I am running various Druid processes as a container on a CentOS machine + DC/OS 1.9. I find that HTTP requests to coordinator endpoints hangs. Querying from within the container with localhost also doesn’t help. I am not observing this on other Druid processes like realtime or historical.
However, one CoreOS, all nodes are giving me proper response. I am trying to find out what does a coordinator do differently than other node types that could cause this issue. Any insights would be helpful.