I am trying to deploy on to OS X Server with capistrano+mongrel+apache2,
I’m trying to do a “cap cold_deploy” as recommended here:
I had a devil of a time getting cap to find the mongrel_rails command
(which is in /opt/local/bin) on the server. I set the path in the
.profile. I set it in /etc/profile. I rebooted the server. I set the
~/.ssh/environment as recommended here:
to the following:
Still nothing. I even went so far as to hardcode the path to
/opt/local/bin/mongrel_rails everywhere I could find "mongrel_rails in
the code: (the following two files on both my dev box and the server)
This worked somewhat.
I know I should not hardcode the path and believe me if that fixed it I
would submit a patch that parameterized it just the way I can use set
:cvs, “/path/to/cvs” in the deploy.rb.
Okay, so then it started complaining that it could not start because the
log path was invalid. Default is log/mongrel.log. Well fooey. I
hardcoded that to be: /Users/me/MyApp/current/log/mongrel.log
Then I had to do the same for the docroot because it could not find
Now, it appears to start the mongrel servers:
- executing task spinner
- executing task start_mongrel_cluster
- executing “sudo /opt/local/bin/mongrel_rails cluster::start -C
[www. myserver.com] executing command
** [out :: www. myserver.com] Starting 3 Mongrel servers…
Looks good, right? Wrong.
There are no files in the log dir – no .pids and no mongrel.log – and
there are no mongrel processes running and nothing listening on 8000.
Even when I run the command by hand, it exits and there is no new
/opt/local/bin/mongrel_rails start -d -e production -p 8000 -l
Anyone have any idea what the issue could be?
Very Much Appreciated,