akram
commented
8 years ago In my case, I wanted to run the OpenShift router on alternative port: 1080 and 10443. And I ended up in this case
-- Checking for available ports ... FAIL
Error: a port needed by OpenShift is not available
Caused By:
Error: ports in use: [443]So maybe the check has been implemented in the meantime or it was already existing. In both case, it prevents from running the router on other ports. Should we add an option? o
csrwng
coreydaley
frobware
openshift-bot
GrahamDumpleton
adambkaplan
mfojtik
ironcladlou
danwinship
If port 80 is already in use by a web server on the host when running
oc cluster up, there is no error or warning when starting up. What happens is that the router pod keeps failing instead and so although the web interface works and can deploy applications, you can't access the deployed applications. If you use docker logs on dead router containers they show no log messages at all.The
oc cluster upcommand should perhaps do a TCP connection to both 80 and 443 when run and warn that the ports are already in use and that routing may not work for applications.So what normally expect to see from
docker psis:If start up Apache httpd server on port 80 and run
oc cluster up, then output all looks normal.The router pod doesn't show as running in
docker ps.If you run
docker ps -ayou can though see many failed router pods.As said above, using
docker logson any of the router pods shows no log messages at all, so would have to guess issue is that something is running on port 80.