Hi to all!First of all, thanks to the NiFi Community for all the
contributions and documentation!We have a five node cluster, with Nifi
1.12.0, where there are running several Uses Cases. Most of them do the same
things, get some files from a server, make some transformations (thanks
QueryRecord!) and then put the records into a Hive table with
PutHive3Streaming.We have a couples of flows design doubts:1. What's better,
¿To have just one PutHive3Streaming processor for ingesting all the use
cases, or one per each use case? We are using the second option, because we
think it's better for debugging purposses, but we are not sure about its
impact in perfoming.2. ¿Whats the best way for managing the failure queue of
the PutHive3Streaming? We have that queue redirected again to the processor,
but sometimes when it fails (e.g. failure conectivity with Hive ) it blocks
the processor and making backpressure in the rest of queues, even inducing
the disconnection of a node.Thanks in advance!Guillermo
--
Sent from: http://apache-nifi-users-list.2361937.n4.nabble.com/