Rollup proxy configuration causes the short circuit in the client-server - proxy

In my rollup.config.js file, I have added the following lines to proxy the requests destined to the client-server to the backend as such:
!production && localdev({
dirs: ['public'],
host: 'localhost',
port: 9876,
proxy: [{ from: '/api/*', to: 'localhost:8080' }],
}),
, but this causes the server to short circuit and give the prompt below:
yarn run v1.22.11
warning package.json: No license field
$ rollup -c -w
rollup v2.58.0
bundles src/main.ts → public\build\bundle.js...
LiveReload enabled
[2021-10-13 12:03:20] ÔÜí´©Ädev-server proxying from /api/* to localhost:8080
Done in 3.63s.
it should be working until it is explicitly interrupted, though.

change localhost:8080 to http://localhost:8080. it's bad that it doesn't shout the error.
See https://github.com/sveltejs/svelte/issues/3717
and https://github.com/pearofducks/rollup-plugin-dev

Related

DDEV: How to etablish websocket connections

I'm having some problems to establish a websocket connection to a running ddev container.
Trying wo etablish the connection per JS for example with wss://websocket.ddev.site:3000 ends always up with connection failed.
Websocket PHP library used: Ratchet (http://socketo.me/)
I tried to set the ext. container port in an own docker-compose.yaml or web_extra_exposed_ports in config.yaml but nothig worked so far.
I have managed to run a Websocket connection.
Therefore, I did an entry in config.yaml of DDEV with following Content:
web_extra_exposed_ports:
- name: ratchet
container_port: 3000
http_port: 3000
https_port: 3001
After DDEV restart, it is now possible to establish a Websocket connections with:
HTTP: 'ws://websocket.ddev.site:3000'
HTTPS: 'wss://websocket.ddev.site:3001'
My working example was build with the tutorial on http://socketo.me/docs/hello-world calling above URL with Browser console.

Lets-encrypt Error: Failed HTTP-01 Pre-Flight / Dry Run

I've set up a redbird based proxy following its README file examples.
By now I've configured single domain both for http and https and it's working well (https still using self-signed certificate).
But now I'm trying to configure it to use letsencrypt to automatically get valid ssl certificates and I'm getting stuck in following error:
{"level":30,"time":1578681102208,"pid":21320,"hostname":"nigul","name":"redbird","0":false,"1":"setChallenge called for 'exposito.bitifet.net'","msg":"Lets encrypt debugger","v":1}
[acme-v2] handled(?) rejection as errback:
Error: Error: Failed HTTP-01 Pre-Flight / Dry Run.
curl 'http://exposito.bitifet.net/.well-known/acme-challenge/test-cf55199519d859042f695e620cca8dbb-0'
Expected: 'test-cf55199519d859042f695e620cca8dbb-0.MgLl7GIS59DPtPMejuUcXfddzNt8YxfLVo5op670u8M'
Got: '<?xml version="1.0" encoding="iso-8859-1"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
<title>404 - Not Found</title>
</head>
<body>
<h1>404 - Not Found</h1>
</body>
</html>
'
See https://git.coolaj86.com/coolaj86/acme-v2.js/issues/4
at /home/joanmi/SERVICES/redbird_domains/node_modules/acme-v2/index.js:49:10
at process._tickCallback (internal/process/next_tick.js:68:7)
As far as I understand, this is telling me that Lets Encrypt is trying to access to the url http://exposito.bitifet.net/.well-known/acme-challenge/test-cf55199519d859042f695e620cca8dbb-0 using the following command:
curl 'http://exposito.bitifet.net/.well-known/acme-challenge/test-cf55199519d859042f695e620cca8dbb-0'
...and that it is getting which seems a 404 HTML Error page which I have no clue wherever it could come.
And, in fact, executing that curl command or just pasting that url in my browser (you can try it: I left the server running), I get the given Expected string so, from my point of view it seems like if my configuration were correct but, for some reason, Lets Encrypt's servers were reaching another server (either because of wrong routing or DNS).
But on the other hand, I suppose it's more probable that I've done something wrong in my configuration.
Here I paste my whole script (ports 80 and 443 are redirected to 1080 and 1443, respectively, through iptables because the script is run by non privileged user):
const Redbird = require("redbird");
const proxy = Redbird({
port: 1080,
xfwd: false, // Disable the X-Forwarded-For header
letsencrypt: {
path: __dirname + '/certs',
port: 9999
// LetsEncrypt minimal web server port for handling challenges.
// Routed 80->9999, no need to open 9999 in firewall. Default 3000
// if not defined.
},
ssl: {
http2: true,
port: 1443, // SSL port used to serve registered https routes with LetsEncrypt certificate.
}
});
proxy.register('exposito.bitifet.net:9999', 'http://localhost:8001', {
ssl: {
letsencrypt: {
email: 'xxxxxx#gmail.com', // Domain owner/admin email
production: false,
// WARNING: Only use this flag when the proxy is verified to
// work correctly to avoid being banned!
}
}
});
proxy.register("exposito.bitifet.net", "http://localhost:8001");
Any clue will be welcome.
Thanks.
SOLVED!!
Many issues were involved at the same time (despite my lack of experience with either redbird and letsencrypt.
The magic 404/Not found page: I guess it came from a lighttpd server that seems to had been preinstalled in my VPS.
Port 80 was redirected via iptables but I suppose in one or other configuration tweak I could have redirected incoming requests to localhost's port 80 (which is not redirected).
My redbird missunderstanding: Looking at examples in its README file, I thought redbird were kinda "multi- reverse_proxy" in the sense that you could redirect http and https requests with single redbird instance.
But I finally realized that the (maybe not so well named) port option which is, in fact, an http port, serves only to configure a built-in unconditional http->https redirector (of which I already had read about, but I thought it were optional).
The actual underlying issue: If your DNS have DNSSEC activated, you should define a CAA register in it pointing to letsencrypt.org.
At the moment I disabled DNSSEC instead because my provider's control panel doesn't allow me to create such register.
I discovered it while trying to get the certificates through certbot (sudo apt-get install certbot which I must say that, If I had known about it before, I wouldn't had care about trying redbird's letsencrypt integration.
It is much more verbose (while redbird is more like a black box when errors arise) and pointed out that I needed the CAA register.
Here the notes I took about it (in case anyone could be interested):
Free SSL Certificates with Certbot
Install certbot:
sudo apt-get install certbot
Create:
sudo certbot certonly --manual --preferred-challenges http -d <domain>
Renew:
sudo certbot renew
Caveats:
DNSSEC
If your DNS server has DNSSEC enabled, you will need to add a CAA
register pointing to letsencrypt.org.
...and your DNS provider my not allow to create it (at least I
couldn't with CDMON. Also not -yet- complained).
production = false is for other kinds of testing: I read that if you put in to true while testing, you may be banned from letsencrypt if you perform too many requests.
Setting it to false you can test redirections, but you will see still errors regarding letsencrypt even you could navigate without a secure certificate (I think kinda self-signed is provided to allow testing). So don't expect a valid one.
ssl port is used for redirection: Not a (big) issue, but if you specify ssl port other than 443, built in redirector will unconditionally redirect you to that port.
Running redbird as root and using standard (80 and 443) ports works fine. But if you, like me, want to use an alternative ports in order to execute redbird with non privileged user, you will get redirected to that alternative port instead of 443 (Even having it redirected through iptables).
Here my (almost*) final redbird script:
const Redbird = require("redbird");
const proxy = Redbird({
port: 1080,
xfwd: false, // Disable the X-Forwarded-For header
ssl: {
port: 1443,
},
letsencrypt: {
path: __dirname + '/certs',
port: 9999,
// LetsEncrypt minimal web server port for handling challenges.
// Routed 80->9999, no need to open 9999 in firewall. Default 3000
// if not defined.
},
});
proxy.register('exposito.bitifet.net', 'http://exposito.bitifet.net:8001', {
ssl: {
http2: true,
letsencrypt: {
email: 'xxxxxx#gmail.com', // Domain owner/admin email
production: true,
// WARNING: Only use this flag when the proxy is verified to
// work correctly to avoid being banned!
},
}
});
(*) I still need to fix the explicit-port redirection issue (5), because I don't want to run redbird as root. But I know is possible to allow uses to listen given ports. Even I would probably better try to patch redbird in order to allow specifying listen and redirection ports separatedly.
EDIT: It is already implemented (and documented) using the (optional) option redirectPort in ssl section. Just added redirectPort: 443 and job done!!
EDIT 2: For the sake of completion, there still was another issue I struggled with.
To get things working I finally configured the redirection to the http port instead of https one.
That is: Incomming https requests gets redirected to my application http port.
It seems weird but it works. At least if you don't need any exclusively https feature such as push notifications (which I plan to use in the future).
But its implies to open an http server at least on localhost. Which isn't a major issue now (this is only a playground server) but I plan to use redbird at work to proxy multiple domains to different servers so that would had forced us to open http at least in our DMZ vlan (which is an additional risk that is better to avoid...).
When I tried redirecting to https I got the DEPTH_ZERO_SELF_SIGNED_CERT error.
Ok: This is telling me that redbird (or node) does not trust my original (self signed) certificate. I know there is an option to tell node to accept those certificates. But maybe it is not the way to go...
So I configured my application to use the same certificate that redbird is obtaining through letsencrypt.
But then I got this other error:
UNABLE_TO_VERIFY_LEAF_SIGNATURE
Researching a bit I found this StackOverflow answer that explains how to get all root and intermediate certificates trusted by Mozilla and make node to trust them.
So, at the end, what I did was:
Installed node_extra_ca_certs_mozilla_bundle package:
npm install --save node_extra_ca_certs_mozilla_bundle
Prepended NODE_EXTRA_CA_CERTS=node_modules/node_extra_ca_certs_mozilla_bundle/ca_bundle/ca_intermediate_root_bundle.pem to the start command in the package.json's scripts section.
Updated my redbird script to point again to the https (protocol and) port:
proxy.register('exposito.bitifet.net', 'https://localhost:4301', {...]);
Here my final redbird configuration:
const Redbird = require("redbird");
const proxy = Redbird({
port: 1080,
xfwd: false, // Disable the X-Forwarded-For header
ssl: {
port: 1443,
redirectPort: 443
// key: "/etc/bitifet/exposito/ssl/private.key",
// cert: "/etc/bitifet/exposito/ssl/public.cert",
},
letsencrypt: {
path: __dirname + '/certs',
port: 9999,
// LetsEncrypt minimal web server port for handling challenges.
// Routed 80->9999, no need to open 9999 in firewall. Default 3000
// if not defined.
},
});
proxy.register('exposito.bitifet.net', 'https://localhost:4301', {
ssl: {
http2: true,
letsencrypt: {
email: 'xxxxxx#gmail.com', // Domain owner/admin email
production: true,
// WARNING: Only use this flag when the proxy is verified to
// work correctly to avoid being banned!
},
}
});
And here my package.json file contents:
{
"name": "redbird_domains",
"version": "0.0.1",
"description": "Local Domains Handling",
"main": "index.js",
"scripts": {
"test": "echo \"Error: no test specified\" && exit 1",
"start": "NODE_EXTRA_CA_CERTS=node_modules/node_extra_ca_certs_mozilla_bundle/ca_bundle/ca_intermediate_root_bundle.pem node ./index.js"
},
"author": "Joanmi",
"license": "GPL-3.0",
"dependencies": {
"node_extra_ca_certs_mozilla_bundle": "^1.0.4",
"redbird": "^0.10.0"
}
}

grpc-go over https: failed rpc error: code = Unavailable desc = transport is closing:

Note: This is running in containers in Kubernetes.
I have successfully done this very short description: https://knative.dev/docs/serving/samples/grpc-ping-go/
Success:
2019/05/08 13:43:56 Ping got hello - pong
2019/05/08 13:43:56 Got pong 2019-05-08 13:43:57.646935391 +0000 UTC m=+1.661567121
But if I run through 443 from a Gateway on knative setup for https, it does not work:
docker run -ti --entrypoint=/client docker.io/{username}/grpc-ping-go \
-server_addr="${SERVICE_IP}:443" \
-server_host_override="${SERVICE_HOST}" \
-insecure
2019/05/08 13:50:28 &{0xc00012e000}.Ping failed rpc error: code = Unavailable desc = transport is closing:
The client code from the sample, and the server code.
The server is not listening for tls, but the connection to the server is over https.
Just to make sure, I know the https is worning from a simpel hello-go text reply.
In your server code, you are not listening on port 443, so this is most likely the reason your example isn't working.
If you want to keep using http and not https, then your code is working just fine.
If you want to get it working with TLS, this overview is a pretty good one.
To get port 80 to redirect to port 443 (I highly recommend it if you are using https), see this SO post.

Cannot connect to ghost server on ec2 instance

I am following the basic Ghost server installation on an ec2 instance, so far I can run ghost server via npm start and I can see that ghost server is up and running:
Ghost is running...
Listening on 127.0.0.1:2368
Url configured as: http://54.187.25.187/
Ctrl+C to shut down
Here is the ghost config config.js:
// ### Development **(default)**
development: {
// The url to use when providing links to the site, E.g. in RSS and email.
url: 'http://54.187.25.187/',
database: {
client: 'sqlite3',
connection: {
filename: path.join(__dirname, '/content/data/ghost-dev.db')
},
debug: false
},
server: {
// Host to be passed to node's `net.Server#listen()`
host: '127.0.0.1',
// Port to be passed to node's `net.Server#listen()`, for iisnode set this to `process.env.PORT`
port: '2368'
}
At the end, I cannot access to anything when I type http://54.187.25.187:2368 on the browser. I really appreciate guidelines on how to setup ghost properly.
EDIT: The problem is solved already, it was a EC2 SG issue that ports remained closed after I have set them to open.
For Amazon EC2 we have found you need to change the port to 0.0.0.0
http://www.howtoinstallghost.com/how-to-setup-an-amazon-ec2-instance-to-host-ghost-for-free-self-install/

How can I configure yeoman (gruntjs) to run as HTTPS?

I have recently been working with Yeoman (http://yeoman.io/) and now would like to set up my local environment to handle HTTPS requests, so that I can have it handle callbacks from OAUTH providers.
Under a non-Yeoman/grunt setup I was able to get node.js configured to handle HTTPS in a following a similar path as directed in this question (How to create an HTTPS server in Node.js?).
Looking at the gruntJS repo on github it appears this has been added as a feature (https://github.com/gruntjs/grunt-contrib-connect/pull/15) but I still am unclear as to where I set the appropriate options.
grunt.initConfig({
connect: {
server: {
options: {
protocol: 'https',
port: 8443,
key: grunt.file.read('server.key').toString(),
cert: grunt.file.read('server.crt').toString(),
ca: grunt.file.read('ca.crt').toString(),
passphrase: 'grunt'
}
}
}
});
see this commit

Resources