I am trying setup a google cloud dataflow pipeline (streaming mode) that read pubsub topic message, extract information (object name in google cloud storage) from published message, then start another pipeline (batch mode) to process the object stored in google cloud storage.
Is it possible to start another pipeline within a pipeline???
There is no technical reason barring this. You would need to be sure to keep your Pipeline objects separate, have sufficient Compute Engine quota to launch all the jobs you need.