Starting Services on Riak Data Platform

You’ve installed Riak Data Platform (BDP), configured your cluster, and added services to your nodes. The setup of your BDP cluster is complete! Now you can begin using your BDP cluster.

Start Services

The very first thing you can do with your BDP cluster is start the services you added. In the last section of the configuration instructions, you added the following services:

  • Spark-master
  • Spark-worker
  • Redis
  • Cache proxy

Spark-Master Service

To start the spark-master service, run the following, using the name and IP of the node you wish to start the service on:

sudo data-platform-admin start-service »NODENAME«@»IPADDRESS« my-spark-group my-spark-master

Then verify that the spark-master service is running:

ps -ef | grep [s]park-master

You should see the output like this:

riak     23418     1  7 20:32 ?        00:00:13 /usr/local/jdk/bin/java -cp /usr/lib/riak/lib/data_platform-1/priv/spark-master/sbin/../conf/:/usr/lib/riak/lib/data_platform-1/priv/spark-master/lib/spark-assembly-1.4.0-hadoop2.6.0.jar:/usr/lib/riak/lib/data_platform-1/priv/spark-master/lib/datanucleus-api-jdo-3.2.6.jar:/usr/lib/riak/lib/data_platform-1/priv/spark-master/lib/datanucleus-rdbms-3.2.9.jar:/usr/lib/riak/lib/data_platform-1/priv/spark-master/lib/datanucleus-core-3.2.10.jar -Dspark.deploy.recoveryMode=CUSTOM -Dspark.deploy.recoveryMode.factory=org.apache.spark.deploy.master.RiakEnsembleRecoveryModeFactory -Dspark.deploy.leadelect.namespace=spark_leaders -Dspark.deploy.leadelect.service=localhost:10012 -Dspark.deploy.riak.metadata.map=spark-cluster-map -Dspark.deploy.riak.consistent.bucket=spark-bucket -Dspark.deploy.leadelect.timeout=10000 -Dspark.riak.connection.host=localhost:10017 -cp /usr/lib/riak/lib/data_platform-1/priv/spark-master/sbin/../lib/* -Xms512m -Xmx512m org.apache.spark.deploy.master.Master --ip 172.28.128.3 --port 7077 --webui-port 8080

If you need to stop the spark-master, the proper sequence is:

sudo data-platform-admin stop-service »NODENAME«@»IPADDRESS« my-spark-group my-spark-master

Spark-Worker Service

To start the spark-worker service, run the following, using the name and IP of the node you wish to start the service on:

sudo data-platform-admin start-service »NODENAME«@»IPADDRESS« my-spark-group my-spark-worker

You can repeat this command (using a different nodename and IP address) to add additional spark-worker services.

To stop a spark-worker service, you need to stop the service first:

sudo data-platform-admin stop-service »NODENAME«@»IPADDRESS« my-spark-group my-spark-worker

Redis Service

To start the Redis service, run the following, using the name and IP of the node you wish to start the service on:

sudo data-platform-admin start-service »NODENAME«@»IPADDRESS« my-redis-group my-redis

Cache Proxy Service

To start the cache proxy service, run:

sudo data-platform-admin start-service »NODENAME«@»IPADDRESS« my-cache-proxy-group my-cache-proxy

Verify The Services Are Running Correctly

Redis and Cache Proxy

Do you use CentOS or Ubuntu?

CentOS

Set the base path of BDP:

export BDP_HOME=/usr/lib64/riak/lib/data_platform-1

Ubuntu

Set the base path of BDP:

export BDP_HOME=/usr/lib/riak/lib/data_platform-1

Verification

First, add the Redis CLI utility to your PATH and specify Ips and ports for the services involved in the test:

export PATH="$PATH:$BDP_HOME/priv/redis/bin"
export RIAK_IP=»RIAK_IP
export RIAK_HTTP_PORT=»RIAK_HTTP_PORT«
export CACHE_PROXY_IP=»CACHE_PROXY_IP«
export CACHE_PROXY_PORT=»CACHE_PROXY_PORT«
export REDIS_IP=»REDIS_IP«
export REDIS_PORT=»REDIS_PORT«

Then, add some test data to your Riak cluster, using the IP:HttpPort of your Riak KV node:

curl -s -X PUT -d "my-value" "http://$RIAK_IP:$RIAK_HTTP_PORT/buckets/my-bucket/keys/my-key"

Retrieve the test data via cache proxy, using the IP address you set for cache proxy:

redis-cli -h $CACHE_PROXY_IP -p $CACHE_PROXY_PORT get my-bucket:my-key

Finally, confirm that the data is now also in Redis:

redis-cli $REDIS_IP -p $REDIS_PORT get my-bucket:my-key