Interpreter Configuration

For JDBC

Update the below proeprty in jdbc.json underOFS_MMG/mmg-studio/server/builtin/interpreters

"propertyName": "default.url"

"defaultValue": "<JDBC_URL>"

For example: jdbc:oracle:thin:@ofss-mum-1033.snbomprshared1.gbucdsint02bom.oraclevcn.com:15 21/MMG19PDB

"propertyName": "default.user",

"defaultValue": "<schameusername>"

For example: The schema user to which you want to connect, for example: datastudio schema name,

"propertyName": "default.password",

"defaultValue": "<schemapassword>"

For example: Password of the provided schema user.

Start the jdbc interpreter by executing below command under //OFS_MMG/mmg-studio/interpreter-server/jdbc-interpreter-22.4.3/bin ./

jdbc-interpreter

If the jdbc interpreter needs to be included in the datastudio startup script remove the below entry from /OFS_MMG/mmg-studio/bin/startup.sh --jdbc -1.

For Spark

Configuration with Kerberos enabled remote spark cluster:

1. Copy the configured Spark directory from hadoop cluster to <MMG Studio>/interpreter-server/spark-interpreter/extralibs. For example: spark-2.4.8-bin-hadoop2.7
2. Copy the below files to the <MMG Studio>/interpreter-server/spark-interpreter/extralibs krb5.conf <keytabfile>.keytab
3. To run Spark in yarn-client mode, configure the following parameters in this file OFS_MMG/mmg-studio/server/builtin/interpreters/spark.json

   spark.master = yarn-client

   spark.driver.bindAddress = 0.0.0.0

   spark.driver.host = <host> -> Apache Spark host

   Note:

   When using the Kubernetes interpreter lifecycle, <host> can be the IP address or hostname of any node in your Kubernetes cluster. When using the Host interpreter lifecycle, <host> should be the IP address or hostname of the node that runs the Spark interpreter.

   Note:

   When connecting to a YARN cluster, the Spark driver authenticates as the UNIX user that runs the Spark interpreter. You can set the HADOOP_USER_NAME environment variable to make the Spark driver authenticate as a different user. If you use the Host interpreter lifecycle, then you can do this by exporting the HADOOP_USER_NAME environment variable before starting the Spark interpreter process. If you us the Kubernetes interpreter lifecycle, then you can do this by setting the HADOOP_USER_NAME environment variable in the resource manifest (spark.yml).
4. Update file spark-defaults.conf keytab location to the location where <keytabfile>.keytab file is copied
5. Update file spark-env.sh with the krb5.conf location to the location where krb5.conf file is copied.

   For example: Djava.security.krb5.conf=/OFS_MMG/mmg-studio/interpreter-server/spark-inte rpreter-22.4.2/extralibs/krb5.conf".