiguration

Methods inherited from class java.lang.Object

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

Constructor Detail

PropertiesFileConfigurationProvider

public PropertiesFileConfigurationProvider(String agentName,
                                   File file)

Method Detail

getFlumeConfiguration

public FlumeConfiguration getFlumeConfiguration()

Specified by:

getFlumeConfiguration in class AbstractConfigurationProvider