Github user vtslab commented on the issue: https://github.com/apache/tinkerpop/pull/721 Could I assume a pseudo-hadoop cluster present during the integration test phase? I thought only the asciidoc processing had that. Anyway, breaking the spark-yarn option will be noticed through the docs processing, but hopefully not until hadoop-3 or spark-3. I´ll correct the TinkerPop naming, off course, and will also add a pointer to the gremlin-plugin-dependencies section of the spark-gremlin manifest file.
---