View on GitHub

RadarGun

Benchmarking framework for data grids and distributed caches

Mapreduce stages

urn:radargun:stages:mapreduce:3.0

map-reduce

Stage which executes a MapReduce Task against all keys in the cache.

collator-fqn (optional) - Fully qualified class name of the org.infinispan.distexec.mapreduce.Collator implementation to execute. The default is null.
collator-params (optional) - A list of key-value pairs in the form of ‘methodName:methodParameter’ that allows invoking a method on the Collator Object. The method must be public and take a String parameter. The default is null.
combiner-fqn (optional) - Fully qualified class name of the org.infinispan.distexec.mapreduce.Reducer implementation to use as a combiner.
combiner-params (optional) - A list of key-value pairs in the form of ‘methodName:methodParameter’ that allows invoking a method on the Reducer Object used as a combiner. The method must be public and take a String parameter. The default is null.
deep-compare-previous-executions (optional) - Compare results of previous executions on entry-by-entry basis. WARNING: This can be lengthy operation on data sets with many distinct keys. On false, only result sizes are checked. The default is true.
exit-on-failure (optional) - If true, then the benchmark stops when the stage returns an error. If false, then the stages in the current scenario are skipped, and the next scenario starts executing. Default is false.
groups (optional) - Specifies in which groups this stage should actively run. The result set is intersection of specified workers, groups and roles. Default is all groups.
mapper-fqn (mandatory) - Fully qualified class name of the mapper implementation to execute.
mapper-params (optional) - A list of key-value pairs in the form of ‘methodName:methodParameter’ that allows invoking a method on the Mapper Object. The method must be public and take a String parameter. The default is null.
num-executions (optional) - The number of times to execute the Map/Reduce task. The default is 10.
print-result (optional) - Boolean value that determines if the final results of the MapReduceTask are written to the log of the first worker node. The default is false.
reducer-fqn (mandatory) - Fully qualified class name of the org.infinispan.distexec.mapreduce.Reducer implementation to execute.
reducer-params (optional) - A list of key-value pairs in the form of ‘methodName:methodParameter’ that allows invoking a method on the Reducer Object. The method must be public and take a String parameter. The default is null.
roles (optional) - Specifies on which workers this stage should actively run, by their roles. The result set is intersection of specified workers, groups and roles. Supported roles are [COORDINATOR]. Default is all roles.
source-name (optional) - Name of the source to execute map-reduce task on. Default value is implementation specific.
store-result (optional) - Boolean value that determines if the final results of the MapReduceTask are stored in the worker state. The default is false.
timeout (optional) - A timeout value for the remote communication that happens during a Map/Reduce task. The default is zero which means to wait forever.
total-bytes-key (optional) - The name of the key in the MainState object that returns the total number of bytes processed by the Map/Reduce task. The default is RandomDataStage.RANDOMDATA_TOTALBYTES_KEY.
workers (optional) - Specifies on which workers this stage should actively run. The result set is intersection of specified workers, groups and roles. Default is all workers.