Folks,
I have raised this issue earlier regarding how to connect to Hortonworks Hadoop cluster using BODS.
Our environment is BODS 4.2 SP6 installed on Windows server and Hoortonworks Hadoop is 2.4.2 installed on Red Hat Linux server. I have seen many posts how to connect and do ETL using BODS to Hadoop where most of them are referring to the BODS server on Linux and Hadoop on Linux which is not the case at our customer scenario. Most of the posts say install the job server on hadoop cluster or install seperately but they are pure Linux servers. What if your BODS server is on Windows and Hadoop on Linux.
Here are 2 different solution;
Using Hortonworks Hive ODBC drivers
1. Download and install the Hortonworks ODBC Hive drivers on BODS server and configure the ODBC datasource on the BODS server to connect to Hadoop Hive.
2. Create Data Store using the ODBC connection; import the tables and load or read. Only drawback is that one can't delete data from the imported Hive tables or can't create Hive tables from BODS designer.
Using Hiveadapter configuration on management console:
I did try this configuring on DS Management console from copying Hortonworks Jar files from Linux server to windows box and located them on MC. It worked and configured the datastore using adapter but failed at some stage and we are still investigating.