Repository files navigation

CPU Mining cryptocurrencies using Apache Spark without any external dependencies.

UsingCPU Miner fromhttps://github.com/pooler/cpuminer in an Apache Spark job that starts tasks on a Yarn/Mesos cluster. Each cluster executor starts an instance of CPU Miner using the provided mining options.

Get yourself an account in any mining pool that supports CPU mining :

Keep an eye onhttps://coinmarketcap.com/ and look for new coins that do not require special mining equipments.

Package the project into a single jar (requiresSBT), in the project root run :

sbt package

You'll get aspark-yarn-miner_2.11-*.jar jar.

Next, usespark-submit to submit your job to the cluster :

spark-submit --master yarn --class fr.marouni.spark.coins.SparkYarnMiner spark-yarn-miner_2.11-0.1.0-SNAPSHOT.jar --url XXX --username YYY --worker-id WWW --password UUUU

Where :

• --url : Your mining pool URL (check mining pool docs). Example : stratum+tcp://stratum.aikapool.com:7915
• --username : Mining pool account username
• --worker-id : Mining pool worker id (check mining pool docs)
• --password : Mining pool account password

Tune your mining with the followingspark-submit options :

• spark-submit --conf spark.executor.instances=10 on a 10 nodes cluster
• spark-submit --conf spark.executor.cores=6 on a cluster with 6 cores per node

Recommended tuning : 1 executor per cluster node with maximum allowed number of cores per executor.