org.mule.routing.outbound
Class AbstractRoundRobinMessageSplitter

java.lang.Object
  org.mule.routing.outbound.AbstractOutboundRouter
      org.mule.routing.outbound.FilteringOutboundRouter
          org.mule.routing.outbound.AbstractMessageSplitter
              org.mule.routing.outbound.AbstractRoundRobinMessageSplitter

All Implemented Interfaces:

MuleContextAware, Disposable, Initialisable, MessageProcessor, MessageRouter, Matchable, MatchableMessageProcessor, MatchableMessageRouter, OutboundRouter, RouterStatisticsRecorder, TransformingMatchable

Direct Known Subclasses:

ExpressionMessageSplitter, ListMessageSplitter, XmlMessageSplitter

public class AbstractRoundRobinMessageSplitter
extends AbstractMessageSplitter

FilteringListMessageSplitter accepts a List as a message payload then routes list elements as messages over an endpoint where the endpoint's filter accepts the payload.

Field Summary

Fields inherited from class org.mule.routing.outbound.FilteringOutboundRouter

expressionManager

Fields inherited from class org.mule.routing.outbound.AbstractOutboundRouter

enableCorrelation, logger, magicProperties, muleContext, replyTo, resultsHandler, routes, transactionConfig

Fields inherited from interface org.mule.api.lifecycle.Initialisable

PHASE_NAME

Fields inherited from interface org.mule.api.lifecycle.Disposable

PHASE_NAME

Constructor Summary

AbstractRoundRobinMessageSplitter()

Method Summary

protected SplitMessage	getMessageParts(MuleMessage message, List<MessageProcessor> endpoints) Retrieves a specific message part for the given endpoint.
void	initialise() Method used to perform any initialisation work.
boolean	isDeterministic() If this option is true (the default) then the first message part is routed to the first endpoint, the second part to the second endpoint, etc, with the nth part going to the (n modulo number of targets) endpoint.
boolean	isDisableRoundRobin() The default behaviour for splitter routers is to round-robin across targets.
boolean	isFailIfNoMatch() If none of the targets match a split message part i.e.
void	setDeterministic(boolean deterministic) If this option is true (the default) then the first message part is routed to the first endpoint, the second part to the second endpoint, etc, with the nth part going to the (n modulo number of targets) endpoint.
void	setDisableRoundRobin(boolean disableRoundRobin) The default behaviour for splitter routers is to round-robin across targets.
void	setFailIfNoMatch(boolean failIfNoMatch) If none of the targets match a split message part i.e.
protected List	splitMessage(MuleMessage message) Method used just to split the message into parts.

Methods inherited from class org.mule.routing.outbound.AbstractMessageSplitter

route

Methods inherited from class org.mule.routing.outbound.FilteringOutboundRouter

addRoute, getFilter, getRoute, getTransformers, isMatch, isTransformBeforeMatch, isUseTemplates, setFilter, setTransformers, setUseTemplates

Methods inherited from class org.mule.routing.outbound.AbstractOutboundRouter

dispose, getEnableCorrelation, getMuleContext, getReplyTo, getResultsHandler, getRoute, getRouterStatistics, getRoutes, getTransactionConfig, isDynamicRoutes, process, propagateMagicProperties, removeRoute, sendRequest, sendRequestEvent, setEnableCorrelation, setEnableCorrelationAsString, setMessageProcessors, setMessageProperties, setMuleContext, setReplyTo, setResultsHandler, setRouterStatistics, setRoutes, setTransactionConfig

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.mule.api.processor.MessageRouter

addRoute, removeRoute

Methods inherited from interface org.mule.api.processor.MessageProcessor

process

Constructor Detail

AbstractRoundRobinMessageSplitter

public AbstractRoundRobinMessageSplitter()

Method Detail

initialise

public void initialise()
                throws InitialisationException

Description copied from interface: Initialisable

Method used to perform any initialisation work. If a fatal error occurs during initialisation an InitialisationException should be thrown, causing the Mule instance to shutdown. If the error is recoverable, say by retrying to connect, a RecoverableException should be thrown. There is no guarantee that by throwing a Recoverable exception that the Mule instance will not shut down.

Specified by:

initialise in interface Initialisable

Overrides:

initialise in class FilteringOutboundRouter

Throws:

InitialisationException - if a fatal error occurs causing the Mule instance to shutdown

RecoverableException - if an error occurs that can be recovered from

splitMessage

protected List splitMessage(MuleMessage message)

Method used just to split the message into parts. Each part should be an entry in the list. The list can contain either MuleMessage objects or just payloads (Mule will automatically convert the payloads into messages).

This method can be overridden by custom implementations of splitter router where the distribution of the message parts will be done using either round robin or endpoint filtering.

Parameters:

message - the source message to split into parts

Returns:

a list of payload objects or MuleMessage objects. Usually, it is sufficient just to return payload objects

getMessageParts

protected SplitMessage getMessageParts(MuleMessage message,
                                       List<MessageProcessor> endpoints)

Retrieves a specific message part for the given endpoint. the message will then be routed via the provider.

NOTE:Implementations must provide proper synchronization for shared state (payload, properties, etc.)

Specified by:

getMessageParts in class AbstractMessageSplitter

Parameters:

message - the current message being processed

endpoints - A list of targets that will be used to dispatch each of the parts

Returns:

a List of message parts. Each part will become the payload of the outgoing message. Note that the parts will be dispatched to

See Also:

SplitMessage, AbstractRoundRobinMessageSplitter

isDeterministic

public boolean isDeterministic()

If this option is true (the default) then the first message part is routed to the first endpoint, the second part to the second endpoint, etc, with the nth part going to the (n modulo number of targets) endpoint. If false then the messages will be distributed equally amongst all targets.

The behaviour changes if the targets have filters since the message part will get routed based on the next endpoint that follows the above rule AND passes the endpoint filter.

Returns:

true if deterministic has been set to true

setDeterministic

public void setDeterministic(boolean deterministic)

If this option is true (the default) then the first message part is routed to the first endpoint, the second part to the second endpoint, etc, with the nth part going to the (n modulo number of targets) endpoint. If false then the messages will be distributed equally amongst all targets.

The behaviour changes if the targets have filters since the message part will get routed based on the next endpoint that follows the above rule AND passes the endpoint filter.

Parameters:

deterministic - the value to set

isDisableRoundRobin

public boolean isDisableRoundRobin()

The default behaviour for splitter routers is to round-robin across targets. When using filters on targets it is sometimes desirable to use only the filters to control which endpoint the split message part goes too. For example, if you have 3 targets where two have a filter but the last does not, you'll need to disable round robin since the 3rd endpoint may end up routing a message that one of the other targets should have routed. Generally it is good practice to either configure all targets with filters or none, in this case there is not need to set this property.

Returns:

true if disabled

setDisableRoundRobin

public void setDisableRoundRobin(boolean disableRoundRobin)

The default behaviour for splitter routers is to round-robin across targets. When using filters on targets it is sometimes desirable to use only the filters to control which endpoint the split message part goes too. For example, if you have 3 targets where two have a filter but the last does not, you'll need to disable round robin since the 3rd endpoint may end up routing a message that one of the other targets should have routed. Generally it is good practice to either configure all targets with filters or none, in this case there is not need to set this property.

Parameters:

disableRoundRobin - true if disabled

isFailIfNoMatch

public boolean isFailIfNoMatch()

If none of the targets match a split message part i.e. each endpoint has a filter for a certain message part. This flag controls whether the part is ignorred or an exceptin is thrown.

Returns:

true if an exception should be thrown when no match is found

setFailIfNoMatch

public void setFailIfNoMatch(boolean failIfNoMatch)

If none of the targets match a split message part i.e. each endpoint has a filter for a certain message part. This flag controls whether the part is ignorred or an exceptin is thrown.

Parameters:

failIfNoMatch - true if an exception should be thrown when no match is found