Getting started with a Pentaho Proof of Concept
I am brand new to trying using Pentaho but have experience with Datamart design and implementation in the non open-source world (MS Sql-Server, Analysis Services, DTS, Informatica Powe Anlyzer)
I am looking a the best path for information on how to get a quick proof of concept up and running in a Linux environment. I am not a Linux system admin but have that resource at my disposal. So far I have gotten a PostgreSQL database setup on the Linux server as well the the Pentaho Demo. I am able to execute the Pentaho Demo successfully.
I am a lttile un clear on where certain development tools reside in the Penatho architecture:
1) If I have the Pentaho demo installed on the Linux server do I need to install the individual compoents as well in order to do my own proof of concept ? Or is Mondrian, report designer, etc. part of the demo installation
2) What tools in the Pentaho suite reside on the Linux server and what components have only client side/developer applications, what have both types of components.. I am thinking of the tools like Kettle (ETL tool), report designer, dashborad builder, etc.
3) I envison the devolpment side of the BI application to taking place in a Windows PC environment or Web based... and just the database / web pages being served form the Linux server.
4) How to connect the Pentaho suite applications to the PostgreSQL database
Can anybody provide me with or point me to a roadmap on how to accomplish a small proof of concept. I don;t need to do anything too fancy. Just be able to demonstarte the OLAP querying, report generation and dashboards with data that is pertinent to our company. SOmething similar to the Pentaho demo, but with my own data would be good.
Thanks in advance,
1) Most people use the demo (Pre-configured install) for their proof of concepts and many times for their final solutions. It is the full blown patform. The demo part refers to the samples.
The demo server comes with all of the engines installed - Mondrian, Kettle, JFreeReport etc. To create new content, you will need to download the client tools for the type of content you are interested in; Report Designer, Cube Designer, Design Studio, Spoon (for Kettle) in order to create new reports, cubes, transforms etc.
2) All of the tools basically create XML files that are used by the engines. They can run on any machine and you can deploy them to the server manually if you want. Most of the tools have a publish feature that helps you with this. As long as we are on the topic - You don't actually need the tools if you are crazy enough to hand edit the XML. I don't recommend this
3) The tools are designed to let you do this, that is the publish
4) There are several posts in the forums for working with Postgres. This link may also help: http://jira.pentaho.org:8080/conflue...ho+BI+Platform
The creating solutions guide should help you get started: http://jira.pentaho.org:8080/conflue...taho+Solutions
Other platform doc:
If I look at what I need to install for the ETL portion for example, this is a 70M download.
Originally Posted by dmoran
Are you saying that this includes the ETL engine and Client portions.
If so, how can I get around not having to download the entire 70 M.
To download the full suite of products, I'm looking to download over a gig of data, if I include the Java SDK and Eclipse.
For me sitting in a remote part of Africa, that's a LOT of bandwidth. (In fact, it's my monthly broadband allocation!).
To continue this dicussion, should I downlading the components on the download page as is, or are there deployements that I could integrate into the existing JBOSS/PCI deployment?