leap_cli should check for couchdb first before checking its stunnel endpoints
When couchdb is down, "leap test" is misleading the user, because it first checks stunnel which fails when couchdb is down and leaves the user debugging the stunnel config first. It should check first for couchdb.
--- pixelated-platform-environments/pixelated-project.org ‹master* M› » leap test dev1 = [dev1.dev.pixelated-project.org] PASS: Network > Can connect to internet? = [dev1.dev.pixelated-project.org] FAIL: Network > Is stunnel running? [network.rb:56]: = [dev1.dev.pixelated-project.org] Failed to open socket localhost:5984 = [dev1.dev.pixelated-project.org] Connection refused - connect(2) = [dev1.dev.pixelated-project.org] The local connect endpoint for stunnel `couch_server` is unavailable. = [dev1.dev.pixelated-project.org] This is probably caused by a daemon that died or failed to start on = [dev1.dev.pixelated-project.org] port `5984`, not stunnel itself. = [dev1.dev.pixelated-project.org] Tests halted on failure (because of --no-continue). = bailing out
root@dev1:~# curl localhost:5984 curl: (7) couldn't connect to host root@dev1:~# netstat -tulpen |grep 59 tcp 0 0 0.0.0.0:15984 0.0.0.0:* LISTEN 0 180690895 8678/stunnel4 tcp6 0 0 :::443 :::* LISTEN 0 209739959 10143/apache2
(from redmine: created on 2016-01-04)