Connection to Hive/BigData through Knowage docker image

Question

Hello,

I'm new to Knowage and I'm experimenting to connect Knowage to my Spark/Hadoop env, but I'm having a lot of difficulty configuring the correct jars for the Hive2 JDBC connector. I'm using the knowage:8.0 docker image, retrieved from docker hub, and I noticed that in $(TOMCAT_HOME)/webapps/knowage/WEB-INF/lib there aren't the connectors to Hive/Hive2/Spark, so I supposed to manually put them there, but It's being really struggle...
My question, is it possible to connect this version of Knowage to any BigData source? I read this on documentation, in the section about Big Data and NoSQL (https://knowage-suite.readthedocs.io/en/master/administrator-guide/configure-data-sources.html):
"Please note that these connections are available for products KnowageBD and KnowagePM only."

Answer 1

Hi,

as you can see from the Dockerfile our official image contains drivers for MySql, MariaDB and Postgres. We saw that those three are the most used drivers in our customers' environments.

If I remember well the JDBC driver for Hive/Hive2/Spark is composed by many JARs, then I understand your struggle.

If you want a really fast fix for this problem I suggest you to create a custom container image where the FROM clause specifies our official image: there you can add all the JARs you need.

From our side, we can think to a directory exposed as a volume where to put all the JARs needed by a certain installation but we surely need some tests and some time.

Comments

Thanks, I also created our custom dockerfile and COPY the required jars for hive-jdbc, but the problem still exists (using another Java application with the same jars, I could connect to Hive and do queries). Finally I resolved editing the class DelegatingConnection of another package (commons-dbcp2-2.2.0.jar) that always launched an SQLException like "Method not supported". After editing and recompiling the package, I could connect to Hive from Knowage. It's really a bad solution, since I had to edit a module from org.apache and customize it.