Ran through Quick Start, but no luck

[Jason Amster]

Hi All,

I just ran through the rubber Quick Start and it seems like everything
went through okay but the rubbertest rails app is not running. I just
get the basic "It Works" page when I try to access the app. I checked
the apache config directory and I don't see any new conf files for the
the app.

Additionally, I created a web_tools role and it also didn't install
the apache config files. Anything I could have missed?

-Jason

[Brandon Casci]

Did you deploy yet? If you did was it a deploy:cold for the first deploy? It's possible mongrels or passenger service wasn't started.

Also, if this is a single instance deployment, make sure the app role was added. If it only has the web role, then the instance will only act as a web server, and only run apache, not the rails app.

--

You received this message because you are subscribed to the Google Groups "rubber" group.
To post to this group, send email to rubbe...@googlegroups.com.
To unsubscribe from this group, send email to rubber-ec2+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/rubber-ec2?hl=en.

--
=========================================
I'm having a bout of spondylosis, so I can't hammer away at the keyboard too long. Please don't mistake short e-mails for rude e-mails..
=========================================

[Jay Amster]

I just did a "cap deploy."  I see the files up on the server for the rails app, but I can't find any config files for apache.  Where in Rubber does it run through the process of parsing the ERB for the configs and uploading them for apache?

-J

[Brandon Casci]

cut and paste in the roles section if rubber/instance-production.yml

Let's see which roles have been added to that instance.

[Jay Amster]

I've kinda run through this a bunch of times, so this is my current try.  This  is following the last set of instructions on how to do a the more complex setup.

http://pastie.org/718224

[Matthew Conway]

Roles look ok.

I take it rubber:bootstrap ran without any errors?  You could try running it again in case there was a failure that didn't stop the sequence - some processes don't exit with error codes, let me know if you find any.  It should be idempotent, so multiple runs are safe.  Try a "rubber:bootstrap", then a "deploy" and see if that helps.  If not, capture the output from both and I'll take a look.

Were you able to do a single instance ok (cap rubber:create_staging)?

In response to your question about ERB, rubber injects "rubber:config" in the capistrano lifecycle after deploy:update_code, but before deploy:restart or deploy:migrate.  This tasks basically just expands the ERB templates from the code that was just pushed to the servers.  It uses the rubber-*.yml, instance-*.yml, that were also pushed to the server in order to introspect your cluster and configure it according to the machine rubber:config is running on.  If you are using an SCM like git, you need to check in all these files before they will be available on the server as part of deploy:update_code, however, if doing the rubber quickstart, it just zips up your directory tree, so gets them that way. If using SCM, as a convenience you can also set "push_instance_config" in deploy.rb to force pushing instance*.yml without checking it in.

Matt

On Nov 28, 2009, at Sat Nov 28, 11:13 AM, Jay Amster wrote:

I've kinda run through this a bunch of times, so this is my current try.  This  is following the last set of instructions on how to do a the more complex setup.

http://pastie.org/718224

[Jay Amster]

I didn't see any errors.  I went ahead and redid the single instance method so that it was less of a pain to debug. 

Main issue here is that it just shows the "It Works"  message from apache's default site and no specific app based apache config in the folder.

attached is the the output of the cap rubber:create_staging command and the instances file.

Thanks!!!

[Matthew Conway]

Looks like the latest rails gem is 2.3.5, but you are creating rubbertest2 with 2.3.4, you either need to comment out RAILS_GEM_VERSION in config/environment.rb, or specify rails version to install in rubber.yml, or update your local rails before generating rubbertest2. Capistrano output isn't the most obvious when it fails, its very easy to overlook the "failed:" at start of last line. I should probably hook into exit and make failures seem more obvious.

.
<snip>
.
** Successfully installed rails-2.3.5
.
.<snip>
.
* executing "sudo -p 'sudo password: ' sh -c 'cd /mnt/bingo-production/releases/20091128165636 && FILE=\"role/mysql_master|role/db/my.cnf\" RUBBER_ENV=\"production\" rake rubber:config'"
servers: ["production.bingobuster.com"]
[production.bingobuster.com] executing command
*** [err :: production.bingobuster.com] Missing the Rails 2.3.4 gem. Please `gem install -v=2.3.4 rails`, update your RAILS_GEM_VERSION setting in config/environment.rb for the Rails version you do have installed, or comment out RAILS_GEM_VERSION to use the latest version installed.
*** [err :: production.bingobuster.com]
** [out :: production.bingobuster.com] (in /mnt/bingo-production/releases/20091128165636)
command finished
failed: "sh -c 'sudo -p '\\''sudo password: '\\'' sh -c '\\''cd /mnt/bingo-production/releases/20091128165636 && FILE=\"role/mysql_master|role/db/my.cnf\" RUBBER_ENV=\"production\" rake rubber:config'\\'''" on production.bingobuster.com

[Jay Amster]

Yep that was def it.  This is an awesome tool!  Thanks for building it and sharing it!

-Jason

[Brandon Casci]

BTW - this has happened to me a few times. It's an easy thing to miss or totally forget about. I'm in the habbit of specifying the gem version of my dev trunk in the gems section of rubber.yml now.