仅生产环境出现ActionController::RoutingError (uninitialized constant Members)求助

Hey there, let's troubleshoot this frustrating production crash you're hitting on Heroku—especially since your Rails 5.1.4 app runs perfectly in development with Ruby 2.5.0. Since you've already ruled out the common dev environment naming issues others faced, let's focus on production-specific gotchas that often trip up Heroku deployments:

1. Check for Missing Environment Variables #

Heroku's production environment doesn't have access to your local .env file (if you use one). Run heroku config in your terminal to list all environment variables set on Heroku, then compare them to what you have locally. Missing keys for databases, APIs, or third-party services are a super common cause of production crashes.

2. Verify Asset Precompilation #

Rails precompiles assets during Heroku deployments, but local dev skips this step (since it compiles assets on the fly). Test precompilation locally first with:

RAILS_ENV=production rails assets:precompile

If this throws errors, fix them before redeploying. Also, make sure any custom JS/CSS files are either included in your application.js/application.css manifests or added to config.assets.precompile in config/initializers/assets.rb.

3. Don't Forget Database Migrations #

It's easy to overlook this step after deploying! Run:

heroku run rails db:migrate

Even if migrations work in dev, production might have schema discrepancies or missing tables that cause your homepage to crash when it tries to fetch data. Also, double-check that your migrations don't include dev-specific logic (like local file paths) that would break in production.

4. Confirm Ruby/Rails Versions on Heroku #

Sometimes Heroku uses a different Ruby or Rails version than what's in your Gemfile. Verify with these commands:

heroku run ruby -v
heroku run rails -v

Make sure your Gemfile explicitly specifies ruby '2.5.0' and gem 'rails', '5.1.4' to avoid version mismatches.

5. Dig Into Heroku Logs for the Exact Error #

The most critical step is getting the full stack trace of the crash. Run:

heroku logs --tail --source app

This will stream real-time app logs, so you can trigger the homepage request and see exactly which line of code is failing—whether it's a missing constant, database error, or something else. The logs will give you the specific detail you need to pinpoint the issue.

6. Rule Out Eager Loading Naming Issues #

While you said others hit naming issues in dev, production uses eager loading by default (unlike dev's autoloading). This means Rails loads all constants upfront, so any naming inconsistencies (like a model file named user.rb but the class is Users, or a module name mismatch) that slip through in dev will crash production. Double-check that all your files follow Rails conventions: snake_case filenames for CamelCase constants, and that your autoload paths in config/application.rb are set correctly.

Once you get the stack trace from the logs, you'll have a much clearer picture of what's going wrong.

内容的提问来源于stack exchange，提问作者Christopher Warrington