We need to set driver as well as executor class path in cluster mode.   Anyways your analysis is quite helpful and we have also used the same for now. Thanks. Keep sharing! ... View more

Hi jeremyw, I am using the same versions of Tika and getting same dependency conflicts.   There are few observations: 1. Dependency tree comparison shows many differences and not only org.apache.poi.ooxml as mentioned by you. 2. Excluding org.apache.poi.ooxml dependency from Tika-parsers, returns blank string after extraction.   Can you please reply if it actually worked in your case or you faced any more issues? ... View more