Mr.x

@spacewander Your test is correct. When I change the hostname to IP, it is normal, but why can't I use the hostname?

@tokers I change the log level to warn, The log I see is like this: ``` 2021/08/12 09:14:38 [warn] 6521#6521: *220612 [lua] healthcheck.lua:1107: log(): [healthcheck] (upstream#/upstreams/1) unhealthy HTTP increment (1/2)...

@spacewander Yes. This phenomenon occurs when the IP(upstream) is changed to the host name. You can test it with the following configuration ``` # apisix.yaml upstreams: - id: 1 name:...

@tokers There are four hosts in my test environment, the actual configuration of the proxy node is like this ``` # apisix.yaml upstreams: - id: 1 name: "nginx" nodes: "nginx1:8090":...

@tokers The path for health check is /active(return 500), and the path for client access is /(return 200) ``` 2021/08/12 09:15:18 [warn] 6508#6508: *234044 [lua] healthcheck.lua:1107: log(): [healthcheck] (upstream#/upstreams/1) healthy...

@tokers ``` $ curl nginx1:8090/active 500 Internal Server Error 500 Internal Server Error openresty/1.15.8.3 $ curl nginx2:8090/active hello, world $ curl nginx3:8090/active hello, world ```

@tzssangglass The access logs of nginx1, nginx2 and nginx3 are like this ``` # nginx1 $ tail -f logs/access.log 192.168.2.11 - - [16/Aug/2021:09:10:33 +0800] "GET /active HTTP/1.0" 500 183 "-"...

> @kuberxy could you please try v0.31.0 (which was released today) and let us know if the issue is gone for you? this issue is not gone. ```shell $ kubectl...