Skip to content

Latest commit

 

History

History
47 lines (31 loc) · 2.73 KB

README.md

File metadata and controls

47 lines (31 loc) · 2.73 KB
Raw

equoid-data-handler

Count instances of data received from AMQP broker

Quickstart on cluster

DataHandler class

  1. Start your cluster oc cluster up

  2. Start app

oc new-app --template=oshinko-java-spark-build-dc -p APPLICATION_NAME=equoid-data-handler -P GIT_REPO_URI=https://github.com/eldritchjs/equoid-data-handler -p APP_MAIN_CLASS=io.radanalytics.equoid.dataHandler -p APP_ARGS='<AMQP Broker Address> <AMQP Port> <AMQP Username> <AMQP Password>  <AMQP Address> <Infinispan Address>  <Infinispan Port> <k> <epsilon> <confidence>' -p APP_FILE=equoid-data-handler-1.0-SNAPSHOT.jar -p SPARK_OPTIONS='--driver-java-options=-Dvertx.cacheDirBase=/tmp'
  1. Publish some data to queue salesq per https://github.com/eldritchjs/equoid-data-publisher

CheckCache test class

  1. Start test
oc new-app --template=oshinko-java-spark-build-dc -p APPLICATION_NAME=equoid-check-cache -P GIT_REPO_URI=https://github.com/eldritchjs/equoid-data-handler -p APP_MAIN_CLASS=io.radanalytics.equoid.checkCache -p APP_ARGS='<Infinispan Address> <Infinispan Port> <Key> <Iterations>' -p APP_FILE=equoid-data-handler-1.0-SNAPSHOT.jar
  1. Check pod log for output either via web console or the following, as data is published to equoid, the value should increase accordingly:
oc logs <Equoid-check-cache pod ID> -n equoid

Quickstart locally

A facility for testing/trying out the DataHandler capabilities without an OpenShift cluster is provided, as well. The class FileDataHandler incorporates textFileStream(<data dir>) to check for files in data dir periodically. This is strictly for testing without AMQ-P and OpenShift and is not meant to be a replacement for the relevant components of the Equoid architecture. The testing can take place as follows.

  1. Build the classes sbt assembly

  2. Recommended: clean out checkpoint area and data files, see for example src/test/bash/cleanlocal.sh

  3. Start an Infinispan server on localhost via Infinispan's provided bin/standalone.sh (easiest for testing functionality in the case where an extant server is not available)

  4. Start FileDataHandler (assumes $SPARK_HOME is set properly) see for example src/test/bash/startlocal.sh $SPARK_HOME/bin/spark-submit --class io.radanalytics.equoid.FileDataHandler --master local[4] target/scala-2.11/equoid-data-handler-assembly-0.1.0-SNAPSHOT.jar

  5. Start copying files to the data directory (default is /tmp/datafiles/) src/test/bash/cpdata.sh <Data File> <Output Directory> <Window Size> <Update Interval>

  6. Optional: start CheckCache in order to view contents of Infinispan as data is received $SPARK_HOME/bin/spark-submit --class io.radanalytics.equoid.CheckCache --master local[4] target/scala-2.11/equoid-data-handler-assembly-0.1.0-SNAPSHOT.jar