This is the first rock-on I’ve tried to install. Tried using two different pools for storage, no luck. Might not be the cause, but the log says something about port 443 being already occupied, and I left the default option for port 8082 in the gui.
[root@weiland log]# more rockstor.log
…
…
[03/Apr/2016 13:22:16] DEBUG [storageadmin.views.rockon:84] Rockon(Discourse) is in pending state
[03/Apr/2016 13:22:18] DEBUG [storageadmin.views.rockon:84] Rockon(Discourse) is in pending state
[03/Apr/2016 13:22:32] DEBUG [storageadmin.views.rockon:84] Rockon(Discourse) is in pending state
[03/Apr/2016 13:22:34] DEBUG [storageadmin.views.rockon_helpers:127] exception while installing the Rockon(13)
[03/Apr/2016 13:22:34] ERROR [storageadmin.views.rockon_helpers:128] Error running a command. cmd = [’/mnt2/discourse2/discourse/launcher’, ‘start’, ‘discourse’]. rc = 1. stdout = [’’, ‘Launcher has detected that port 443 is in use.’, ‘’
, ‘If you are trying to run Discourse simultaneously with another web server like Apache or nginx, you will need to bind to a different port.’, ‘See https://meta.discourse.org/t/17247 for help.’, ‘’]. stderr = [‘which: no docker.io in (/
usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin)’, ‘’]
Traceback (most recent call last):
File “/opt/rockstor/src/rockstor/storageadmin/views/rockon_helpers.py”, line 125, in install
globals().get(’%s_install’ % rockon.name.lower(), generic_install)(rockon)
File “/opt/rockstor/src/rockstor/storageadmin/views/rockon_discourse.py”, line 94, in discourse_install
run_command([’%s/launcher’ % repo, ‘start’, rockon.name.lower()])
File “/opt/rockstor/src/rockstor/system/osi.py”, line 92, in run_command
raise CommandException(cmd, out, err, rc)
CommandException: Error running a command. cmd = [’/mnt2/discourse2/discourse/launcher’, ‘start’, ‘discourse’]. rc = 1. stdout = [’’, ‘Launcher has detected that port 443 is in use.’, ‘’, ‘If you are trying to run Discourse simultaneousl
y with another web server like Apache or nginx, you will need to bind to a different port.’, ‘See https://meta.discourse.org/t/17247 for help.’, ‘’]. stderr = [‘which: no docker.io in (/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin)’,
’’]
Does anyone else have this issue? I would really like to use this plugin and I’m not sure how troubleshoot any further.
I would look at /mnt2/discourse/discourse/launcher and see what this is doing. I would look, but rockstor is currently up and running for me atm. See if there is something specifying 443, or if it is looking at a script or another config file that is pointing to 443.
Or maybe here /opt/rockstor/src/rockstor/storageadmin/views/rockon_discourse.py
But basically if you see 443, change it to the a secondary port like 44333 and see if it succeeds.
Thanks for the ideas. I didn’t find any references to port 443 in there. I don’t really code but the port seems to be a variable in the error message.
Is this a clue? I’m a linux newb and just learned grep yesterday, but I think this is a clue.
more /mnt2/discourse/discourse/launcher | grep '${1}'
local valid=$(netstat -tln | awk ‘{print $4}’ | grep “:${1}$”) <----where port 443 came from?
echo “Launcher has detected that port ${1} is in use.” <----line in log
The 0.0.0.0:443 route is definitely output by netstat on my system. Should this command (or is this command) run inside a docker container with it’s own netstat? Now that I think about it, I’m not sure exactly how isolated a Docker is…
I tried to change the webgui port in nginx config file and reboot, but that caused all sorts of problems including the rock-ons service refusing to start (I was hoping to install the plug-in and then change it back). After moving it back to 443 I had a few disconcerting failures to boot. When I plugged a monitor into it this morning there was some sort of btrfs timeout message but it booted right away without complaint on the next try. Just wanted me to watch it, I guess. It appears to be functioning normally again.
welp, I broke it. thought I would get around this by removing netstat entirely. Unfortunately I don’t know what else is included in net-tools but apparently it is important.
yum remove net-tools
No more web interface.
yum install net-tools
Still no web interface
reboot
still just ssh access, nginx won’t start
Process: 3759 ExecStart=/usr/sbin/nginx (code=exited,
status=1/FAILURE)
I think that’s my cue to reinstall. Fingers crossed for Discourse magically working this time.
Looks like the same thing on this fresh install again.
[06/Apr/2016 11:38:59] DEBUG [storageadmin.views.rockon:84] Rockon(Discourse) is in pending state
[06/Apr/2016 11:39:09] DEBUG [storageadmin.views.rockon_helpers:127] exception while installing the Rockon(14)
[06/Apr/2016 11:39:09] ERROR [storageadmin.views.rockon_helpers:128] Error running a command. cmd = [’/mnt2/discourse
/discourse/launcher’, ‘start’, ‘discourse’]. rc = 1. stdout = [’’, ‘Launcher has detected that port 443 is in use.’,
’’, ‘If you are trying to run Discourse simultaneously with another web server like Apache or nginx, you will need to
bind to a different port.’, ‘See https://meta.discourse.org/t/17247 for help.’, ‘To continue anyway, re-run Launcher
with --skip-prereqs’, ‘’]. stderr = [‘which: no docker.io in (/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin)’, ‘
’]
Updated to latest testing release and got the same thing.
I want to use the --skip-prereqs option but I’m not sure how. tried cd to /mnt2/discourse/discourse/ then running ‘launcher --skip-prereqs start discourse’ but get a launcher: command not found error.
I think I’m going to give up now. Probably way easier to just install this in a Centos VM.