Details
Description
I am getting a java.sql.SQLException: No suitable driver found for jdbc:mysql://dbhost/test when using df.write.jdbc.
I do not get this error when reading from the same database.
This simple script can repeat the problem.
First one must create a database called test with a table called table1 and insert some rows in it. The user test:secret must have read/write permissions.
testJDBC.scala:
import java.util.Properties
import org.apache.spark.sql.Row
import java.sql.Struct
import org.apache.spark.sql.types.{StructField, StructType, IntegerType, StringType}
import org.apache.spark.{SparkConf, SparkContext}
import org.apache.spark.sql.SQLContext
val properties = new Properties()
properties.setProperty("user", "test")
properties.setProperty("password", "secret")
val readTable = sqlContext.read.jdbc("jdbc:mysql://dbhost/test", "table1", properties)
print(readTable.show())
val rows = sc.parallelize(List(Row(1, "write"), Row(2, "me")))
val writeTable = sqlContext.createDataFrame(rows, StructType(List(StructField("id", IntegerType), StructField("name", StringType))))
writeTable.write.jdbc("jdbc:mysql://dbhost/test", "table2", properties)}}
This is run using:
spark-shell --conf spark.executor.extraClassPath=/path/to/mysql-connector-java-5.1.35-bin.jar --driver-class-path /path/to/mysql-connector-java-5.1.35-bin.jar --jars /path/to/mysql-connector-java-5.1.35-bin.jar -i:testJDBC.scala
The read works fine and will print the rows in the table. The write fails with java.sql.SQLException: No suitable driver found for jdbc:mysql://dbhost/test. The new table is successfully created but it is empty.
I have tested this on a Mesos cluster with Spark 1.4.0 and the current master branch as of June 18th.
In the executor logs I do see before the error:
INFO Utils: Fetching http://146.203.54.236:50624/jars/mysql-connector-java-5.1.35-bin.jar
INFO Executor: Adding file:/tmp/mesos/slaves/.../mysql-connector-java-5.1.35-bin.jar to class loader
A workaround is to add the mysql-connector-java-5.1.35-bin.jar to the same location on each executor node as defined in spark.executor.extraClassPath.
Attachments
Issue Links
- is duplicated by
-
SPARK-9985 DataFrameWriter jdbc method ignore options that have been set
-
- Resolved
-
- links to
