Recently I upgraded an application from Rails 3 to Rails 4. In the deploy scripts I precompile the assets locally and then rsync them up to the server(s). In Rails 4 the asset pipeline now produces manifest- < random > .json instead of a manifest.yml. Since the manifest files are named differently, this adds multiple manifest.json files to the shared assets directory. The application then picks up the wrong manifest file, and serves old assets.
I have read about various issues related to this in some github pull request threads:
https://github.com/capistrano/capistrano/pull/412
https://github.com/capistrano/capistrano/issues/210
https://github.com/capistrano/capistrano/pull/281
My options seem to be:
Don't share the asset directory.
This would break old clients asking for old resources.
Switch to compiling assets on the servers.
This would add complexity to the server.
Move the manifest file outside of the shared asset directory.
I have since learned that this option was removed in Rails 4.
Are there other solutions to this problem?
I found the best answer after looking at the standard capistrano rails asset precompile task. I added a command to the local precompile task that moves the old asset manifest to the current release as asset_manifest.json. This leaves only one manifest when the new one is uploaded.
run "mv -- #{shared_manifest_path.shellescape} #{current_path.to_s.shellescape}/assets_manifest#{File.extname(shared_manifest_path)}".compact
Moving the manifest-.json to the current_dir as asset_manifest.json allows capistrano to restore the correct manifest file on rollback.
Related
I've started learning rails and I've already built two apps, one simple blog app and one store app. Now I ran into a term precompile assets when uploading to heroku, can someone explain it to me is that necessary when deploying an app to production, because i've uploaded my store app to heroku without any problems?
Assets is your css + JS. Precompile assets mean that they get joined into single .css and another single .js. file (to load it in one HTTP request). And special mechanism of minifying get applied to both these files (to make them smaller). Rails by default is setup in a way, that it uses average files in dev and compiled files in prod. You can easily change this in configs, but you shouldn't do this unless you really know what you do.
If you want you can compile this files locally running rake assets:precompile and then put it into git. I think that you can disable/enable precompile during heroku deploy in heroku config. But, in general, I would stick with the very defaults.
More info on asset pipeline: http://guides.rubyonrails.org/asset_pipeline.html
Rails has an assets pipeline which consists of Sprockets and the assets helpers.
The assets pipeline will concat and minify your CSS and javascript and takes care of setting the correct paths to images and other assets. This is known as compiling the assets.
In development this is done on the fly for each request which lets you immediately see changes.
In production this would be far to slow so instead the assets should be compiled once at deploy time. Heroku does this automatically for you in a post-commit hook.
Pre-compiling is when you run rake assets:precompile locally and then upload or push the result to a server. This is done if you are deploying to a server without the support for the assets pipeline. For example if the production server does not have a javascript runtime which is required to run uglifier.
It adds tons of noise to the git change history and manually doing anything is a common source of user error. So pretty much it sucks and you only do it if you have to.
I pre-compiled my assets i.e CSS,javascript and fonts files in order to reduce the file size. Using RAILS_ENV=production rake assets:precompile.
I already have compressed assets on the server in the following path:
shared/public/assets
But anyway I went ahead and pre-compiled them again on my local and they got generated inside public/assets folder. I noticed that the compressed files are exactly as same as the ones on server shared/public/assets. But the tester in my team is been testing it on some online tools. And they all say "Your java script files need to be minified." So, would this "minifying" issue be solved if I push these locally pre-compiled assets to production source code?
You should not push locally pre-compiled assets to production source code, no.
Your build process should include precompiling assets for production during a deploy. If you're using Rails 5 then that is already turned on by default. As long as the production environment has a proper environment variable (again, by default, the environment variable 'production' takes care of all these things), the assets will be precompiled.
You say that your compressed files, after running rake assets:precompile, are exactly the same as the ones on the server. That means your javascript should have and should be minified and uglified (again, default for Rails 5). To confirm that's the case, open dev tools in chrome, hard-reload (ctrl+shift+r) and check with JS files are being loaded under "network" tab. If the asset pipeline was used as it should be, you should only see minified and uglified js files here. If you open them up, they should already be minified and uglified. If that's not the case then either the asset pipeline wasn't used for fetching the JS files or your build process has been changed.
If most of these JS files are minified and uglified with - check those which aren't. Is some library being added outside of asset pipeline? Are those files minified?
If all of your JS files are actually minified, then tell your tester to use a different tool ;)
Setup: Rails 4.2.5, Ubuntu 12.04, nginx 1.6, unicorn 5, capistrano 2.15
The setup did work with Rails 3.2 but once i started to migrate to rails 4.x i noticed an odd behaviour with the assets.
Inside the application the links to the asset files like application.js are not the ones inside the manifest.yml
If i do a deploy with asset cleanup, i will get many 404. The fingerprints used by the app and the manifest.yml differ.
If i do right away a second deploy without a cleanup, the assets will be served. They wont match with the manifest again, but the old ones are still in the folder. Killing unicorn and making a full restart did not change anything.
I have the same problem in the 4.0 and 4.1 branch of my app.
The assets are compiled on the server during deploy (capistrano scripts)
update:
I checked via console which fingerprints are generated and which files are present. Result: hashes are fine.
Problem is that the cleanup script deletes files older than a week or something and when doing a deploy (asset precompile) i get a wrong timestamp for some files.
If i do a precompile just now, some files have a datestamp 4th January and get deleted by the following cleanup script. Capistrano shows me some warning when writing that file but without any meaningfull information
I have a number of assets in my rails public/assets directory that I do not want compiled - specifically all the image files I want to use in my ember app. However, I cannot figure out how to get capistrano to deploy them. They are checked in to the public/assets directory but they do not appear in the deployed shared or release directories on my server. In fact I searched the whole capistrano directory tree on my server and the assets are nowhere to be found. I have confirmed that the assets are in git for the branch that capistrano is deploying. Any suggestions would be much appreciated.
I couldn't find a way to configure capistrano / capistrano-rails to allow me to use git source controlled files under public/assets, however, it did work to create an img directory under public and put the images there in git. That avoids the asset compilation process and the md5 hashes for the images and works well with ember.
In rails 3.1, when you precompile the assets, rails create public/assets directory and add files there.
Do you version-control public/assets/*?
I use Capistrano to deploy. The last step is compiling the assets. Nothing like that gets checked into version control.
https://github.com/capistrano/capistrano/wiki/Documentation-v2.x
Checking in compiled assets, .gz files/etc, will just clutter up version control.
I was looking for an answer to this too. I found the official Rails Guide has some thoughts on this:
http://guides.rubyonrails.org/asset_pipeline.html#local-precompilation
Here's a quote of the relevant section (emphasis added):
There are several reasons why you might want to precompile your assets locally. Among them are:
You may not have write access to your production file system.
You may be deploying to more than one server, and want to avoid duplication of work.
You may be doing frequent deploys that do not include asset changes.
Local compilation allows you to commit the compiled files into source control, and deploy as normal.
There are three caveats:
You must not run the Capistrano deployment task that precompiles assets.
You must ensure any necessary compressors or minifiers are available on your development system.
You must change the following application configuration setting:
In config/environments/development.rb, place the following line:
config.assets.prefix = "/dev-assets"
The prefix change makes Sprockets use a different URL for serving assets in development mode, and pass all requests to Sprockets. The prefix is still set to /assets in the production environment. Without this change, the application would serve the precompiled assets from /assets in development, and you would not see any local changes until you compile assets again.
In practice, this will allow you to precompile locally, have those files in your working tree, and commit those files to source control when needed. Development mode will work as expected.
So, it looks like it might be a good idea to put precompiled assets into VCS on occasion.