org.gridgain.grid.spi.topology.nodefilter
Class GridNodeFilterTopologySpi

java.lang.Object
  org.gridgain.grid.spi.GridSpiAdapter
      org.gridgain.grid.spi.topology.nodefilter.GridNodeFilterTopologySpi

All Implemented Interfaces:

GridSpi, GridSpiJsonConfigurable, GridSpiManagementMBean, GridTopologySpi, GridNodeFilterTopologySpiMBean

@GridSpiInfo(author="GridGain Systems",
             url="www.gridgain.com",
             email="support@gridgain.com",
             version="4.3.1e.10112012")
@GridSpiMultipleInstancesSupport(value=true)
public class GridNodeFilterTopologySpi
extends GridSpiAdapter
implements GridTopologySpi, GridNodeFilterTopologySpiMBean

This class provides implementation for topology SPI based on GridPredicate2. The implementation returns nodes that are accepted by GridPredicate2 provided in configuration. If no predicate filter was provided, all nodes, local and remote, will be included into topology.

This topology allows for fine grained node provisioning for grid task execution. Nodes can be filtered based on any parameter available on GridNode. For example, you can filter nodes based on operating system, number of CPU's, available heap memory, average job execution time, current CPU load, any node attribute and about 50 more metrics available in GridNodeMetrics. Here some of the methods on GridNode interface which may be used for filtering:

• GridNode.attributes() - attributes attached to a grid node. Node attributes are specified in grid configuration via GridConfiguration.getUserAttributes() parameter. Note that all system and environment properties are automatically pre-set as node attributes for every node.
• GridNode.metrics() - about 50 node metrics parameters that are periodically updated, such as heap, CPU, job counts, average job execution times, etc...

Apache JEXL Predicate Filter

GridGain also comes with GridJexlPredicate2 implementation which allows you to conveniently filter nodes based on Apache JEXL expression language. Refer to Apache JEXL documentation for specifics of JEXL expression language. GridJexlPredicate2 allows for a fairly simple way to provide complex SLA-based task topology specifications. For example, the configuration examples below show how the SPI can be configured with GridJexlPredicate2 to include all Windows XP nodes with more than one processor or core and that are not loaded over 50%.

Configuration

Mandatory

This SPI has no mandatory configuration parameters.

Optional

This SPI has following optional configuration parameters:

• GridNodeFilterTopologySpi.setFilter(GridPredicate2) - Node predicate filter that should be used for decision to accept node.

Java Example

GridNodeFilterTopologySpi needs to be explicitely configured.

 GridNodeFilterTopologySpi topSpi = new GridNodeFilterTopologySpi();

 GridJexlPredicate2<GridNode, GridTaskSession> filter = new GridJexlPredicate2<GridNode, GridTaskSession>(
     "node.metrics().availableProcessors > 1 && " +
     "node.metrics().averageCpuLoad < 0.5 && " +
     "node.attributes().get('os.name') == 'Windows XP'", "node", "ses");

 // Add filter.
 topSpi.setFilter(filter);

 GridConfigurationAdapter cfg = new GridConfigurationAdapter();

 // Override topology SPI.
 cfg.setTopologySpi(topSpi);

 // Starts grid.
 G.start(cfg);
 

Spring Example

GridNodeFilterTopologySpi can be configured from Spring XML configuration file:

 <bean id="grid.custom.cfg" class="org.gridgain.grid.GridConfigurationAdapter" singleton="true">
       ...
       <property name="topologySpi">
           <bean class="org.gridgain.grid.spi.topology.nodefilter.GridNodeFilterTopologySpi">
               <property name="filter">
                    <bean class="org.gridgain.grid.lang.GridJexlPredicate2">
                        <constructor-arg>
                            <value>
                                <![CDATA[elem1.metrics().availableProcessors > 1 &&
                                elem1.metrics().averageCpuLoad < 0.5 &&
                                elem1.attributes().get('os.name') == 'Windows XP']]>
                            </value>
                        </constructor-arg>
                    </bean>
                </property>
           </bean>
       </property>
       ...
 </bean>
 

Constructor Summary

GridNodeFilterTopologySpi()

Method Summary

GridPredicate2<GridNode,GridTaskSession>	getFilter() Gets node predicate filter for nodes to be included into topology.
Collection<GridNode>	getTopology(GridTaskSession ses, Collection<? extends GridNode> grid) This method is called by GridGain right before calling GridTask.map(List, Object) to obtain a topology for the task's split.
void	setFilter(GridPredicate2<GridNode,GridTaskSession> filter) Sets filter for nodes to be included into task topology.
void	spiStart(String gridName) This method is called to start SPI.
void	spiStop() This method is called to stop SPI.
String	toString()

Methods inherited from class org.gridgain.grid.spi.GridSpiAdapter

assertParameter, checkConfigurationConsistency0, configInfo, createSpiAttributeName, getAuthor, getConsistentAttributeNames, getGridGainHome, getLocalNodeId, getName, getNodeAttributes, getSpiContext, getStartTimestamp, getStartTimestampFormatted, getUpTime, getUpTimeFormatted, getVendorEmail, getVendorUrl, getVersion, injectables, onContextDestroyed, onContextDestroyed0, onContextInitialized, onContextInitialized0, registerMBean, setJson, setName, startInfo, startStopwatch, stopInfo, unregisterMBean

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.gridgain.grid.spi.GridSpi

getName, getNodeAttributes, onContextDestroyed, onContextInitialized

Methods inherited from interface org.gridgain.grid.spi.GridSpiJsonConfigurable

setJson

Methods inherited from interface org.gridgain.grid.spi.GridSpiManagementMBean

getAuthor, getGridGainHome, getLocalNodeId, getName, getStartTimestamp, getStartTimestampFormatted, getUpTime, getUpTimeFormatted, getVendorEmail, getVendorUrl, getVersion

Constructor Detail

GridNodeFilterTopologySpi

public GridNodeFilterTopologySpi()

Method Detail

getFilter

public GridPredicate2<GridNode,GridTaskSession> getFilter()

Gets node predicate filter for nodes to be included into topology.

Specified by:

getFilter in interface GridNodeFilterTopologySpiMBean

Returns:

Node predicate filter for nodes to be included into topology.

setFilter

@GridSpiConfiguration(optional=true)
public void setFilter(GridPredicate2<GridNode,GridTaskSession> filter)

Sets filter for nodes to be included into task topology.

Parameters:

filter - Filter to use.

See Also:

GridJexlPredicate2

spiStart

public void spiStart(@Nullable
                     String gridName)
              throws GridSpiException

This method is called to start SPI. After this method returns successfully kernel assumes that SPI is fully operational.

Specified by:

spiStart in interface GridSpi

Throws:

GridSpiException - Throws in case of any error during SPI start.

Parameters:

gridName - Name of grid instance this SPI is being started for (null for default grid).

spiStop

public void spiStop()
             throws GridSpiException

This method is called to stop SPI. After this method returns kernel assumes that this SPI is finished and all resources acquired by it are released.

Note that this method can be called at any point including during recovery of failed start. It should make no assumptions on what state SPI will be in when this method is called.

Specified by:

spiStop in interface GridSpi

Throws:

GridSpiException - Thrown in case of any error during SPI stop.

getTopology

public Collection<GridNode> getTopology(GridTaskSession ses,
                                        Collection<? extends GridNode> grid)
                                 throws GridSpiException

This method is called by GridGain right before calling GridTask.map(List, Object) to obtain a topology for the task's split.

Specified by:

getTopology in interface GridTopologySpi

Throws:

GridSpiException - Thrown in case if topology cannot be obtained.

Parameters:

ses - Current task's session. If implementation does not depend on task's information it may ignore it.

grid - Full set of all grid nodes.

Returns:

Topology to use for execution of the task represented by the session passed in.

toString

public String toString()

Overrides:

toString in class Object