Description of problem: Making the port 5000 (default one in /etc/katello/thin.yml) busy (ssh -L 5000:localhost:80 root@localhost) brings then `service katello start` to react as "[OK]" while the thin instance was failed to start up. Would be better to do a netstat check before returning from the init/katello->start procedure. It would be important especially for the case when there would be servers = 1 there (one thin instance which failed) Fortunately `service katello status` tells the truth :) Version-Release number of selected component (if applicable): katello-0.1.117-1.el6.x86_64 How reproducible: always when the thin port is busy Steps to Reproduce: 1. see above 2. 3. Actual results: `service katello start` returns 0 - instead the failure status would be appreciated Expected results: error code (with logging the failure cause) Additional info:
Possibly related: https://bugzilla.redhat.com/show_bug.cgi?id=752863
Thin does not report any error when port is already used. Reported upstream: https://thin.lighthouseapp.com/projects/7212-thin/tickets/150-thin-does-not-report-error-when-port-is-used We might do an explicit check, but since we would need to load configuration (port numbers) and do this in a loop, it is not an easy task. Setting low priority.
https://github.com/Katello/katello/pull/575
Merged. Commit 6da9ee9
Part of Katello 1.1 release