Hyeonmin-Ha / onyx

Onyx: A Flexible Data Processing System

Onyx

Onyx prerequisites and setup

Prerequisites

• Java 8
• Maven
• Latest REEF snapshot
• YARN settings
  • Download Hadoop 2.7.4 at http://apache.tt.co.kr/hadoop/common/hadoop-2.7.4/
  • Set the shell profile as following:

    export HADOOP_HOME=/path/to/hadoop-2.7.4
    export YARN_HOME=$HADOOP_HOME
    export PATH=$PATH:$HADOOP_HOME/bin

• Protobuf 2.5.0
  • Downloadable at https://github.com/google/protobuf/releases/tag/v2.5.0
  • On Ubuntu:
  1. Run sudo apt-get install autoconf automake libtool curl make g++ unzip
  2. Extract the downloaded tarball and run
     • sudo ./configure
     • sudo make
     • sudo make check
     • sudo make install
  3. To check for a successful installation of version 2.5.0, run protoc --version

Installing Onyx

• Run all tests and install: mvn clean install -T 2C
• Run only unit tests and install: mvn clean install -DskipITs -T 2C

Running Beam applications

Running an external Beam application

• Use run_external_app.sh instead of run.sh
• Set the first argument the path to the external Beam application jar

./bin/run_external_app.sh \
    `pwd`/onyx_app/target/bd17f-1.0-SNAPSHOT.jar \
    -job_id mapreduce \
    -executor_json `pwd`/tests/src/main/resources/sample_executor_resources.json \
    -user_main MapReduce \
    -user_args "`pwd`/mr_input_data `pwd`/onyx_output/output_data"

Configurable options

• -job_id: ID of the Beam job
• -user_main: Canonical name of the Beam application
• -user_args: Arguments that the Beam application accepts
• -optimization_policy: Canonical name of the optimization policy to apply to a job DAG in Onyx Compiler
• -deploy_mode: yarn is supported(default value is local)

Examples

## MapReduce example
./bin/run.sh \
	-job_id mr_default \
	-executor_json `pwd`/tests/src/main/resources/sample_executor_resources.json \
	-optimization_policy edu.snu.onyx.compiler.optimizer.policy.DefaultPolicy \
	-user_main edu.snu.onyx.examples.beam.MapReduce \
	-user_args "`pwd`/examples/src/main/resources/sample_input_mr `pwd`/examples/src/main/resources/sample_output_mr"

## YARN cluster example
./bin/run.sh \
	-deploy_mode yarn \
  	-job_id mr_pado \
	-executor_json `pwd`/tests/src/main/resources/sample_executor_resources.json \
  	-user_main edu.snu.onyx.examples.beam.MapReduce \
  	-optimization_policy edu.snu.onyx.compiler.optimizer.policy.PadoPolicy \
  	-user_args "hdfs://v-m:9000/sample_input_mr hdfs://v-m:9000/sample_output_mr"

Resource Configuration

-executor_json command line option can be used to provide a path to the JSON file that describes resource configuration for executors. Its default value is config/default.json, which initializes one of each Transient, Reserved, and Compute executor, each of which has one core and 1024MB memory.

Configurable options

• num (optional): Number of containers. Default value is 1
• type: Three container types are supported:
  • Transient : Containers that store eviction-prone resources. When batch jobs use idle resources in Transient containers, they can be arbitrarily evicted when latency-critical jobs attempt to use the resources.
  • Reserved : Containers that store eviction-free resources. Reserved containers are used to reliably store intermediate data which have high eviction cost.
  • Compute : Containers that are mainly used for computation.
• memory_mb: Memory size in MB
• capacity: Number of TaskGroups that can be run in an executor. Set this value to be the same as the number of CPU cores of the container.

Examples

[
  {
    "num": 12,
    "type": "Transient",
    "memory_mb": 1024,
    "capacity": 4
  },
  {
    "type": "Reserved",
    "memory_mb": 1024,
    "capacity": 2
  }
]

This example configuration specifies

• 12 transient containers with 4 cores and 1024MB memory each
• 1 reserved container with 2 cores and 1024MB memory

Monitoring your job using web UI

Onyx Compiler and Engine can store JSON representation of intermediate DAGs.

• -dag_dir command line option is used to specify the directory where the JSON files are stored. The default directory is ./dag. Using our online visualizer, you can easily visualize a DAG. Just drop the JSON file of the DAG as an input to it.

Examples

./bin/run.sh \
	-job_id als \
	-executor_json `pwd`/tests/src/main/resources/sample_executor_resources.json \
  	-user_main edu.snu.onyx.examples.beam.AlternatingLeastSquare \
  	-optimization_policy edu.snu.onyx.compiler.optimizer.policy.PadoPolicy \
  	-dag_dir "./dag/als" \
  	-user_args "`pwd`/examples/src/main/resources/sample_input_als 10 3"