502 bad gateway with nginx in network calls

Question

I set up load balancing on NGINX using the Round Robin for apache tomcat servers with two servers in my proxy.conf file:

upstream appcluster1 {
    server IP_ADDRESS1:8000;
    server IP_ADDRESS2:8001;
}

server {
    location / {
        proxy_pass http://appcluster1;
    }

}

For more details, I am running Geoserver with each of the tomcat server. For the first IP_ADDRESS1, I am successfully able to make calls to it and get layers returned (say if geoserver on the IP_ADDRESS2 is down). However, if I switch it and try to make calls to IP_ADDRESS2 (and IP_ADDRESS1 is down), I am unable to get all of the tiles back from Geoserver. In the network calls, I get a mixture of 200 and 502 errors.

The two geoserver is exactly the same and not any different from one another. Does anyone know why this is the case -- one Geoserver is able to return calls successfull but the other one only gets partial results back. I am doing all of this using NGINX and the round robin method..

I also tried to set proxy_next_upstream to off but that didn't seem to work. Any help will be greatly appreciated!

Is there a Tomcat instance running at `IP_ADDRESS2:8001`? Try connecting to this Tomcat directly. If both Tomcat's are configured in the same way, isn't the second one also listening on port `8000`? — Piotr P. Karwasz, Feb 02 '21 at 09:13
@IanTurton I don't see anything particular in the Geoserver logs.. perhaps it has to do with the way nginx is doing the round robin method — fairlyMinty, Feb 04 '21 at 18:28
If the requests don't get to GeoServer then it's an nginx issue — Ian Turton, Feb 05 '21 at 10:26

502 bad gateway with nginx in network calls

0 Answers0