How To Run Graph In Endeca Outside of Endeca Server
Posted by sranka on January 15, 2014
Recently at one of the client, we had a situation , where in Hadoop was taking lot longer that anticipated time to generate a file. The graph needed the file as an input, but since file was not getting generated on time, Endeca graph was picking up partially created file, causing data issue. After looking into the issue, the best bet was to have a task dependency, we looked into running clover ETL from command line, but due to some product limitation, we were not able to run the graph from command line.
After discussing with Chris Lynskey from Oracle (original Endeca team) , I found that following simpleHttpApi could work:
Description: Text Id, which is unique in specified sandbox. File path relative to sandbox root .
e.g. graph%2FLoadViewDefinitions.grf , the “/” needs to encoded to “%2F”
Description: Sandbox code
In cluster mode it’s ID of node which should execute the job. However it’s not final. If the graph is distributed, or the node is disconnected, the graph may be executed on some another node.
Description: MESSAGE | FULL
For more HTTP API please refere to Latitude Data Integrator Server Guide .
Hope this helps
Sunil S Ranka
“Superior BI is the antidote to Business Failure”