So... I've started over, using the standard install location for the time being in order to decrease the number of changes to track. And I've noticed a couple of things as I work through this:
1.) I'm having to source rvm.sh in order to get some of this stuff to work (such as unicorn). Is there something I'm missing that should be causing this to not be an issue? 2.) Many places where you are calling rake tasks and the like, I am having to substitute </path/to/>bundle exec <rest of command>. The path isn't necessary if I've sourced the rvm.sh, which makes sense. Is this normal? I can see how using bundle exec may make sense, given all the gems being installed by bundler, but I also notice you don't have it in the puppet recipe you posted in your last response, and it's not working without it. Apart from these two things, and having to do a copypasta job to recreate the CentOS lines as OracleLinux instead, it seems to be going well. But I don't want to veer quite so far afield as I was before. Appreciate any help you can give me. I'm quickly narrowing in on the target, my git diff is nice and small at the moment, so I know I'm sticking much more closely to how the installer was originally written. Just need to settle out the last few things. Thanks in advance, Damian On Friday, June 28, 2013 4:31:33 AM UTC-4, Marius Mårnes Mathiesen wrote: > > > [email protected] <javascript:> writes: > > Sorry, Chris posted this for me as my messages didn't seem to be going > > through. I see both of them now, though. > > :-) > > > Regardless, the majority of the changes are to move the application to a > > new location on the filesystem to keep within our standards. The stack > as I > > currently have it configured is nginx, unicorn, stompserver, resque, > native > > git-daemon with the git-proxy, thinking-sphinx (bundle exec rake > ts:start), > > and mysql. > > Okay. You won't be needing the stompserver though, Redis+Resque takes > care of the messaging stuff now. > > > gitorious_status claims everything is up except for git-daemon and > mysql. > > I've logged into mysql with the username and pass which gitorious is > > configured to use and I see that the database is there. And I have > cloned > > from the lone repository I've been testing with using the git:// url, so > > the git-proxy -> git-daemon interaction seems to be working. > > We have been seeing some issues with Monit not handling the > git-{daemons,proxy} setup properly. We will probably be switching to > using Upstart for these two instead of Monit, here's a Puppet recipe > which relieves Monit of this responsibility and uses Upstart for it > instead: > > http://pastie.org/private/b3pu9sx4bvuecztx6jlvbw > > The way this works is: > - the resque-worker is started when runlevel 3 (or 5) is entered. > - once the resque-worker is up, the git-daemons are started too > - once the git-daemons are started, the git-proxy is started > - when shutting down, the opposite happens > > There seems to be an edge case where both the resque-worker and > git-proxy will shut down the first time they're started (maybe an > init.d-based service hasn't started yet?) so the recipe above has a > respawn instruction which handles this. > > > I did have to create init scripts for the git daemon, as running it > through > > xinet.d wasn't working (it will not accept port/listen arguments in inet > > mode). I also created one for stompserver, as it was not starting along > > with everything else either. > > Re. the inet.d setup, I've been experiencing the same trying to use > Systemd socket activation for Ruby-based servers. Shame. > > Cheers, > - Marius > -- -- To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] --- You received this message because you are subscribed to the Google Groups "Gitorious" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
