After more analysis for the log, it appears the issue is likely in Jetty as
the stacktrace shows the relevant threads during the 30 seconds timeout is
sitting in sun's EpollSelector in native code.
I've dug as deep as I can. It can be reliably recreated in nifi-1.8.0 and
in nifi-1.9.2 in my
Alright, round two occurred today, and after much frustration, I believe I
know what exactly is causing the problem, and the solution is probably just
upgrade to nifi-1.9.2 which comes with an upgraded jetty.
Java 8u202 - Nifi 1.8.0 - Jetty 9.4.11 (same Jetty version is in Nifi 1.9.2)
BLUF, a
I have a new 3 node cluster i'm trying to stand up.
On most of the API calls (to create proxy the proxy user) and others for
debugging, i get a "java.net.SocketTimeoutException: timeout" returned. In
debugging the issue, the timeout occurs when the API replicates the request
to the same instance