Recent Posts by David Smalley

Topic: Passenger / Passenger+Nginx

@pet3r – actually, nginx+passenger is so new that none of us have had a chance to battle test it in a production environment. I do believe that john has been building the package for it though, although our current version of the gem is unlikely to support quick setup out of the box (nothing to stop you doing the vhost config manually though).

We do have a new version of the gem in the pipeline, but its still in the planning phase so unlikely to be out for a few months.

Having not deployed any apps to nginx+passenger I can’t tell you how long or complicated it will be to move from apache+passenger to nginx+passenger. However, given that passenger is stupid simple to setup, I would guess not too long unless you’re doing anything custom in your vhost (rewrite rules etc.)

Topic: Brightbox Help / new mongrels

Hi Volker,

The gem is actually configured for a standard rails app deployment. We do the kind of configuration you talk about ourselves sometimes for our internal projects and it does always have to be done by hand.

Topic: Brightbox Help / new mongrels

Hi Volker,

It sounds like you don’t have rubygems and/or the brightbox-server-tools package – try a ‘sudo apt-get install rubygems’ and a ‘sudo gem install brightbox-server-tools’.

Topic: Brightbox Help / new mongrels

Volker.

It shouldn’t matter – migrations that have already been run won’t run again. Otherwise, I’m not sure how you’d stop them running, you could dig into the source of the brightbox gem and try commenting out the part that runs migrations. We’re working on a new version of the gem and every report like this is very useful for us to plan what extra features need adding.

Topic: Brightbox Help / new mongrels

Hi Volker.

If you change the number of mongrels you’ll need to run “cap deploy:initial” again. This is the task that does the initial setup of mongrels and the Apache configuration.

Topic: Brightbox Help / deployment problems

ant284 By default cap deploy:initial will tar the files in your local application directory and upload them to your brightbox. If /home/rails/newapp/shared and /releases are getting created then the deploy is almost succeeding but it sounds like you are hitting some issue that is causing the deploy to fail and not finish up by symlinking /current to the most recent release.

If you can perhaps let us know what failure message is for Capistrano then it might help us to diagnose. If you have to paste lots of text then try just posting back with a Pastie link.

Topic: Brightbox Help / will_paginate gem issues?

Hi David,

We’re aware of issues with installing gems with the Brightbox gem. On the whole it does work but there are some corner cases where it seems to fall down. We’ll soon be looking at version 3 of the Brightbox gem and how we can overhaul the gem dependency management. I’m sorry it caused you trouble.

Topic: Brightbox Help / will_paginate gem issues?

David, are you able to pin down the precise cause of this 500 error? Check /home/rails/yourapp/current/log/production.log and also have a look in /var/log/web/ for error logs generated by your application.

If you can figure out the exact reason that the web server is producing a 500 error and whether the error is coming from Rails or Apache we should be able to help you pin it down.

Topic: Brightbox Help / deployment problems

@ant284 Can you clarify this step:

“then i upload the Allapp to my subversion on brigthbox”

Are you committing the code to subversion via svn+ssh? What happens if you type:

svn ls svn+ssh://MYACCOUT-001.vm.brightbox.net/home/rails/subversion/<reponame>

Topic: Brightbox Help / Webistrano Compatible with Brightbox Gem?

Hi David.

We haven’t ever actually tried using the brightbox gem with Webistrano so I’m not too sure – you basically include a ‘require ’brightbox’’ line in your deploy.rb and it adds in the magic. You can give it a go and let us know how you get on.

David

Topic: Passenger / Is there a timescale to support Rails 2.3 on Passenger?

@joshhart – we have the beta version of Passenger in testing at the moment. I’ll let john comment on the timeline as he’s handling the packaging.

Topic: Passenger / Passenger crashes the server - how do I remove?

@dw you should just be able to do

apt-get remove libapache2-mod-passenger

Topic: Brightbox Help / Gem 2.3.1 and gem dependencies

@sam.

This is my fault completely. Back to the drawing board on how to handle custom dependencies. I thought I’d properly tested this fix but clearly I hadn’t.

I’ve filed a bug in our internal tracker and will try to get this resolved asap.

Topic: isitruby19.com / Delete my rating

@Chris Lloyd I’ve deleted your erroneous comment now.

Topic: isitruby19.com / GitHub Gem Import

@elliottcable that’s a good idea. We’re looking in to how best to do this and hopefully we’ll have something live soon.

Topic: isitruby19.com / API is added

We have just added a basic JSON and XML API.

Just hit any resource with the appropriate extensions/accept headers and you are away.

curl http://isitruby19.com/rails.json

{"code": {"name": "rails", "code_type": "gem", "rubyforge": "rails", "permalink": "http://isitruby19.com/rails", "summary": "Web-application framework with template engine, control-flow layer, and ORM.", "description": "Rails is a framework for building web-application using CGI, FCGI, mod_ruby, or WEBrick on top of either MySQL, PostgreSQL, or SQLite with eRuby-based templates.", "homepage": "http://www.rubyonrails.org"}}

curl http://isitruby19.com/hpricot/comments.json

[{"comment": {"name": "Jai-Gouk Kim", "body": "Ubuntu 8.1, x86, installed from github.", "works_for_me": true, "url": "http://jaigouk.blogspot.com", "version": "0.9.198", "platform": {"name": "GNU/Linux"}}}, {"comment": {"name": "postmodern", "body": "gcc -I. -I/usr/include/ruby19-1.9.1/x86_64-linux -I/usr/include/ruby19-1.9.1/ruby/backward -I/usr/include/ruby19-1.9.1 -I. -DHAVE_STDIO_H -fPIC -O2 -g -Wall -Wno-parentheses -fPIC -o fast_xs.o -c fast_xs.c\r\nfast_xs.c: In function \u2018fast_xs\u2019:\r\nfast_xs.c:169: error: \u2018struct RArray\u2019 has no member named \u2018ptr\u2019\r\nfast_xs.c:170: error: \u2018struct RArray\u2019 has no member named \u2018len\u2019\r\nfast_xs.c:175: error: \u2018struct RArray\u2019 has no member named \u2018ptr\u2019\r\nfast_xs.c:176: error: \u2018struct RArray\u2019 has no member named \u2018len\u2019\r\nmake: *** [fast_xs.o] Error 1\r\n", "works_for_me": false, "url": "http://houseofpostmodern.wordpress.com/", "version": "0.6.164", "platform": {"name": "GNU/Linux"}}}, {"comment": {"name": "Jeremy Kemper", "body": "fast_xs has been fixed: <a href="http://github.com/why/hpricot/commit/2cf87f4">http://github.com/why/hpricot/commit/2cf87f4</a>", "works_for_me": true, "url": "http://bitsweat.net/", "version": "0.6.171", "platform": {"name": "Mac OSX"}}}, {"comment": {"name": "Ryan Neufeld", "body": "Fails to install on building fast_xs.c", "works_for_me": true, "url": "", "version": "0.6.164", "platform": {"name": "Mac OSX"}}}, {"comment": {"name": "Rahoul Baruah", "body": "Installs but tests won't run", "works_for_me": false, "url": "http://www.brightbox.co.uk", "version": "0.6.164", "platform": {"name": "Mac OSX"}}}]

Just replace .json with .xml for an XML flavour.

You can also access the author model.

curl http://isitruby19.com/authors/why-the-lucky-sti…

{"author": {"name": "why the lucky stiff", "codes": [{"name": "hpricot", "permalink": "http://isitruby19.com/hpricot"}, {"name": "camping", "permalink": "http://isitruby19.com/camping"}, {"name": "metaid", "permalink": "http://isitruby19.com/metaid"}, {"name": "markaby", "permalink": "http://isitruby19.com/markaby"}], "permalink": "http://isitruby19.com/authors/why-the-lucky-stiff"}}

And carry out a search

curl http://isitruby19.com/codes.json?s=activemerchant

[{"code": {"name": "activemerchant", "code_type": "gem", "rubyforge": null, "permalink": "http://isitruby19.com/activemerchant", "summary": "Framework and tools for dealing with credit card transactions.", "description": null, "homepage": "http://activemerchant.org/"}}]

And list all the code…

curl http://isitruby19.com/codes.json?page=1

Topic: isitruby19.com / GitHub Gem Import

I have nearly readied the GitHub gem import, hopefully once this is in then I will setup the Gem imports to run on an automated basis once every few hours, to make sure we have the latest gems ready for review.

One question on my mind, while I really think we should have GitHub gems in here, will it dilute the data? Will it be confusing to see that davidsmalley-eventmachine doesn’t work while bobjobs-eventmachine does work? I am slightly worried that it might dilute the overall impression we get of Ruby 1.9.1 compatibility. Are there any features that can help overcome this or should I just go ahead and get the data imported already :)

Topic: isitruby19.com / automatic gem testing

@grosser we had been thinking about how this could work. We wanted to get the site launched as quickly as possible so we put this feature back on the shelf.

We were keen for as many eyes as possible to be on each gem, so we can be sure that a gem really works before it gets the green light. If we do automated testing then we run the risk of giving a pseudo-“official” answer on the compatibility of a gem and there is no way we could guarantee it. I think best for now that as many people as possible check out each gem.

Topic: isitruby19.com / Delete my rating

@Chris Lloyd To keep things simple we don’t actually log anyone in, just use the captcha. But I agree, it would definitely be useful to remove a rating that you’ve made incorrectly. That said, we want to keep this site moving without getting too bogged down in little features so perhaps for now you can just do as you did and vote again with an explanation :)

We’ll definitely look at it as a future feature.

Topic: isitruby19.com / JRuby gems?

@martinbtt Good point, we personally don’t use JRuby but would hate to see it neglected. Do you think we need another site or could we use tags/some other metadata to split the site up a bit and make it easier to get at the different types of gems?

Topic: isitruby19.com / Metric to decide on compatibility?

Nice idea.

Topic: isitruby19.com / Welcome!

I’ve just hacked up JSON and XML actually, will be adding shortly.

Topic: isitruby19.com / Score board

We’d have to add some kind of login/user verification for that to happen.

Topic: isitruby19.com / Metric to decide on compatibility?

We didn’t add this in the end as we couldn’t think of the best way to do it, but if anyone has any ideas…

Basically we’re looking for a score that tells us how likely a gem is to be compatible, now obviously we could do the number compatible divided by the number incompatible but then that’s likely to penalise the gems that have just been fixed.

One idea was to only count the last X (say, 10) comments towards the score, but any other ideas welcome.

Topic: isitruby19.com / Welcome!

We’re all as excited as you about the first stable release in the Ruby 1.9 series!

The big question now, is when can we all switch?

We’ve put this site together to try and keep track of which gems and code are working with Ruby 1.9 so people can make an informed decision about when to switch. We’ve already got some great ideas for features on the site but we’re interested to hear your suggestions and bug reports.

Some feature ideas in the pipeline:

• Create a new entry (so you can add your favourite plugin as well as official Rubygems)
• A widget that shows the support status for a particular piece of code

We would love to hear more…

Topic: Brightbox Help / Letting Apache server public/sitemaps/* files?

I think that the setup of the Apache config is supposed to only proxy things which do not exist as physical files.

If you look in /etc/apache2/brightbox-common you will see:

Odd that it seems to be sending your sitemap request through to the balancers anyway though.

Glad you have a solution now.

Topic: Brightbox Help / Shared Memcached Cluster?

Hi Jörg.

Actually, I think the main issue here is security. Memcached uses a completely un-authenticated access model so there would be no way to stop someone else with access setting/getting/clearning your cache. Indeed, the only way we could restrict access would be to setup firewall rules in front of memcached and limit who could access the port but that would still allow unfettered access to the whole cache for anyone who signed up.

Depending on your caching needs, you could always run memcached alongside your existing app server. Most standard apps can make do with 64 or 128mb of Memcached RAM quite easily.

Topic: Brightbox Help / long query time on brightbox

Hi Ernesto,

“In any case, would it be reasonable that the very same code takes 10 times longer on the server than my computer?”

Actually yes this is perfectly reasonable. You’d be surprised how many support queries I’ve dealt with in my time (I’ve worked for two Ruby on Rails hosting companies) where people are able to run a query fine on their 2.2 ghz dual core Macbook Pro with 4gb of RAM and then are baffled when the same inefficient query won’t run as fast on their much lower powered virtual machine. Unfortunately all I can suggest is you walk through debugging this and try to ignore the fact that it seems to run quickly on your local machine.

Try using Ruby-prof and bleak_house to analyse the memory usage of the intensive action and which methods are being called and how often. This can give you a good idea of whether you action is doing something that might take orders of magnitude longer on a much slower box, such as using REXML or doing lots of Regex’s.

Also remember that if you are on our smaller Brightbox plans then it may simply be that your box is not powerful enough for what you are trying to achieve without serious optimisation.

I think Filesort is normally a sign that you haven’t got a good index for a query – but don’t quote me on that as I am not a database expert myself.

I’m fairly confident that what you are experiencing it nothing to do with the stability of your particular Brightbox itself and I can assure you that our database cluster is very carefully monitored for load and is currently under utilised.

Topic: Brightbox Help / no such file to load -- rubygems

Yeah you need to specify the Rails environment.

It’s available as a local variable in deploy.rb so you could do…

run “RAILS_ENV=#{rails_env} cd #{release_path} && lib/daemons/parser_ctl restart”

Topic: Brightbox Help / no such file to load -- rubygems

You can add custom tasks to your Capistrano recipe and set them to run before or after other tasks. E.g. your daemon and worklings might be…

after "deploy:restart", "custom:restart"

namespace :custom do

  task :restart do
    run "cd #{release_path} && script/daemon restart"
  end

end

You can find plenty examples of how to write your own Capistrano tasks online.