Premable: as a Rails 6.1.3 application, on deployment, I noticed that migrations were not launched, thus there may be a capistrano issue at hand.
Being an initial deployment, shared files were copied over, the database manually created, migrated and populated with a postgresql archive.
Launching the server however leads to a complaint
Webpacker can't find application.js
and in examining the directory structure, the public/packs directory is absent.
On localhost, these are created on first page call-up; but this is not occuring on a VPS.
Upon consultation of the Rails guide for webpacker neither command
bundle exec webpacker:compile nor bundle exec assets:precompile run, both with the same error:
bundler: command not found: assets:precompile
Install missing gem executables with `bundle install`
yet bundle install runs successfuly repeatedly, invoking webpacker 5.4.0.
config/webpacker.yml has
development:
<<: *default
compile: true
So this behaviour is contrary to the assertion in the guide "By default, Webpacker compiles automatically on demand in development when a Rails page loads. "
How should this be configured for proper deployment, in both development and production modes?
Related
I'm trying to deploy an API-only Rails 5 application to AppEngine Flex w/ the standard Ruby runtime, and I'm getting the following error at the very end:
Updating service [default] (this may take several minutes)...failed.
ERROR: (gcloud.app.deploy) Error Response: [9]
Application startup error:
/usr/bin/env: 'ruby2.5': No such file or directory
I'm specifying ruby '2.5.1' in my Gemfile and I've added an explicit .ruby-version file to the root of my project set to 2.5.1 as well.
I have no other debugging information available to me in the logs, no other fancieness. My entrypoint command is:
bundle exec rails server Puma -p $PORT
I can provide more details if needed, not sure what else might be relevant. Any pointers? As far as I can tell, nothing on my side is asking for a version of ruby that specifically at execution time.
Thanks!
EDIT: Here's my app.yaml file
entrypoint: bundle exec rails server Puma -p $PORT
env: flex
runtime: ruby
UPDATE:
I can verify that I'm having similar problems when trying to exec rake tasks like db:migrate:
--------- EXECUTE COMMAND ----------
bundle exec rake db:migrate
/usr/bin/env: 'ruby2.5': No such file or directory
ERROR
ERROR: build step 0 "gcr.io/google-appengine/exec-wrapper:latest" failed: exit status 127
--------------------------------------------------------------------------------------------------------------------------------------------------------
OK I now see what has happened after debugging the docker image locally. Because I was on ubuntu and had used system ruby to install gems it had embedded /usr/bin/env ruby2.5 into every executable script that was bundled into my app. I deleted all gems and switched to rbenv for managing Ruby version which mitigated this odd behavior between unbuntu's ruby and my app.
I´m having trouble to access the rails console in production.
I used Capistranoto deploy the app to a VPS
If I cd to deploy#myapp:~/myapp/current$and run bundle exec rails cthere I always get the option list for creating new rails project, like rails new
I've also tried bundle exec rails console --productionand rails consoleetc all with the same outcome.
the thing is I must be able to access the console because I have to create an admin user for active admin
might be worth adding that I'm using Passenger/Capistrano and Nginx on Ubuntu 16
Does anyone know what is going on here? Am I doing something wrong?
*EDIT
After running RAILS_ENV=production bundle exec rails c I get this message
Looks like your app's ./bin/rails is a stub that was generated by Bundler.
In Rails 4, your app's bin/ directory contains executables that are versioned
like any other source code, rather than stubs that are generated on demand.
Here's how to upgrade:
bundle config --delete bin # Turn off Bundler's stub generator
rake rails:update:bin # Use the new Rails 4 executables
git add bin # Add bin/ to source control
You may need to remove bin/ from your .gitignore as well.
When you install a gem whose executable you want to use in your app,
generate it and add it to source control:
bundle binstubs some-gem-name
git add bin/new-executable
Loading production environment (Rails 4.2.5)
irb(main):001:0>
You are missing executable files of bin folder in Production after Capistrano deployment.
You need to remove bin from set :linked_dirs from your Capistrano deploy.rb in order to avoid symlinking it.
You can again try cap production deploy, it would take all the executable files from bin to Production.
Now, you can access the rails console using:
RAILS_ENV=production bundle exec rails c
If you are using Capistrano 3, you can include the rails:console option which will allow you to do this from your local machine to gain access to the console on the remote host:
bundle exec cap production rails:console
https://rubygems.org/gems/capistrano-rails-console
The Rails project is deployed in /deploy/your_project_name/current on the server by default. So, you can access it via SSH or ... and run bundle exec rails c to access the Rails console. It works for me!
check if rvm and regarding gemset is suitable for the app
rvm gemset list
rvm list
and then do
bin/rails rails c -e production
Sorry if this question sounds basic. But I haven't been able to find an answer anywhere on the web...
I'm currently running my Rails app on a Ubuntu server. Until now I've always shut down the production app before I pull the changes, run rake assets:clean assets:precompile, and only boot up the Rails app again once the process is finished.
I'm not sure whether the shutting down of app is necessary(i.e. if I don't do it, my app will behave erratically). It induces about 5 minutes of down time.
If that's a must, then maybe I should try to do local precompilation/more advanced deployment procedure, in order to reduce downtime? (Tried local compilation according to http://guides.rubyonrails.org/asset_pipeline.html#local-precompilation, but then after deleting original public/assets and pulling locally precompiled public/assets from the repo, the production server was having rack timeout all the time and won't render anything.)
YES you run rake assets:precompile Rails looks through your assets folder and copies over everything that is not Javascript or CSS into public/assets. It then creates application.js by reading app\assets\javascripts\application.js, and application.css by reading app\assets\stylesheets\application.css, loading up all the "require" files it finds in there.
So yes..you can do it ..but if you ran rake assets:clean..and then precompile...then public/assets will be updated with new compiled assets.
Dont forget to restart the server :)
I have a directory structure like the following:
my-app/
.git/
db/ <-- Database stuff
lib/ <-- Business logic
spec/
web/ <-- Rails
Rails is in the web/ subdirectory of the repository. Heroku doesn't like this by default.
I have a partially working heroku buildpack. It basically calls Dir.chdir to change into the APP_SUBDIR environment variable (if it exists) before running any tasks.
I'd like to deploy to Heroku, but I can't get the asset pipeline task to work. When it comes time to precompile assets, the customized buildpack changes to the web/ directory and attempts to run the task.
However, I get this error, with the following important bit:
Could not detect rake tasks
ensure you can run `$ bundle exec rake -P` against your app with no environment variables present
and using the production group of your Gemfile.
Your Ruby version is 1.9.2, but your Gemfile specified 2.0.0 (Bundler::RubyVersionMismatch)
Which is weird since Heroku clearly says I'm running 2.0 in that very same pastie. In fact, when I run:
$ heroku run 'cd web; bundle exec rake -P'
it work just fine.
TL;DR (Summary)
Heroku thinks my ruby version is 1.9.2 when it's really the 2.0.0 shown in its own build process. Why might this be?
You can push only the subdirectory the Rails app lives in to Heroku:
git subtree push --prefix web heroku master
Where web is the subdirectory.
I have deployed a new Rails site to a Linux VM using Capistrano. I am using nginx as the front end and running my Rails app using unicorn.
If I try to run rake routes on the server, I get an error telling me that Rails is not installed, even though Rails is installed. The problem seems to be that the gem search directory is different for the app and the logged in user.
How do I load the Rails environment that my app is seeing as the logged in user?
Just use:
RAILS_ENV=production bundle exec rake routes
The RAILS_ENV part sets your environment variable so your app is loaded in full production mode, including database settings and so on.
The bundle exec part is necessary so that any commands that come after that are executed within the environment of the gems installed in your Gemfile.