PDA

View Full Version : Pentaho CDE with Cloudra Impala (on top of Hadoop)



šmrc
07-08-2013, 04:52 AM
Hi,

First of all I don't know in which part of forum to put my question. I have a problem with Pentaho CDE connection towards Cloudera Impala which offers near to real time query response on top of Hadoop. The thing is that I want to develop dashboards on Hadoop Hive data. We started of without Impala, just on Hadoop Hive but the performance was very poor (queries lasted couple of minutes, but the amount of data is negligible). Soooo we went to search the vast wilderness of internet :) to find what could speed up our reports, and bumped into Cloudera Impala and decided to give it a try.
The reports now last couple of seconds ;) but we had problems with parameters because the syntax ${} just doesn't go :(. When I hard code the values, it works like a charm (this is another problem that's bugging me, but ignore this currently).
I added several jar drivers to /pentaho/server/biserver-ee/tomcat/lib/ which were added before to the Report designer (a colleague of mine uses these for same combination just for Kettle) . And when I try to create a CDC dashboard no matter which component I use i get error processing component: The data source connection is: 11294

The query has no parameters and I know it should work because I copied it from Report Designer where it works. I have a hunch it's a driver thing but then again I am no system admin. When I first saw the problem I was like :confused:.... The Pentaho log is: 11295

Any help,ideas?