View Full Version : Scheduling options on a cluster
03-28-2012, 06:55 PM
New to Pentaho, but very interested in the potential.
I had a question regarding to scheduling on a Hadoop cluster - what are my options?
In particular, I would be interested in any scenarios for scheduling Pentaho ETL jobs using Oozie.
03-29-2012, 05:06 PM
Pentaho ETL jobs can be executed from the command line so it should be easy to schedule them to run from oozie or any scheduler.
We use "pan" to run transforms:
We use "kitchen" to run jobs:
Some advanced concepts for scheduling:
You can also use the Pentaho BI Server to schedule ETL jobs which have been configured to run on the cluster. I am currently working on the roadmap for Kettle as it related to Big Data and we have been wanting to look into Oozie integration but haven't heard much from people with real use cases. I am very interested in your specific use case and what your "wish list" for Oozie integration is. Are you already using Oozie and wanting to orchestrate from there? Are you assuming you nee to use it but are open to other technology?