I'm running a NodeJS app on Openshift using Jenkins for building deployments (and I'm pretty new to both Node and cloud-based servers). My app depends on a package that has a binary component, so I can't just check it into git - it fails when it's executed on the server. I'm wondering what's the best way to deploy these sorts of dependencies. I see that there is an $OPENSHIFT_DEPENDENCIES_DIR (as well as $OPENSHIFT_BUILD_DEPENDENCIES_DIR), but I can't find any information about how (or if) these can be utilized for node modules. It would be great if I could keep all my dependencies on the server and out of my source tree.
Thanks!
Update: I forgot to mention that I need to apply a patch to the package in question, which is why I can't just rely on it being auto-installed via package.json. Plus, it seems awfully redundant/slow to rebuild all your dependencies on every deployment.
I'm also new to nodejs. I've been playing with nodeJs for about 6 months from now. As for my personal experience nodejitsu is the best cloud-hosting service for nodejs. As I said so due to the following reasons.
You can simply install jitsu command line in your terminal
Your app can be deployed with all the dependencies and databases using the package.json file
They support all the types of sockets either
A very good alternative for jitsu is heroku But sometimes heroku fails with Socket.IO and stuff.
Related
Apologies for the broad scope of the question and if there is a better sub for this question please let me know.
So my requirement is that we want to make sure that the Node app that is developed by the dev team goes through testing, QA, and production without any changes, very deterministically.
The strategy is simple:
Package the Node app and digitally sign it. This package will go through testing, QA, and production.
One of the requirements of this strategy is that when we go to the production server, we do not want to do npm install.
[Pkg][1] seemed to be the silver bullet until I found out it has trouble packaging the BigNumber library.
The other alternative seemed was doing npm pack. However, I found out that when I do npm install package.tgz it actually downloaded from the internet, and not actually package everything in it.
So I would like to ask for advice on how we can move a Node app from dev to production guaranteeing that we deploy the exact same package/binary that was built originally.
I understand Docker is the obvious choice, but we are trying to find out other ways that going down the Docker route.
Thank you very much.
This question already has an answer here:
Socket.io Official Chat application not work on firebase
(1 answer)
Closed 6 years ago.
I have a simple web app made using Socket.io, Express and Node.js.
Have successfully managed to push it to Heroku via Git, and it works fine. Yay!
However, I am trying to make it work on Firebase. I seem to be having some trouble installing the NPM dependencies. I didn't use Webpack or Browserify, and for Heroku, the various NPM dependencies are installed on their server, and I can leave my node_modules folder on my local computer. The docs don't seem to talk about NPM dependencies. Am I trying to do the impossible at this time? or am I missing a few critical steps?**
I believe Firebase hosting should support the various JS dependencies because if we were just doing static websites might as well use FTP and Wordpress or something. And there's much mention of web apps on their site too..
I have tried using the Firebase CLI, and reading the docs, but it doesn't say anything about specifying my various NPM Dependencies, or even ask for my package.json contents. I am wondering if it even takes notice of the package.json file so that it will install them the way Heroku does.
Am not trying to use any Firebase features at the moment, just their hosting for now, and later on will install their login and auth, which is my final intention.
Cheers guys and thanks for your help! :D
To quote one of the Firebase developers, from this forum thread:
Firebase Hosting is for static files only. As such, you cannot run Node.js scripts on it. You should look into using something like Webtask or just run your own Node.js server on something like Google Compute Engine or Heroku.
We plan to provide a solution for you to run things like you are suggesting in the future. No timeline for it at the moment.
I am from the NodeJS/JavaScript world where I have npm and dependencies written down in the package.json. When I deploy it, I know that I just need to run npm install and all the dependencies consumed by the app will be installed.
How is it supposed to be done for a Go project? Suppose I have a source code of the app which I deploy remotely by, say, running git pull. Now, how do I make sure the dependencies are present? What I see is I need to install a package manager manually then install dependencies using it?
What's a standard way of deploying a Go app on a server?
First of all, you're indeed thinking like a JS developer. Go is compiled, and thus the proper way to deploy a Go app is not to use the source code at all - you build it on your build server, and deploy a binary. So on the server level you simply don't care anymore, the only place where you need dependencies is the build system.
Now, the standard way to do this in go is to vendor dependencies with your source, that is make sure they are included in the git repo. Another approach is the express them in a manifest file and fetch them with an external tool. These are both more reliable than the naive approach, of simply using go get in build time, fetching the current version of your dependencies (this requires no manifest file).
There are many tools for vendoring management, to name two: Godep and gb
I'm working on including Ember into an already deployed Node/Express/EJS application. I don't want to disrupt any of the existing application behavior, but instead, want build out any additional feature to the app using Ember. The server side code for these new features has already been built, and each endpoint returns the JSON format that Ember Data expects. I've been looking into Ember App Kit and Ember-cli, but I'm not sure how to include these tools into my existing directory structure, and I'm not certain if these are in face the right tools for my use case. Does anyone have any experience with this particular use case?
For example, navigating to /foo returns the existing express route that renders an ejs template, but /bar would be an Ember route that hits the api endpoint of the same name.
Use ember-cli (ember-cli.org). It's perfect for this situation as it allows you to rapidly prototype out your ember app. It even comes with an expressJS based testing suite and mocks server.
Once you are ready to incorporate it to your NodeJS, Flask, or whatever other application all static files should be available in the ember-cli dist directory.
Just don't forget to build the ember-cli project before porting via the means of ember build. After that it's just a simple matter of moving the files in the ember projects dist folder into wherever you need in your environment.
Just to embellish a bit: Ember-cli has a great work-flow for use while building your ember app. Try ember serve for a quick example. I mention this because it speaks to your question of how to incorporate this into your existing project (by project, I assume you may mean workflow). I typically will build ember projects purely using ember-cli and consider the back-end (usually a REST-API exposed via either Flask or NodeJS) a separate concern. When importing the app all I have to concern myself with is making sure my server serves the correct static dist files.
I would not recommend using the Ember App Kit (EAK) as it has been deprecated in favor of ember-cli. It really is.. much, much better.
Ok so I'm going to try to be more complete in this answer. Let's start with the isolated question - ember-cli or eak? Definitely Ember-Cli, but why?
EAK is officially heading for deprecation in favor of ember-cli.
Ember-cli produces more structured, cleaner, maintainable ember code.
Ember-cli integrates your entire ember-app workflow.
Managing all types of dependencies and assets is made simple via bower install --save and Brocfile.js edits. (See the ember-cli docus for explanation)
Now the more complicated part of the question. How do I integrate this with an existing workflow? I recently ran into this when building a webrtc-included ember app. It just so happened that this was my first real use of ember as well. So, not yet realizing the full potential of my new hammer I wrote the REST API, Backend ORM layer, the signalling service, the session cache, and built a complete CI workflow first. Then I was ready to build my ember app and ended up in your exact position.
To short circuit a long story - the lesson I learned was that I should treat my ember-cli app like a completely separate concern. What I mean here is - there's my backend (NodeJS, Apache, Nginx... whatever) and what I code here is built, tested and integrated separately. It normally even lives in its own git repository. It's a separate concern to my front end equation which typically would consist of several components itself. My I-Phone Native app would have its own workflow from build-to-test and integrate to my backend via a REST API. My Android native app another. My web app another. For all intents and purposes, in my workflow these are entirely separate workflows that only tie together when we start talking Continuous Integration.
There's a lot of arguments for why you'd want to do this. Most importantly - it scales.
The beauty of ember-cli is that it makes it fairly trivial to get a workflow for your ember app going and trivial to redeploy your app + workflow on a new box/instance. I would certainly recommend referring to the official ember-cli setup instructions but I'm going to include them here in case the URL goes bad one day:
No really, refer to the link my instructions will suck in comparison...
Deploying a new Ember App
Install NodeJS, NPM and Git (ember-cli will as a default load git going for you on new apps) on your system via sudo apt-get install nodejs and sudo apt-get install npm, sudo apt-get install git.
Note: On Ubuntu 14.04 and some other Debian systems use sudo apt-get install nodejs-legacy instead. If in doubt, use legacy. If you experience problems using the node command after install, it's definitely that you need to use nodejs-legacy. Don't bother trying to do the linking manually.
Install required node modules globally: sudo npm install -g ember-cli, sudo npm install -g bower, sudo npm install -g phantomjs
Create new ember-cli app: cd <Desired Directory>, ember new my-app-name
Now you can look at ember help to begin learning how to use ember-cli. Hint: The --dry-run flag is your friend. You'll notice that when you installed ember-cli all the scaffolding was taken care of for you. You'll see that you can add things with simple ember generate commands and they will not only create the required objects, but the test files as well. Best of all, using ember serve you can start scaffolding your app and via simple flags you can configure the test server to actually proxy and use your already-existing REST API (if you have one) or the expressJS mocks server to build a psuedo-API.
Integrating it with your larger workflow from here is a simple matter of configuring whatever tools you use (I use Jenkins and Ansible for this kind of stuff) to distribute the dist folder of ember-cli to where it should go to be served as static content (it is just a single page webapp in the end).
If you want to instead play with an existing ember-cli app that operates in an isolated workflow and already makes use of most of the goodies in order to get some familiarity - as I suspect you'll quickly realize how to fit this into whatever your current structure is - feel free to clone and play with this one here.
And so finally - to answer the more specific question of how this might fit into an existing directory structure, I would break this down into two categories. When we're talking src - I would have it in it's own "structure", separated at least by being in a separate sub-directory of its own. When we're talking built and deliverable I would include the contents of the /dist folder in whatever static web server directory you want to serve your ember app from.
EDIT: I Added some more detail - hopefully useful details below the line break. Let me know if you have more questions or if I can explain anything better.
I am facing a similar situation. I am planning to use EAK as a "prototyping tool" in a separate project folder. Then build the distribution directory from EAK using grunt dist and insert that into the assets folder of my main Node.js project.
I am building an app in nodejs and making use of a library called nodify-shopify which is used to interact with the shopify api
The library hasn't implemented all the api calls so i had to implement it myself
when i push my app to heroku, it doesn't go with the modified library in node_modules so my app breaks
What's the best way to solve this?
I have considered making a pull request, but the project seems abandoned and i need to do this fast.
I have also considered adding the node_modules to version control, but I'm wondering if there's a better way?
I need to get this done fast too, so time is of essence
thanks
Publish you version of the library somewhere it could be found during the heroku deployment. Npm supports a variety of possibilities to load node module:
(g)zip packages
git checkout