health checks for orderer unreachable from kubelet: causing pod to crash

[gWOLF3]

health checks on the orderer are unreachable on the kubelet.

expected: should be reachable from kubelet. same health checks are working fine on peer.

whats happening?

health check is working fine when inside container but unreachable from kubelet. tested on both minikube and gke.

how to replicate: check out healthcheckv2 branch. this functionality was added in 1.4.1 fabric release/

[gWOLF3]

this issue is FIXED. plus learned some really useful shit:

1. port 8443 indeed does have conflicts with minikube api server. i tried changing with minikube start --apiserver-port <someotherport> command, but was still having trouble. this is not the best way anyhow.

2. better way: change the address that healthchecks are exposed on in values file. this was tricky because i was using the wrong env variable, namely PEER_OPERATIONS_LISTENADDRESS - after a helpful discussion on rocket chat, I learned that these variables are leveraging viper's template engine - and the correct way to specify orderer is with ORDERER_OPERATIONS_LISTENADDRESS. anyways - more info on this - check out my convo on hyperledger rocket chat with user btl5037.