All, I'm attempting to setup a Load Balancer (LB) in front of my NiFi cluster using another NiFi server. I've seen a successful example used online and it was very straight forward. 1. Receive data on LB 2. And then send it to the cluster using RoundRobin
Pretty straight forward. FYI, here is my setup: All machines are CentOS7 I'm using NiFi version 1.12.1 on LB and Cluster LB's PutTCP: - Max Size of Socket Send Buffer: 10MB Cluster ListenTCP and ListenSyslog: - Receive Buffer Size: 128 MB - Max Size of Message Queue: 1,000,000 - Max Size of Socket Buffer: 100 MB On my Load Balancer, I'm using Round Robin and sending my TCP data to my NiFi Cluster. I understand this isn't the best solution for the long term, and it's not easily scalable. Regardless, I'm sending the TCP data to my NiFi Cluster using PutTCP. On my NiFi Cluster, I'm trying to receive this TCP traffic (sent on a specific port), but nothing is coming through on my ListenTCP (same port). This confused me, and my first assumption was I'm not receiving TCP data at all. After running a tcpdump, I can confirm I'm receiving TCP data over the specified port from my Load Balancer. I even ran tcpdump with the -A flag so I can see the actual data. So it's not empty TCP traffic (even though I can see the length of the data is greater than 0). So, because I can see the TCP traffic on my NiFi nodes, why wouldn't I be able to receive the data using ListenTCP? I'm even using ListenSyslog to see if that would work. Nothing. Thoughts? I really appreciate any time you put into this. Thank you in advance.
