I am using node & expressjs as an app. I have problem in deploying it to serverless with dotenv (without using any dotenv plugin).
My serverless.yml is this:
functions:
server:
handler: index.server
events:
- http: ANY /
- http: "ANY /{proxy+}"
environment:
ENV_VAR1: ${env:ENV_VAR1}
and my .env.staging is this:
ENV_VAR1='ENV_VAR1 using .env.staging'
ENV_VAR2='ENV_VAR2 using .env.staging'
I use AWS Lambda and deploy it using sls deploy --stage staging. So far I can read the ENV_VAR1from .env.staging by calling process.env.ENV_VAR1 because I explicitly define it in the serverless.yml file. But that's not the case with ENV_VAR2.
My question is, how to load all the .env.{stage} file without defining the key one by one?
I hope to load it similar to something like this:
environment: ${file(.env.${self:stage})}
Thank you in advance.
As I know, there is no way, you'll have to do it one by one (you have to have an object with key-value pairs inside environment, but you are trying to give it a string - even if it is going to be interpreted as a file containing something, it is still a string when you are referencing it).
I'm not an expert, so there still might be some tricks.
Related
I have an Expo managed react native application. I created my .env file in the root of my project, installed react-native-dotenv and set up babel to use it. After a while I managed to get it to work.
I have my environment variable
ENDPOINT=http://127.0.0.1:8000/api
and i use it with
process.env.ENDPOINT
After a while I decided to test the android version of the app, so i changed the endpoint url to my LAN ip and restarted the server. The problem is that even after restarting the server, the cache and the computer, when I call process.env.ENDPOINT it keeps the first url I set.
Here's a list of the things i tried:
restarting the server
restarting the server and the cache
restarting the whole computer
change the variable name to REACT_APP_ENDPOINT as many suggested (I get undefined, it's still stuck to ENDPOINT)
empty expo cache
The strange thing is that I already changed that same variable twice (from 127.0.0.1:8000 to 127.0.0.1:8080 and back for a problem with backend) and had the same problem, but it went away by itself after a couple of minutes (and server restarts).
This time I've been trying to get it to work for 7 hours and nothing has changed.
Any idea?
I had the same issue and managed to run the app with .env changes after using the following command.
expo r -c
reference: https://github.com/goatandsheep/react-native-dotenv/issues/75#issuecomment-728055969
After a couple hundred more tests I gave up and implemented a "custom" solution, without any external library:
Switched .env files to TypeScript files (E.g. .env.development -> env.development.ts)
Set up an object named env that has all environmental variables as properties
export const env = {
VAR1: 'foo',
...
}
Imported this constant inside the application entry point (in my case App.tsx)
Inside the main constructor assign env to global.env
Use global.env instead of process.env
Not sure if this is the best practice, but solved my problem for now, works like a charm and doesn't require me to reload my application at every change. I'm a bit concerned by the security aspect of having the environment in a global variable inside a js project, so any suggestion is still welcome
I'm using serverless framework and using bitbucket-pipeline to configure CI/CD.
I have the following configuration in the serverless.yml file
provider:
name: aws
runtime: nodejs10.x
region: ${opt:region, ${env:AWS_REGION, 'ap-south-1'}}
memorySize: ${opt:memory, ${env:MEMORY_SIZE, 512}}tracing:
apiGateway: ${env:TRACING_API_GATEWAY, true}
I want to be able to pass the variables from the CLI as well as environment variables.
I have set up environment variable for AWS_REGION in the bitbucket pipeline variables but not for MEMORY_SIZE as want to use the default value.
But this gives error while running the pipeline.
Serverless Warning --------------------------------------
A valid option to satisfy the declaration 'opt:region,ap-south-1' could not be found.
Serverless Warning --------------------------------------
A valid environment variable to satisfy the declaration 'env:MEMORY_SIZE,512' could not be found.
Serverless Warning --------------------------------------
A valid environment variable to satisfy the declaration 'env:TRACING_API_GATEWAY,true' could not be found.
First, those are warnings not errors. A Serverless error would be something like:
Serverless Error ---------------------------------------
Trying to populate non string value into a string for variable ${opt:stage}. Please make sure the value of the property is a string.
This specific error happens because in the custom tag I've declared: var: ${opt:stage}-something which should be changed like:
provider:
stage: ${opt:stage, 'local'}
custom:
var: ${self:provider.stage}-something
I think in your case, you need to update region like this:
region: ${opt:region, env:AWS_REGION, 'ap-south-1'}
I couldn't reproduce the last warning though but I reckon ENV variables should be defined in bitbucket-pipelines.yml (or similar CI pipeline YAML) under args or variables and then they can be accessed using ${env:VAR}.
I'm using Vue-cli V3, In my UI I need to pass an environment variable that states if I'm in test mode or not.
I know I can use .env files to define variables, but I have a problem (that is related to our Jenkins build process) that prevents me from using it.
Is there a way to access system env variables?
Yes, you access them the same way that you would inside any normal JS file.
// server.js
const port = process.env.PORT;
console.log(`Your port is ${port}`);
vue-cli only processes environment variables with VUE_APP prefix, with NODE_ENV being an exception. Use environment variables with VUE_APP_ prefix, only then it will work. If you have variable TEST make it VUE_APP_TEST.
const test = process.env.VUE_APP_TEST
console.log(test);
I was also struggling with the same problem it almost took an hour to solve this, finally found this document.
reference -
https://cli.vuejs.org/guide/mode-and-env.html#example-staging-mode
i want to have a better overview on the phinx migration files. i want something like this
/db/migration/1.8.5/ID-2065/my_file_name_1234567890
So i can use
'migrations' => '%%PHINX_CONFIG_DIR%%/db/migrations/'. $_ENV['APP_VERSION'],
In the docs only is something like this
migrations: %%PHINX_CONFIG_DIR%%/module/*/{data,scripts}/migrations
But how can i use there maybe a param from the command line.
See you
If your using the default YAML based configuration you can try using Phinx ENV vars (PHINX_ prefix) and then use a %%PHINX_VARNAME%% replacement. Note: I haven't actually tried this before. Read more about them here: http://docs.phinx.org/en/latest/configuration.html#external-variables
Otherwise if your using a PHP-based configuration file you can definitely access the $_ENV superglobal as you have described. Just be sure to call your bootstrap/init scripts so your application version is injected.
Rob
i've been studying up on webpack for a couple of weeks now, and i've seen many examples of front end setups, and probably just this one setup for a backend.
i'm trying to setup a react app with a node back-end (e.g. express, koa, hapi, etc) where i would need at least one transpilation step for the back-end (e.g. babel, coffeescript, etc), and i think it would be nice to use webpack there for consistency vs adding another build mechanism to the mix (e.g. gulp, grunt, etc).
it would also be great if i could make changes to the backend and have the server restart automatically (watch style).
i'm wondering if the best way to do that is to basically have two distinct project setups with their own package.json and webpack.config files. maybe nest the back-end one under a server folder in the top level project folder, and use one or more script directives in the top level package.json file to control the two.
i guess i might have to proxy one server to the other to avoid CORS issues as well.
looking for any guidance from those more webpack battle tested than i for a decent setup.
regards,
tony.
Easiest way is to split this into two tasks: a build step that outputs to a folder (e.g. 'server'), then watch the output folder for changes and restart server task.
1. Build task
This can be in the same webpack.config as client building code - you can export an array and webpack will watch all of them.
Example webpack.config.js (top half is for server)
module.exports = [
{
name: 'server code, output to ./server',
entry: './index.js',
output: {
filename: './server/index.js'
},
target: 'node'
},
{
name: 'client side, output to ./public',
entry: './app.js',
output: {
filename: './public/app.js'
}
}
];
2.Watch Step
For the watch step, nodemon monitor changes and restart. Otherwise you could add a watch task to your server.js manually using something like fs.watch or node-watch.