Hello,
I’m looking for a solution to use jruby on top of spark. The only tricky point is that I need that every worker thread has a ruby interpreter initialized. Basically, I need to register a function to be called when each worker thread is created : a thread local variable must be set for the ruby interpreter so that ruby object can be deserialized. Is there any solution to setup the worker threads before any spark call is made using this thread ? Regards, Lionel