I am very interested with porting one of our large-scale machine learning systems to Yarn. As this system is written in Go, I think I would rely on github.com/hortonworks/gohadoop.
However, it seems that gohadoop requires me to patch Hadoop source code, rebuild and redeploy. But I have no such kind of control to Hadoop clusters in our company.
Is the purpose of patch to disable authentication in RPC? If so, could this be done via editing the configuration file instead of patching the source code?
Support from the Experts
A HDP Support Subscription connects you experts with deep experience running Apache Hadoop in production, at-scale on the most demanding workloads.
Become HDP Certified
Real world training designed by the core architects of Hadoop. Scenario-based training courses are available in-classroom or online from anywhere in the world