streaming.core.message.MLSQLMessage.scala Maven / Gradle / Ivy
The newest version!
package streaming.core.message
import org.apache.spark.MLSQLConf
/**
* 2019-01-13 WilliamZhu([email protected])
*/
object MLSQLMessage {
val MLSQL_CLUSTER_PS_ENABLE_NOTICE = s"""
|------------------------------------------------------------------------
|MLSQL detects that ${MLSQLConf.MLSQL_CLUSTER_PS_ENABLE.key} is enabled.
|
|Please make sure you have the uber-jar of mlsql placed in
|1. --jars
|2. --conf "spark.executor.extraClassPath=[your jar name in jars]"
|
|for exmaple:
|
|--jars ./streamingpro-mlsql-spark_2.x-x.x.x-SNAPSHOT.jar
|--conf "spark.executor.extraClassPath=streamingpro-mlsql-spark_2.x-x.x.x-SNAPSHOT.jar"
|
|Notice that when you configure `spark.executor.extraClassPath`,please just assign the jar name
|without any path prefix.
|
|Otherwise the executor will fail to start and the whole application will fails.
|------------------------------------------------------------------------
""".stripMargin
val PYTHON_REQUEIRE_MLSQL_CLUSTER_PS_ENABLE = s"""
|[Warning]
|
|MLSQL detects that you are not in local mode,
|but ${MLSQLConf.MLSQL_CLUSTER_PS_ENABLE.key} is not set true.
|
|PythonAlg and PythonParallelExt will fail if you are in cluster mode and
|without enabling ${MLSQLConf.MLSQL_CLUSTER_PS_ENABLE.key}
|
""".stripMargin
}