tools-and-devops

[sturdy seal]

We aren't just handling accounting now, though.

[thorny shell]

heh. imo there aren't any

there are only horribly-complex ways, and less-complex ways

[sturdy seal]

So python across the board was decided and I was hired. 😄

[thorny shell]

I still don't understand your setup, but if you're using an EC2 instance and nginx, then you should look into "lego" -- it's what I use

https://github.com/go-acme/lego

[rapid sparrow]

cases like that make me scarried ever leaving current company. Sure... our code is written may be far from perfect way. (in python too)
But still we uphold a lot of things for dev comfort
It sounds... impossibly challenging to maintain system like you have.

[thorny shell]

that's the first thing I found that was simple enough for me to understand, and get working

[heavy knot]

the installation says to use docker, does it still work for AWS?

[grim osprey]

Let'sEncrypt seems to be good in general, but I haven't used it so can't say.

[heavy knot]

suggestions on installing docker?

[thorny shell]

It doesn't "say to use docker". It says you can use docker.

I am using it, without docker, but on AWS

it's just a program you run

[heavy knot]

nvm i am big brain

[indigo zenith]

It says Docker isn't running. It's already installed

[heavy knot]

lego is installed, what to do with it?

wait oh no, is there a way to get a free certificate?

also is this good pseudo code running the certificate?
lego --email="[e-mail]" --domains="example.com" --dns="route53" --path="/path/to/store/certificates" run

[indigo zenith]

Let's Encrypt

[grim osprey]

https://letsencrypt.org/getting-started/

[heavy knot]

is AWS with or without shell access?

[grim osprey]

If you're at a company, maybe IT or some other department can provide you with a certificate. If it is a completely internal service, usually certificates are issued from the domains CA server. If it's external, you need to demonstrate that you own the domain.

In order to get a certificate for your website’s domain from Let’s Encrypt, you have to demonstrate control over the domain.

[crimson spruce]

https://caddyserver.com/

[heavy knot]

I own a domain through NameCheap, how do i demonstrate it?

[heavy knot]

should i run cert bot as well?
sudo apt-get update
sudo apt-get install certbot

[indigo zenith]

It depends on what resources you are asking about. You have shell access to EC2 instances, probably some other things

[indigo zenith]

Certbot is generally a good idea unless you're using something else that takes care of it

[sturdy seal]

We're working toward it. I know for sure we won't use Docker here, though. I'm not sure why. Someone that's been here longer than me hates it, I think.

[grim osprey]

https://letsencrypt.org/getting-started/ Read through the documentation. You're the one that's going to maintain this over time and it would be much more helpful to future you if you read the documentation. Certs have an expiry-time, so you're going to want to be familiar with the process of requesting/renewing them. It's not a one-time thing. When the certificate expires, the service is (likely) going to be inaccessible.

If there is a particular step you have issues with I can try to assist.

https://certbot.eff.org/instructions

Seems they have pretty good selection of instructions for different OS'es and webservers. Cool

[thorny shell]

read the docs

[heavy knot]

where is the documentation? also should i follow the ACME certbot instructions?

[grim osprey]

Yes. Under the section where it says With Shell Access you have the line Visit the Certbot site to get customized instructions for your operating system and web server.

[heavy knot]

what is wrong with the domain?

https://tenor.com/view/coding-typing-pc-laptop-power-gif-21599707

[heavy knot]

Learning Documentation one sec

8000 is not a port you linked to http in your incoming ports, is it?

[heavy knot]

it is

if it returns an HTML document but it looks different check the DevTools->Network Tab for errors in the css requests.

[heavy knot]

i am following the cert bot instructions for nginx on pip, it is not going well

[grim osprey]

Doesn't seem to be an issue with the domain. It looks like you got the certificate and private key /etc/letsencrypt/live/harmonyapp.org/*.

The installer looked through the nginx-configuration and didn't find a serverblock listening to "harmonyapp.org", so it exited. It couldn't figure out where to put the configuration lines for the certificate.

[heavy knot]

Should i manually update the nginx configuration to include the necessary server block for harmonyapp.org and then configure certbot to use a configuration?

[grim osprey]

Yes, that's probably the easiest solution

[heavy knot]

wait heck what is the command for editing the conf.d?

wdym? isn't that vim?

i enables editing, :wq for write & quit

[heavy knot]

do i use http://harmonyapp.org or do i use www.harmonyapp.org

[heavy knot]

thank you

`
server {
listen 80;
server_name harmonyapp.org www.harmonyapp.org;

# Redirect HTTP to HTTPS
location / {
    return 301 https://$server_name$request_uri;
}

}

server {
listen 443 ssl;
server_name harmonyapp.org www.harmonyapp.org;

ssl_certificate /etc/letsencrypt/live/harmonyapp.org/fullchain.pem;
ssl_certificate_key /etc/letsencrypt/live/harmonyapp.org/privkey.pem;

#What do i add here?

}
`

[heavy knot]

any thoughts on proxy code?

i haven't edited nginx, probably Netherwolf can help you

i'd say this info should be easy to search though

paging doctor Netherwolf

[grim osprey]

Life happened so I'm a bit busy.

See https://docs.nginx.com/nginx/admin-guide/web-server/reverse-proxy/

You want to proxy all traffic from nginx to your webserver:port, otherwise nginx will think that it is responsible for serving the webpages.

[edgy basin]

trying to setup fuzzing and am confooz

[edgy basin]

fuzzing make me oof

[edgy basin]

okay so im trying to write fuzz tests

and the thing im fuzzing against

uses randomness in the function

😔

i feel like im just ranting here

[heavy knot]

How do I get the nginx to see the ssl certificate?

[gentle solstice]

Does the nginx user have read access to that folder?

[heavy knot]

i will assume not since i do not remember doing that

[thorny narwhal]

did You guys ever automated configuration of the user profile and computer settings in windows with python ?

[gentle solstice]

That's best done with powershell

[thorny narwhal]

well i do not know powershell and i would not like to use it

[thorny shell]

and yet it's ideally suited for maniuplating Windows, wheras python isn't 🤷

I myself hate and fear WIndows, and yet ... I have a gruding respect for Powershell.

[gentle solstice]

powershell lets you cd into HKLM:\

registry keys are treated as files in powershell

[thorny shell]

stuff like that, ya

[gentle solstice]

basically, microsoft saw linux and /sys and /etc/system and decided that was a good idea.

[arctic flicker]

I have opened a folder in VSCODE and want to add it to github. Why is it stuck like that? With the commit button greyed out

[arctic flicker]

what does staging even mean?

[willow pagoda]

https://git-scm.com/book/en/v2/Git-Basics-Recording-Changes-to-the-Repository
staging is simply picking which new/changed/deleted files you want in your commit, normally done with the git add command

when it prompts you with a file to edit like that, it wants you to type something appropriate, save, and then close the file to perform an action

[arctic flicker]

Ah I see!

Thank you

[burnt ginkgo]

Why do we use -e .

[astral apex]

https://setuptools.pypa.io/en/latest/userguide/development_mode.html

When you install your package, it creates a copy of your code, so every time you edit it you have to install it again

Editable links to your current folder instead of making a copy
So you don’t need to install over and over again

[rapid sparrow]

they probably have some obscure folder/package structure

if u use flat structure. when at root folder u have package_name/__init__.py, u usually don't need -e because current folder is already... sensed as installed package

and if need to reuse your library in other application in real time edit, usually -e is not very convinient too because it does not propagate to docker environments.
Easy to use overshadowing trick, where u just mount volume your library into app project folder 😁 real time editing too then. compatible with docker (its own docker volumes could be utilized for this hack)

-e pretty much useful only if u develop applications straight in a current machine filesystem without extra complications

🤔 technically nothing really forbids using -e mod with docker, if volume forwarding appropriate library project to it

[gentle solstice]

e for editable

[rapid sparrow]

question araises though, what is the benefit of using -e if u can just volume/mount forward library folder itself into end project folder and it will be already working as editable lib

[arctic flicker]

say I have an executable c:\python\python.exe
how do I use pip to install to that specific python instance?

[rapid sparrow]

u are supposed to have active pip in current terminal...
...but if u wish to ensure... u could run it as c:\python\python.exe -m pip install smth like this

[arctic flicker]

thank you!

[heavy knot]

stil no idea on how to fix this or what I should change in the config

[mint cipher]

Can anyone share their terraform script that deploys Django made in Docker to GCP or AWS?

[heavy knot]

python -m pip install ... @arctic flicker

[heavy knot]

basically creates a symbolic link to your project i.e adds it to the path

SO has a nice explanation https://stackoverflow.com/questions/42609943/what-is-the-use-case-for-pip-install-e

[heavy knot]

how do I save and test a config in AWS?

[rapid sparrow]

this answer is too broad in approaches

1. u could just open it with nano, it has obvious simple interface

2. u could connect with program like vscode, using plusin remote ssh stuff and edit in real time through vscode (assuming u know how to use ssh)

3. u could learn how to use 👀 probably it is vim? how to edit and save. i never bothered to remember. too over complex unintuitive program

4. you could learn (i like the best option) running your nginx at your local environment for more rapid feedback
   by just using Linux as main OS, or at least windows with WSL2
5. you could apply local code to modify remote server by using things like ansible (or other configuration management tools), including terraform and etc

[heavy knot]

new error, I just need to chang redirect settings, idk

[heavy knot]

do i have to install these on the server? is the remote server a different server than the AWS server?

[rapid sparrow]

1. if u wish to use nano, u have to install it at remote sever(AWS server)
2. vscode should be present only at your local dev machine
3. should be already present there vim at remote
4. well, definitely installing to local environment different OS
5. this can work from being installed at local machine only, but some of them require modifications at remote machine as well 😅

[heavy knot]

can these instructions help me fix the bad request error? because i fixed the server name error

[heavy knot]

also thank you as always oh honorable and knowledgeable Darkwind The Dark Duck : ]

[rare anvil]

bro theyre just ways to edit your nginx file and iterate quicker

like if u take some time to understand what he said...

[wet plover]

i want these full number can tell me anybody

[heavy knot]

new error unlocked

you have an extra } don't you (edit, meant need an extra })

i dont think so?

i removed the }

and it gave me this

when i readd the } it gives me this still

there is a bracket mismatch, one } missing it indicates.

why aren't you using vscode with remote ssh?

chat gpt may find the error

[heavy knot]

I want to set up the server on AWS to keep it up with nohup

[lethal pond]

pre-commit is slow and is extra work to keep it's version synced with my pyproject

so I'm looking to find a decent replacement

something that can run ruff and mypy for me

oh yea pre-commit's mypy is an issue too so that's another reason

I use poetry

[thorny shell]

I have my editor do "ruff format" on every save; it's fast enough.

I only do "mypy" at the same time I do unit testing

[lethal pond]

am i doing something wrong then?

I have the ruff extension but it's not doing anything from what I can see

[thorny shell]

on every keystroke? That seems like overkill

Anyway that looks like Visual Studio Code, but I don't use that, so can't help with it

[gentle solstice]

Just run pre-commit autoupdate

[heavy knot]

do i have to create the www, i am trying to set up nginx

[lethal pond]

Oh no that was just what the default showed. Setting it on save had the same issue

I don't even see a "format" on save

[heavy knot]

how do i use socket?

[gentle solstice]

varwww is for apache httpd

[gentle solstice]

Check the directory by looking at /etc/nginx/sites-enabled/default

It should look something like this. ```cfg
server {
listen 80 default;
server_name localhost;

    access_log  /var/log/nginx/localhost.access.log;

    location / {
            root   /var/www/nginx-default;
            index  index.html index.htm;
    }

[heavy knot]

wait do i not sudo vi /etc/nginx/sites-enabled/default?

[gentle solstice]

Oh, this is amazon linux.

Check /etc/nginx/conf.d

it's a directory

[heavy knot]

how do it edit it?

[gentle solstice]

use ls to list the files

There's nothing in there...

Ok, check /etc/nginx/nginx.conf

[heavy knot]

oh no, um recover, edit anyway?

[gentle solstice]

Do you have another instance of vi(m) running?

[heavy knot]

i think so, maybe, how do i check?

[gentle solstice]

run pkill vi

[gentle solstice]

It's running as root

use sudo

[heavy knot]

seems to work

[dire rivet]

Does anyone know any good frameworks?

I normally use tkinter

[gentle solstice]

pyside6

wxpython

[dire rivet]

Thanks

[warm rock]

Having a little trouble with a project. I made a script to webscrape product information from wholefoods site, but in the all products page there is a Load More button at the bottom and my current script using the Requests library I'm not sure how to get around this to pull all the data. Do I need to switch to selenium or is there a workaround?

[median saddle]

Hello I am using lazyvim but I can't run my Python code due to this error:
Error invoking 'python_execute' on channel 3 (python3-script-host): Traceback (most recent call last): File "<string>", line 1, in <module> NameError: name 'caesar_cipher' is not defined

This is my caesar_cipher.py :

`#text = input("Enter your message: ")
#cipher = ''
#for char in text:

if not char.isalpha():

continue

char = char.upper()

code = ord(char) + 1

if code > ord('Z'):

code = ord('A')

cipher += chr(code)

#print(cipher)

Caesar cipher - decrypting a message.

cipher = input('Enter your cryptogram: ')
text = ''
for char in cipher:
if not char.isalpha():
continue
char = char.upper()
code = ord(char) - 1
if code < ord('A'):
code = ord('Z')
text += chr(code)

print(text)`

So what is wrong about this code and why can I not run it in lazyvim? In vscode it works fine, but not in lazyvim..

[indigo zenith]

The correct solution is to use an API instead of scraping. I don't think Whole Foods has a public one, but Instacart does, so hopefully that has what you need.

Selenium is technically an option but usually against terms of service. If you go that route you won't get help here per the #rules

[warm rock]

Sorry no that doesnt help. Instacart has different prices on their app than what whole foods has in store. Please help with my question

[heavy knot]

yes, you can't handle interactivity with a static /text version of the page

[warm rock]

Can you help via Requests using the JSON link on the network page

[heavy knot]

it's a lot more hassle, just use a lib that's for the purpose, try selenium.

[warm rock]

It ends in ...offset=0 and the number increases by 60 each iteration the iterations part im just having trouble figuring out

[heavy knot]

yeah it takes time to learn that. that's why you start easy.

[warm rock]

Tells me the count of products 11348 on the site so I know I have to iterate the offset= number +60 until it reaches a number just above 11348

That's why im asking for help/guidance

[heavy knot]

i get it, but try selenium. in fact chatgpt probably can help a lot better

[warm rock]

I would really prefer to use requests I just need a nudge in the right direction

There are 190 load more screens to reach the bottom of the all products list I dont want to click load more 190 times via selenium

[heavy knot]

Requests to the json api endpoint is ok. Assuming they allow it

but normally only requests initiated from the page's origin (the main domain name) are replied to, to avoid people stealing I guess @warm rock

[heavy knot]

new error? how to get websocket to work idk

[hollow storm]

idk what you mean with the websocket thing

this error is probably because epel 7 is EOL

[heavy knot]

how do it update it?

[hollow storm]

how did you arrive at the conclusion that you should run that command? the yum install amazonlinux.com etc

if im not sure whats your end goal it's hard to help

[heavy knot]

i have a Django project that i was to run on an AWS server, the project does not work because they server lacks the packages to make the websocket the project uses work. idk how to get the packages?

[hollow storm]

these are the rough steps i would take to do this:

• make sure you have python and a reasonably modern one (you can check with python -V, i would say anything higher than 3.7 is fine)
• create a virtualenv, preferably in the same folder as your code (python -m venv venv for example)
• activate the virtualenv (source venv/bin/activate)
• im gonna assume you dont have a requirements.txt. if you do, ignore the next 2 steps and do pip install -r requirements.txt instead
• try to run the program. you'll get an error like Could not find module django
• run pip install django and keep trying to run the program and installing what's missing until it works

once you have it running, you need to figure out a way to leave it running even when your terminal dies. the easiest way is probably nohup, there are some fancier alternatives like creating a systemd service or using pm2

[heavy knot]

is the v suppose to be capital

[hollow storm]

yes

in any case you have the version there

it's 3.9

[heavy knot]

also i have a venv already look at the side

[hollow storm]

that works too. idk your setup

[heavy knot]

is it already activated, is there a way to check that?

also requirements did not work?

[heavy knot]

Running the server does not run the Django project, instead it runs the ngxin conif. How do I change that?

[spice field]

It looks like you're not connecting to the app on the port 8000

nginx looks like it's serving port 80

You would most likely need to configure a reverse proxy to serve port 8000 to the webpage

At least that's what i've always done for things that use ports

Ex in your nginx domain config, your server block might look something like this:

server {
    listen 80; # Listen on Port 80 for IPv4 Addresses
    listen [::]:80; # Listen on Port 80 for IPv6 Addresses

    server_name harmonyapp.org www.harmonyapp.org; # The  domain of your website to include the www subdomain
        
    location / { 
        proxy_pass http://127.0.0.1:8000; # The host:port combo Nginx should serve to the website on Port 80
        include proxy_params;
    }
}

This is configured for a non-SSL certified website. If you end up getting an SSL certificate, make sure to switch the listen port 80 to 443

While this guide is used for Ubuntu, it should apply to most flavors of *nix

[heavy knot]

how do i configure my nginx domain config?

so run django on port 80
then reverse proxy it using proxy_pass
and then running django and daphne at the same time, turn them into systemd services

[quick lance]

i need help man

i have an error

i am a beginer python

[spice field]

If you follow that guide it should tell you how to do it

ex nano /etc/nginx/sites-available/harmonyapp.org and then paste what I put above. Then hit CTRL + O and then enter to save and then CTRL + X to quit

[brazen forge]

nano's key bindings are so weird

[spice field]

Surround it in quotes or apostrophes

such as

first = float(input('What is your first number?'))

[spice field]

I personally can't stand VI lmao

I know it's super powerful, but I just don't find it to be efficient for me

[heavy knot]

I am having some trouble with the guide, what does it mean does not exist?

[spice field]

Looks like you didn't edit the command

on the ln -s ... command

You left it as your_domain

[heavy knot]

but then i reran it correctly

[spice field]

So it's looking for a file that has that filename

[heavy knot]

or did i not?

[spice field]

Yes, however nginx still things the other one is enabled

you mispelled the second time too though it looks like

harmonapp.org instead of harmonyapp.org

Oh i see

So open nginx.conf using nano /etc/nginx/nginx.conf and find that line that has your_domain

I don't use nginx on my server so I'm not sure what it's going to look like

[heavy knot]

here?

[spice field]

Real quick, press CTRL + X

then type ls and send a screenshot

[spice field]

type rm -r sites-enabled/ right there

[heavy knot]

it worked?

[spice field]

Looks like it

Thought it was a directory

@heavy knot can you type ls /etc/nginx/sites-available/

[heavy knot]

thats not confusing

[spice field]

Alright now
ln -s /etc/sites-available/harmonyapp.org /etc/nginx/sites-enabled/

[heavy knot]

worked?

[spice field]

Yup, ln commands don’t give any output

Now try a systemctl restart nginx

Are there any errors?

[heavy knot]

where in the guide are we?

[spice field]

We kinda went off track of that to troubleshoot

You add some errors and typos and so we’re just kinda fixing that right now

[heavy knot]

ran the journalctl

[spice field]

cat /etc/nginx/nginx.conf

Might’ve enabled it twice and nginx is having a fit

[spice field]

Are you able to copy the contents of nginx.conf and put it in https://paste.pythondiscord.com/

[heavy knot]

i could do such a thing : ] https://paste.pythondiscord.com/5MIQ

[spice field]

Haha thanks I’m on my phone now so trying to help the best I can

Okay remove all of this

    server {
        listen       80;
        listen       [::]:80;
        server_name  harmonyapp.org;
        root         /usr/share/nginx/html;

        # Load configuration files for the default server block.
        include /etc/nginx/default.d/*.conf;

        error_page 404 /404.html;
        location = /404.html {
        }

        error_page 500 502 503 504 /50x.html;
        location = /50x.html {
        }
    }

Lines 39-55

[heavy knot]

i love you

[spice field]

We’re doing this because we have enabled the config in the sites-enabled directory. So we don’t want it in the main file

The line “include /etc/nginx/sites-enabled/*” is what’s letting nginx know we have that configuration available, just in another place

[heavy knot]

i got rid of the lines

i saved and exited the conf

[spice field]

Restart nginx again

See if it errors out again

[heavy knot]

i see an error

its a new errors?

[spice field]

Can you run that journalctl command?

[heavy knot]

https://paste.pythondiscord.com/PTUA

lmk if that works, i am not sure paste enjoys the journalctl command

[spice field]

ls /etc/nginx/sites-*

[spice field]

Alright let’s start over real quick.

rm -f /etc/nginx/sites-enabled/*

[heavy knot]

oh no

[spice field]

rm -f /etc/nginx/sites-available/harmonyapp

Run it as sudo

[heavy knot]

heck?

[spice field]

Use sudo on this one too

[heavy knot]

happy!

[spice field]

Beautiful

Let’s confirm nginx can be happy now. Restart it once more

It shouldn’t have anything conflicting

[heavy knot]

: ]

[spice field]

Alright, let’s confirm your original config. Let’s cat /etc/nginx/sites-available/harmonyapp.org

[spice field]

Looks fantastic

[heavy knot]

no u

[spice field]

Now let’s do that ln command again:

ln -s /etc/nginx/sites-available/harmonyapp.org /etc/nginx/sites-enabled/

[spice field]

This is creating what’s called a symbolic link. Anytime you edit the file in sites-available, it will make the changes to the file in sites-enabled

[heavy knot]

symbolic? what is this 6th grade english class? jk lol

[spice field]

Now let’s try restarting nginx once more.

[heavy knot]

womp womp?

[spice field]

Man wtf lmao

[heavy knot]

ran journalctlhttps://paste.pythondiscord.com/F7YQ

[indigo zenith]

It's telling you where to check the log to see what's wrong. Any clues there?

[spice field]

Oh lmao

I see now lol

[spice field]

You didn’t put the host:port combo

[indigo zenith]

What stands out to me is: invalid URL prefix in /etc/nginx/sites-enabled/harmonyapp.org:8

[spice field]

Exactly

Need to put 127.0.0.1:8000

[heavy knot]

in the nginx conf file?

in the TPLS settings thing?

[spice field]

No in the sites-available/harmonyapp.org file

Look for the line “proxy_pass”

See if you can find the issue

[heavy knot]

wait wrong file?

oh sites-availble

one sec

https://tenor.com/view/blues-cluesand-you-blues-clues-gif-14902469

[heavy knot]

this me

i found it!

should this be 127:something something

or is this my private or public ip?

[spice field]

Should be http://localhost:8000

This is the app on the server you want to serve

[heavy knot]

is my "localhost" 0.0.0.0?

[heavy knot]

how do i vertify this?

[spice field]

every host has this

It’s a standard/default address that points to your servers internal self if that makes sense lol

When you start the app it says it’s being hosted on 0.0.0.0:8000 right? You should be able to use 127.0.0.1:8000 or localhost:8000 to reference this

Since you’re not reaching out to a remote server

[heavy knot]

oh so i do not have to replace localhost in http://localhost:8000/ ? wack

[spice field]

No, keep it as localhost

[heavy knot]

saving and exiting

[spice field]

Remove the last slash on that line

I think discord automatically put that in

So just …:8000;

[heavy knot]

done?

i did the restart command?

time for proxy params?

[spice field]

Everything should be good

You don’t have to redo the ln command, it already knows where to look.

If you restarted, it should be good.

[heavy knot]

but error message?

but it works when i run sudo?

does that mean it works yay?

[spice field]

Yup @heavy knot

No errors should be good

[heavy knot]

I want to run django on port 80
then reverse proxy it using proxy_pass
and then running django and daphne at the same time, turn them into systemd services

how we gotten steps 1 and 2 done?

[spice field]

The Django?

Is the app right

[heavy knot]

yes the django project is an app i made

it runs locally but the end result means it should work on a server

[spice field]

Yeah, the reverse proxy is what will make it be served on port 80

[heavy knot]

i think the daphne i need does not work, thoughts? how to enable http/2 what is that?

[rare spire]

Anyone in this Discord an expert in Prefect/Prefect Cloud?

[spice field]

not sure what daphne does or is, or what the http/2 package is for

forgive me if my responses are delayed im in a class rn

[heavy knot]

np

[heavy knot]

ok i am back i got the package and now daphne works, do I have to log into the server through daphne or do i have to use the python manage.py runserver command?

[spice field]

At this point I'm not sure what else you have to do

nginx and reverse proxy is as far my knowledge goes for what you're working on - i don't know anything about djano or daphne

[heavy knot]

any idea how to solve dependecy issues, i am following this guide to deploy on gunicorn

https://youtu.be/tnJBalPzeAo

[spice field]

what version of python are you on?

[gentle solstice]

is ufw a requirement? Can you use firewalld instead?

That should be installed by default on amazonlinux

amazonlinux is based on redhat, so will use the redhat defaults.
https://www.redhat.com/sysadmin/firewalld-linux-firewall

[spice field]

Thanks for teaching me something new! I don't use AWS so I didn't know that but yeah, firewalld is a great thing to know

[gentle solstice]

though I also read that amazonlinux has all ports open by default because you can use aws itself as a firewall

[spice field]

I used to use ufw for everything because I started on debian/ubuntu. I work in a RHEL environment now and all I use is firewalld now

[gentle solstice]

look into setting up rules through the AWS security group

https://devcoops.com/install-firewalld-on-amazon-linux-2/

[spice field]

(thats for you backup 🙂 )

I dont work in a cloud environment but it's good to know if I ever do

[gentle solstice]

The recommendation is to not use a firewall, but to use aws security console to control what can access your server

a firewall adds overhead

[heavy knot]

the firewall prevents the server from running at all!

[heavy knot]

so this tutorial stops working for me because: sudo ufw allow 8000 does not work

did i write the correct allowed hosts?

[spice field]

from what I'm reading, thats acceptable, but I don't think it does what you think it does

It looks like it protects from certain web vuln attacks

[heavy knot]

does not accept all hosts?

how do i determine what hosts i need?

[spice field]

I don't think that setting is applicable to the issue you're having

At what point did the website stop letting you connect?

[heavy knot]

sometime between when i make the directory for the venv and when i made the migrations for the directory

[spice field]

wait im confused you CTRL + C'd it

[heavy knot]

so i named y directory gdirectory

and it is in the venv of the repo

[heavy knot]

i copied the firewall?

[spice field]

is directory meant to be called gdirectory?

[heavy knot]

you can name it anything

[spice field]

copied the fw from where?

[gentle solstice]

you need to configure the firewall in the aws console

[heavy knot]

projectdirectory

[spice field]

cause to me you stopped the django project from running

with CTRL + C

again i dont know django, just using t/s skills. unalive sounds more confident at this point than I do

[heavy knot]

have i not done that?

[gentle solstice]

No, that's just instructions on how to stop

[spice field]

oh, I misinterpreted how it was written

[heavy knot]

ohhhh yeah i control c to stop the server but when it is suppsoed to be running it does not run

[gentle solstice]

Can you connect via curl localhost:8000?

[heavy knot]

no

[gentle solstice]

in a new terminal console, right?

[heavy knot]

i reopened the terminal yes

sometimes it times out

[gentle solstice]

don't close the terminal

[heavy knot]

ok but if it hypothetically times out can i refresh it ?

[gentle solstice]

keep it open by making a systemd service file

[heavy knot]

hard life backup

[heavy knot]

in the gdirectory or the regular venv i make in thought_sync_django-channels

here

[gentle solstice]

are you familiar with systemd?

[heavy knot]

or here

[heavy knot]

i have used systemd before yes

but idk where to write commands because in the venv or not in the venv does different things

[gentle solstice]

systemd service files go in /etc/systemd/system/

[heavy knot]

what are we looking for?

or are we making one of these?

[gentle solstice]

Follow this guide. https://www.digitalocean.com/community/tutorials/how-to-set-up-django-with-postgres-nginx-and-gunicorn-on-ubuntu

skip what you don't need

So start at step 7

[heavy knot]

question about step 6

do i have to run this in gdirectory?

also what happens when i control c and close this, does it just not work?

[gentle solstice]

The directory containing manage.py

[gentle solstice]

This is just testing, so it won't matter.

[heavy knot]

oh ok

[heavy knot]

i think i finished step 7

[gentle solstice]

keep going

[heavy knot]

oh no

the preset in the guide is enabled

did i miss a step

[gentle solstice]

Did gunicorn.service start?

[heavy knot]

is that not shown

do i have to eanble?

[gentle solstice]

Run systemctl staus again

[heavy knot]

interesting?

[gentle solstice]

You only enabled the socket. You also need to start it

[gentle solstice]

your service file may be invalid.

[heavy knot]

how do i fix?

[gentle solstice]

share your service file.

cat /etc/systemd/system/gunicorn.service and share here

[gentle solstice]

Use your own username

or make your own

All the boxed strings in the textbox should be replaced with your own

WorkingDirectory should be where your project is located.

[gentle solstice]

ExecStart should be the fully qualified command to start your server

User should be your own user, like ec2-user

[heavy knot]

what does that mean python manage.py runserver 0.0.0.0:8000

[gentle solstice]

Fix the working directory and exec start

we won't be using manage.py. gunicorn is used to start the server

/path/to/your/project/.venv/bin/gunicorn

You can run which gunicorn to get the full path

[heavy knot]

what does that start with? ec2/tsupdate/thought_sync_django-channels

[gentle solstice]

^

are you familiar with how absolute paths work?

[heavy knot]

no

[gentle solstice]

there's basically the root / which is the parent of all files. Your home directory is located at /home/yourname

That's the default directory when you log in

the absolute path is the path starting at root

[heavy knot]

noted, now how do i find out the absolute path?

at least in this context

[gentle solstice]

^^

[gentle solstice]

Ok run this. readlink -f $(which gunicorn)

[heavy knot]

i have to go home now but i sure be on in 2 hours : ] thanks for everything

[heavy knot]

I have returned, I shall run this command

[gentle solstice]

Use that as the execstart

[fleet nebula]

hey, 2 questions, what programming language does a keyboard use? like a normal laptop keyboard, and can i program it to behave like "Razer Huntsman Mini V3"

basically my computer can't even run shooter games ( could get proof of my bad laptop ) it just seems fun making it, if anyone does not know, ppl hate it because it helps on shooter games

i know that Razer Huntsman Mini V3 has other features that compliment and make the shooting better, but i'd like to focous on the key canceling and not allowing a and d be pressed at the same time

not only a and d, almost every key, might be interesting to see how it effects avarage typing and such

does github of this exist? seems like a good idea, ppl would like to use this for hacks

[indigo zenith]

A lot here depends on the OS. Here's a note about fiddling with the Windows API with Python: https://stackoverflow.com/questions/71370107/how-to-change-input-keyboard-layout-programmatically-in-pyqt5

[spice field]

Anyone have any good resources on git pipelines and how to do local dev and push to a remote prod environment?

[gentle solstice]

dokku!

https://dokku.com

[spice field]

oh this is neat! Would this work for a discord bot?

[gentle solstice]

Don't see why not

it will be a worker app

[spice field]

I like the scaling feature, not sure how that would work for a bot because I know they use shards. But I think Dokku sounds like it's the right track.

Is what I'm essentially doing is committing my changes to my GH repo, and then go sync it with the dokku container, it detects the changes and pushes them?

[gentle solstice]

yes.

[spice field]

this sounds perfect

thank you so much

[burnt ginkgo]

Hi guys now if we are creating a virtual env for a system , it would be conda create -n env python 3.8

This would available in the whole system right?

Now what if I want a virtual env in directory how do I that?

I have used the --prefix method , but each time I want to activate it wants me to put the path

Is there any way to replace that path with the env name?

[heavy knot]

don't spam every channel...

(this is the right one though, not the others)

did you read did you read https://docs.python.org/3/library/venv.html ? @burnt ginkgo

[solid widget]

is there a way in python where when running a script it prints out whether the machine it's running on is a phone, PC, tablet or some other type of device?

[distant raven]

Is there a way to declare the build dependencies and dependencies to be installed locally for development purposes in the same place?
Right now I’m using setuptools in a venv and if I want my IDE to recognize the dependencies I have to install them with pip, which I’d do with a requirements.txt.
But if I do that I have to list the dependencies in two places (pyproject.toml and requirements.txt). Is there another way?

[distant raven]

if anybody wants to suggest conda I don’t want to use conda

[astral apex]

If you have dependencies required for build, they should go inside the requires in the [build-system].
Then they’ll automatically get installed when building your package.

A somewhat common way to add dev dependencies is to add them as optional dependencies
For example — https://github.com/letsbuilda/imsosorry/blob/main/pyproject.toml#L19-L25>
Then you install them with pip install -e .[dev]

One thing to keep in mind is that if you publish your package to PyPI that dev extra will be included, but it’s only installed if people explicitly ask for it, so it doesn’t affect anything.

[rancid schooner]

pyproject.toml lines 19 to 25

[project.optional-dependencies]
dev = [
    "pip-tools",
    "pre-commit",
    "ruff",
    "mypy",
]```

[distant raven]

Thanks.

How can I install runtime dependencies into the development environment?

[astral apex]

You start with pip install .
The . installs your pyproject.toml, it will use the requires to build your package, and then also install everything in dependencies [under [project]], all of your runtime dependencies should be in there.

Then if you also want to install your dev dependencies, you add [dev] to your ., and you get pip install .[dev]

You can install more than one at a time too, for example I put my docs building dependencies in a different group, so I would use pip install .[dev,docs]

[distant raven]

Can I install the runtime dependencies without building the package? I just need them so mypy etc. can find them while I code.

[astral apex]

That’s not really a thing currently

Some people are arguing for it

Most people say there’s no reason to not just let your project be installed too

You need your project to be available for mypy too

You can kinda hack it with Hatch if you really want it

[distant raven]

Why do I need my project to be installed in the venv for mypy to work?

It has worked fine before (but I didn't package anything then)

[rapid sparrow]

But if I do that I have to list the dependencies in two places (pyproject.toml and requirements.txt). Is there another way?
we use the choice of project building through setup.py. and just read requirements.txt file into its configuration 😉
as well as making dynamic reading of current README.md and changing version

[heavy knot]

can't poetry do that?

[thorny shell]

poetry doesn't require the dependencies to be in two places -- they're just in pyproject.toml (well OK they're also in poetry.lock, but that's autogenerated)

[heavy knot]

I meant symlinking the project without building it, so it 'live updates' i.e editable mode. But that too.

[alpine horizon]

where should I store Python CI scripts that are called from a GH workflow?

[astral apex]

Generally scripts go in a scripts/ folder

If they’re only used by CI, some projects create a ci folder
https://github.com/pandas-dev/pandas/tree/main/ci

[distant raven]

I am considering Hatch

does anybody have experience with it and would like to en- or discourage that choice?

[gentle solstice]

hatch is a fine choice.

It's great if you're looking for a modern alternative to setuptools

among other features

[heavy knot]

i have a new error this week! i have no idea what to do

here are my settings if that helps

[vernal mirage]

We have Jenkins CI running various pipeline via Docker images. These pipelines require numerous and sometimes conflicting Python packages from public and private repos. There's a recent push to solely install these requirements at runtime, but previously we were just baking them into the image at build time. What would be best practice here?

• IMO more time on build vs. run always makes sense and there's increased reliability since sometimes access to these repos can be flaky. There are different opinions in the org that images should be as small as possible which would remove the baking of these packages in.
• Additionally, we're looking to utilise a shared cache for these Python packages across pipelines. Our first iteration of this just uses a pip install --dry-run to warm the cache, but since we have some conflicting requirements we need a better way to do this. My first thought was to set up a local pypi server within the image for pip to install from. Is this a common usecase? Can you see this working?

[rapid sparrow]

We have Jenkins CI running various pipeline via Docker images. These pipelines require numerous and sometimes conflicting Python packages from public and private repos. There's a recent push to solely install these requirements at runtime, but previously we were just baking them into the image at build time. What would be best practice here?

The whole point of Docker images is creating frozen artifacts u can build once, test, check in stagng and be sure it will be the same running in prod now and at multiple servers and in few months.
So when u build new image and deploy it, and u screwed it up, u could rollback to previous docker image that was built before and everything will be okay (so u could relax and fix bugs for new version at your own temp, instead of rushing to fix prod bugs forward only right away at evenings)

You forfeitted THE Biggest point of Containerization.

TLDR: u are supposed to bake all dependencies on build

[vernal mirage]

Well said, that makes things a lot clearer. I think this is known but the reason it’s being suggested is because of a high frequency of these requirements changing combined with a long build time for our C software (which is also baked into the image).

[rapid sparrow]

mm... let me copy u instruction i wrote how to take care of it...

[rapid sparrow]

We speed up all our CI by reusing already built Docker image layer where they are already installed.
Our CI pipelines only write layers with code usuaally, and Heavy dependencies are all the time reused by pulling already built version from docker registry. We achieved it with this algo (Multi stage building pretty much, where first build stage has calculated hash based on used dependencies in Pipfile.lock, and pulled in every CI if exists first, and if it does not exist for this Pipfile.lock only then it is built and saved again to remote docker registry for reusage across all CI)

our CI in a nutshell:

1. Calculate md5 hash from dependencies file lock onto libraries. Whatever your language is. Lets call the value as builder_base_hash (we do it by python in Github Actions run 😉 )
2. Pull image image {{ builder_base_hash }} if it exists in docker registry, if not then build up to stage --builder. Save result to docker registry under tag {{ builder_base_hash }}
   (Thus we implemented speed up twice for CI), as we are were able to cache half of longest CI in a way that its CI jobs can run at different runners, as we use remote for persistence
3. run full building of an image, til the code capable to run unit tests and push to docker registry under tag build_${{ github.run_id }}
4. at unit test stage: pull the image build_${{ github.run_id }} and run unit tests and other tests
5. if it passed them, than save the image as service_name_{{ github.run_number }} as fit for deployment 🙂, also mark it as latest and etc whatever tags u need

So we used approach to cache building docker image stage through Docker Registry.

feels like i should write article about it 🤔

[vernal mirage]

That’s a very interesting approach. Using that approach, you’re, in a sense, splitting the requirement on that CI image being rebuilt into multiple parts. I can see that improving CI times quite a bit depending on where those “split” stages are defined. Presumably the effectiveness would be maximised at points of low dependency from one stage to the next.

Alternatively would a shared cache not offer the same benefit as your dependencies-are-installed image would?

[rapid sparrow]

Alternatively would a shared cache not offer the same benefit as your dependencies-are-installed image would?
sharing cache could but it is bothersome to setup working reliably and very CI provider depended. And Also default shared cache is not having ability to propagate for reusage across multiple CI runners.
Our solution is pretty much CI agnostic, will work the same at any CI tool. And works great for our autoscaled self hosted github runners

last time i tried using cache in Gitlab CI at least, it worked very unreliably, and the problems of cache persisting on same host only is an issue too, nullifying the benefits of caching

may be today it is different and shared cache became useful normally, at some of CI providers

[vernal mirage]

The caching I’m thinking of is more language-specific rather than some OOTB CI-provided tool e.g. ccache, pip cache, Python build cache. But I can see the value in rolling your own in the way you’ve described. Thanks for sharing.

I’ll definitely be more forward about the potential value being lost if we were to go forward with runtime installs.

[heavy knot]

wait can daphne run in venv?
I am in the project
I login into the server
i cd tsupdate
cd thought_sync_django-channnels
pipenv shell
and then i am in the project
would the directory be:
ec2-user@ip-172-31-19-47 thought_sync_django-channels
or ec2/tsupdate/thought_sync_django-channels?

[placid nacelle]

Checkout rye maybe

[distant raven]

I still don’t understand why mypy needs to have the project installed? In my experience it has worked fine without…

[alpine horizon]

I like the ci folder solution, thanks

[heavy knot]

how do i sudo yum install django, do i need epel? how do i install epel?

[hollow storm]

why do you want epel? you were asking the same thing the other day

[heavy knot]

should i want epel? i want django to work lol

[hollow storm]

if you have python+pip installed already, you could just install it with pip, e.g. pip install django outside a venv (or do it inside a venv and source it before starting ur server)

[heavy knot]

Pip install Django is not working

[radiant pollen]

try
py -m pip install django

[heavy knot]

i guess django already works?

then why can i not reach my server?

[hollow storm]

django works as a server, having it installed is not the same thing as having it running

im not very familiar with django but i think you should have a file called manage.py in your project

once you find it, try running python manage.py runserver

[heavy knot]

same error

[hollow storm]

okay that's an improvement

my bet is that your instance doesn't have the right firewall port open

if you open another terminal on the same instance (or connect via ssh), can you run curl localhost:8000?

[topaz idol]

Hello

I am a freelancer.

[hollow storm]

if it works, problem is most likely firewall and you'll have to open the port

[topaz idol]

Help in solving error in python

[hollow storm]

hi dangi, welcome to the server. as a friendly note, remember that asking for (or offering) paid work in this discord is against the rules

[patent raven]

https://github.com/python-screen-builder/python-screen-builder

[heavy knot]

no i can not:
$ curl -v 'http://harmonyapp.org' curl: (7) Failed to connect to harmonyapp.org port 80 after 259 ms: Couldn't connect to server

[hollow storm]

run it as i wrote it

[heavy knot]

that also failed:

[hollow storm]

did you close this process to run the command? you need 2 terminals open, one with the process and another one running curl

[heavy knot]

how to i open two AWS terminals?

[hollow storm]

well you either open two tabs or connect with ssh instead of the web interface

[heavy knot]

the error is too large to copy

um

how do i share a message over 2000 words?

[hollow storm]

i think discord can convert it to a file when you paste it

otherwise pastebin or something like that

[heavy knot]

https://paste.pythondiscord.com/LBUA

got it

here you go

um why is the error in html?

[hollow storm]

as a general comment, you should think about what you're trying to do and how the system works before doing anything. us feeding you commands and you copy pasting is not going to translate to any useful skill if you dont understand why each step is necessary

[hollow storm]

that is not an error, it's the website you want

this means that the django service is running properly and you can reach it locally. now it's a matter of being able to reach it from the outside

[heavy knot]

outside meaning aws config? what do i do next?

[hollow storm]

outside meaning from your computer, or anyone's computer

[heavy knot]

ok, so how do i do that?

[hollow storm]

give this a try

why can you access it locally but not remotely?

[heavy knot]

idk but i can i guess?

[hollow storm]

sure, guessing is fine

[heavy knot]

i think it is because it works locally but something is wrong with the AWS settings

[hollow storm]

science is basically coming up with a theory and then giving it a try

[heavy knot]

i do know however it works locally and not in AWS so that is a start

[hollow storm]

yes, good

[heavy knot]

i have limited time and therefore limited tries : [

[hollow storm]

your EC2 instance in amazon is in it's own network. AWS calls this VPC, and restricts their inputs/outputs

this is done with something called security groups, which are basically a firewall. my theory is that your instance's firewall doesn't have the right input port open (8000)

[heavy knot]

are my security groups set up correctly?

[hollow storm]

second to last group seems correct to me. is this security group applying to your instance?

[heavy knot]

yes, this is the security group inbound rules used on the server

outbound is "all"

[hollow storm]

can you show me the network tab of the instance? i dont remember the exact name of the tab, but it's on ec2 -> instances -> your instance

[heavy knot]

this thing?

[hollow storm]

i cant see whether the security group is applying to the instance here

[heavy knot]

oh that would be in security

[hollow storm]

firewall seems fine

new theory: django doesn't bind to external ip by default

should be easy to test. run it with python manage.py runserver 0:8000 instead

im not sure if this has security implications, for the final setup it might be better to setup a reverse proxy instead

[heavy knot]

still nothing

[smoky pivot]

👋 Hey all, would anyone be able to help setting up traefik on kube? I've been stuck on it for the past 4 hours and I desperatly need help. I can't for the life of me get the traefik dashboard working with HTTPS. I always get errors with ingressRoutes or something else which is extremely frustrating 😦

[rapid sparrow]

🥲

i am a person configuring similar stuff

[smoky pivot]

Have fun 🤣 😭

[rapid sparrow]

i would say, tough up and debug step by step where it is working

i enjoy using kind instrument for local debugging

[smoky pivot]

That's what I've been doing for the past day...

[rapid sparrow]

and k9s to navigate rapidly over kubernetes things and seek the clues

[smoky pivot]

Yeah I already am using k9s it's awesome

[rapid sparrow]

obviously though ingress will not work in kind unless u apply metallb btw

[smoky pivot]

error: resource mapping not found for name: "traefik-dashboard" namespace: "default" from "ingressroute.yaml": no matches for kind "IngressRoute" in version "traefik.containo.us/v1alpha1" -> Which leads me to try and define IngressRoute but then other errors ensue ...

[rapid sparrow]

dev env for kubernetes requires too many things, i just use kind in terraform-opentofu locally 😋
That helps me locally to go through all issues before i try to intergrate it working in our staging (and facing new arrays of problems)

[smoky pivot]

And when I try to do another type of ingress for the traefik dashboard it just doesn't work...

[rapid sparrow]

resolving those issues can indeed take a whole 4h or smth

it is just literally beyond the... limits of this channel 😅

we try to help in a reasonable some time

but debugging such issues can indeed take 4 hours and more 🥲

[smoky pivot]

Yeah I know, that's what I've been stuck on since this morning 🥹

[heavy knot]

hey Darkwind if you are available my server is not running and i do not know why?

[rapid sparrow]

that totally depends on type of your deployment strategy/target (where u deploy it)

[heavy knot]

I am deploying a Django project in AWS EC2

[rapid sparrow]

okay then it makes it compatible with my ways. i am using AWS too 😅

1. curl localhost of your website from within EC2 sever within and check if it works

2. check that u binded Django server to 0.0.0.0 for public exposure

[heavy knot]

1. it does not: https://paste.pythondiscord.com/LBUA

[rapid sparrow]

3. check that your EC2 security stuff configurations allow traffic from IPs to ports u wish

4. check that your EC2 is even having public IP

[heavy knot]

what is biniding, is that not security groups?

[rapid sparrow]

5. check any other custom networking that u touched for it

[heavy knot]

i think they do, so i have to ssh into port 80, i can only ssh into port 22

[heavy knot]

it has an elastic IP

[heavy knot]

i dont think i did that but idk

[rapid sparrow]

gunicorn your_app.smth -p 0.0.0.0 (rough example, or python3 manage.py runserver -p 0.0.0.0)
smth like that stuff exists, if you bind your web python server to 127.0.0.1, it will be acessable locally only
if using 0.0.0.0, it will be exposed beyond host network to all attached networks to the host

[heavy knot]

what do i have to change?

[rapid sparrow]

that is a rough example

u are supposed to fix it where u launch your app inside host

to check to which port u binded

ergh... u can actually validate it by googling to which port/ip my app is binded and making scan over your host if desired (through the found commands)

[heavy knot]

what are we looking for?

[rapid sparrow]

we can see from this stable your app running at 0.0.0.0:80 ip listening to everything

so your are good at this department

also it listens to all ipv6 too (potential screw ups to be prepared, there are always some kind of screw ups with ipv6 😅 )

[heavy knot]

what does this mean? run it at port 80?

new error lets go!

[rapid sparrow]

only sudo level users can access ports below 1000 or smth like that

login as sudo if u wish binding to 80

sudo -s

[heavy knot]

does that not take me out of the venv?

[rapid sparrow]

it does take you out of a venv. relogin to venv again source path_to_venv/bin/activate

or run your app as path_to_venv/bin/python3 manage.py runserver (if wishing running with venv without its activation)

[rapid sparrow]

you haven't used the commands I provided

u made mistakes both times in missing parts

[heavy knot]

what did i miss?

[rapid sparrow]

u haven't wrote source the first time
and u wrote python with space in the beginning the command to be path_to_venv/bin/ python instead of requested path_to_venv/bin/python3

are you able to read english directly hmm? 🤔

[heavy knot]

new error!

[rapid sparrow]

u managed to make mistake once again in executing my initial request to use one thing or another one (but not both at the same time!)
#tools-and-devops message
#tools-and-devops message

i have to go now, gl

[heavy knot]

is it activated?

[heavy knot]

now my terminal is broken

anyone out there?

[heavy knot]

what is my endpoint ID?

[smoky pivot]

@rapid sparrow since we talked about it earlier: https://github.com/traefik/traefik/issues/10986

[muted iron]

Hey if I want to get going with spark, should I deploy all the spark, kafka, zookeeper etc in docker images? Someone once told me to always do kafka in docker never local, is it the same for others as well?
even the database?

[heavy knot]

why does epel-release not install normally?

(thought_sync_django-channels) [ec2-user@ip-172-31-19-47 thought_sync_django-channels]$ sudo yum install epel-release
Last metadata expiration check: 1 day, 2:29:14 ago on Tue Aug  6 17:53:05 2024.
Error: 
 Problem: conflicting requests
  - nothing provides redhat-release >= 7 needed by epel-release-7-14.noarch from epel
(try to add '--skip-broken' to skip uninstallable packages)
(thought_sync_django-channels) [ec2-user@ip-172-31-19-47 thought_sync_django-channels]$ sudo dnf install epel-release
Last metadata expiration check: 1 day, 2:32:09 ago on Tue Aug  6 17:53:05 2024.
Error: 
 Problem: conflicting requests
  - nothing provides redhat-release >= 7 needed by epel-release-7-14.noarch from epel
(try to add '--skip-broken' to skip uninstallable packages)
(thought_sync_django-channels) [ec2-user@ip-172-31-19-47 thought_sync_django-channels]$ 

[hollow storm]

i think it's because you are in centos7

and its EOL

[heavy knot]

how do i install for centos7?

[gentle solstice]

You broke your repos

[heavy knot]

how do i fix them?

(thought_sync_django-channels) [ec2-user@ip-172-31-19-47 thought_sync_django-channels]$ sudo yum install epel-release-7-14.noarch
Last metadata expiration check: 1 day, 9:36:12 ago on Tue Aug  6 17:53:05 2024.
Error: 
 Problem: conflicting requests
  - nothing provides redhat-release >= 7 needed by epel-release-7-14.noarch from epel
(try to add '--skip-broken' to skip uninstallable packages)
(thought_sync_django-channels) [ec2-user@ip-172-31-19-47 thought_sync_django-channels]$ sudo dnf install epel-release-7-14.noarch
Last metadata expiration check: 1 day, 9:36:20 ago on Tue Aug  6 17:53:05 2024.
Error: 
 Problem: conflicting requests
  - nothing provides redhat-release >= 7 needed by epel-release-7-14.noarch from epel
(try to add '--skip-broken' to skip uninstallable packages)
(thought_sync_django-channels) [ec2-user@ip-172-31-19-47 thought_sync_django-channels]$ python install epel-release-7-14.noarch
python: can't open file '/home/ec2-user/tsupdate/thought_sync_django-channels/install': [Errno 2] No such file or directory
(thought_sync_django-channels) [ec2-user@ip-172-31-19-47 thought_sync_django-channels]$ python3 install epel-release-7-14.noarch
python3: can't open file '/home/ec2-user/tsupdate/thought_sync_django-channels/install': [Errno 2] No such file or directory
(thought_sync_django-channels) [ec2-user@ip-172-31-19-47 thought_sync_django-channels]$ 

[gentle solstice]

Repos are configured in /etc/yum.repos.d

[heavy knot]

what should i add to the repo?

[gentle solstice]

You are on amazonlinux. No epel for you

[heavy knot]

then what is wrong with the server?

[gentle solstice]

Django isn't installed

Use a venv

[heavy knot]

i have i think?

[gentle solstice]

You didn't activate your venv after becoming root

[heavy knot]

i ran pipenv shell? should i run something else?

[gentle solstice]

Do it after sudo

sudo -s ignores changes to your path

[heavy knot]

did i do it wrong?

[gentle solstice]

It means pipenv isn't installed for root.

Install it globally

[heavy knot]

so cd out of the root?

[gentle solstice]

Stop installing epel

[heavy knot]

what security group am i missing?

i thought i opened port 80?

[gentle solstice]

You should provision a new vm

[woven phoenix]

Hey, where should i ask questions regarding Big data like Apache Hadoop or Spark?

[heavy knot]

what does that mean? make the venv somewhere else? where?

[gentle solstice]

kill your ec2 vm and recreate it

it needs to be put out of its misery

[heavy knot]

i can kill a port how does one kill a vm?

[gentle solstice]

delete it

[heavy knot]

https://tenor.com/view/smg4-gun-we-cant-expect-gif-25209316

[heavy knot]

my live reaction to this

[gentle solstice]

two in the heart, one in the head

[heavy knot]

wait so i get into my venv using:
cd tsupdate
cd thought_sync_django-channels
and then running
pipenv shell
where in this process do I run
rm -rf somethingsomething how do i path to the venv?

[gentle solstice]

do you know the difference between a vm and venv?

[heavy knot]

oh virtual machine versus venv

how do i kill a vm then? i am on a vm? is that not the instance? you mean terminat the instance i make a new instance?

[gentle solstice]

do it in the ec2 console

[heavy knot]

is there a virtual machine killing option?

[gentle solstice]

That's the security groups

You want instances

[heavy knot]

terminate?

[gentle solstice]

state is for shutdown/startup

go ahead and terminate it

[heavy knot]

now is there a specific type of instance i should make?

like know that i know more is amazon linux the best option?

[gentle solstice]

just stick to amazonlinux

[heavy knot]

are my security groups good? i think it might be reverse proxy but idk how to do that, should i get nginx?

[gentle solstice]

Using aws dns needs to use port 80

https://stackoverflow.com/a/69456960/2351110

Try using the ip

[heavy knot]

export the security group?

[gentle solstice]

connect to the ip:port

[heavy knot]

in the terminal or in here?

[gentle solstice]

in your web browser

https://tenor.com/view/coffee-beer-simpsons-gif-13247227

[heavy knot]

wait we should set up elastic ip first!

ok till not working

[heavy knot]

it working

now we have to get daphne and redis working

daphne is up

redis is not working

[hollow storm]

lmao

[bitter prairie]

What's the difference between the cloud and the arrows in VS Code? I pushed both branches, and I can see both of them in my GitHub account. When I pushed the master branch I used the -u command (git push -u origin master), and when I pushed ap, I didn't use it.

As I understand, -u makes git remember the branch as the default branch to push future commits to. I didn't want to make ap the default branch, so I didn't use -u.

By the way, when I created my first repository through VS Code's GUI, it called the main branch main, and when I used git init in a different repository (the one in the screenshots), it called it master 🤔

[thorny shell]

yeah, some code that uses git has changed the name of the default branch from "main" to "master". There's nothing significant about that; it's just a convention

[gentle solstice]

If you don't push with -u, vscode won't know there's a remote tracking branch, so it will offer to "publish" the branch.

[gentle solstice]

Run git config --global init.defaultBranch main

[stable hearth]

is there a way to hide config file which includes my API key ( i didnt push the files in yet) in my repository on github? i heard about gitignore but idk how to implement it

[bitter prairie]

is that bad, or can I ignore it?

[inner pollen]

Use .env file, read about environment variables on internet

[stable hearth]

Aight

[thorny shell]

No.

[subtle quarry]

Like make it not tracked by git? Then create file .gitignore in root of your repo and create entry there with your config file path

If you want to have that in GitHub or similar there are options to store keys secure in secrets or something like that can't remember exactly the name now. By doing that. These keys are then available to GitHub actions.

[dim eagle]

I wanted help making a dependabot.yml file that would update workflows and also make a pull request when packages need to be updated

where would be ask to this?

packages should be specifically pull requested
workflows can be automatically updated

version: 2
updates:
  - package-ecosystem: "devcontainers"
    directory: "/"
    schedule:
      interval: "monthly"

  - package-ecosystem: "github-actions"
    directory: "/"
    schedule:
      interval: "monthly"
    groups:
      ci-dependencies:
        patterns:
          - "*"

  - package-ecosystem: "pip"
    directory: "/"
    schedule:
      interval: "monthly"
    groups:
      python-dependencies:
        patterns:
          - "*"

I have this rn but I want it to make a pull request for package updates.
But I want it to update when a workflow is changed

The reason why I want dependabot to automatically update workflows is because it's harder for me to update them but package updates is something I myself can do so I tend to do it myself

[bitter prairie]

some people recommend changing the default terminal in VS Code to git bash. why? how is it better than powershell?

[astral apex]

Git Bash installs Bash on Windows, but then it also installs a bunch of really useful command line tools that aren’t available on Windows.

However
Windows has WSL now, which is the same thing but a million times better, and you should use that instead.

[brazen forge]

I didn't really understand what you mean here
could you rephrase your question?

[dim eagle]

the wolf/dog guy above helped me out thankfully

https://github.com/JDJG-Holding-Team/Melody-Api/blob/main/.github/dependabot.yaml
gave me that file

[brazen forge]

okay
what's automatic in that?
also, isn't that the same Dependabot config you already shared?

[alpine horizon]

how do I trigger a GH actions, on a cronjob or if the commit message contains some word?
I found this to accomplish the latter, but how can I integrate it with my cronjob?

[rapid sparrow]

https://docs.github.com/en/actions/writing-workflows/choosing-when-your-workflow-runs/events-that-trigger-workflows
Cronjob

on:
  schedule:
    - cron: "15 4,5 * * *"   # <=== Change this value

[rapid sparrow]

if u use hackery like this for if conditions, then u need to create two Workflow files.

one that is working on commit, and uses this if condition

and second one working on cronjob

they both can be reusing same shared workflow code

jobs:
  check-translation-fuzzy:
    name: Prevent fuzzy translation
    uses: Orgname/actions/.github/workflows/check-fuzzy.yml@master

as example

the reusable workflow will need to have permission being called on workflow_call

name: Checks .po files for fuzzy translations

on:
  workflow_call: {}

jobs:
  lint:
    name: Prevent fuzzy translation
    runs-on: ubuntu-latest
    steps:
      - name: Check out source repository
        uses: actions/checkout@v2
      - uses: Orgname/actions/locales-check-fuzzy/@v4

[alpine horizon]

I see, so basically 3 files. one for the actual job definition, and two for each condition

didnt think about that solution

are there any alternatives tho?

[rapid sparrow]

yes

u can do it with a single workflow file

if u change your regarding if: "!contains(github.event.head_commit.message, 'build')"
Replace it with stuff like

on:
  release:
    types: [published]

or

on:
  release:
    types: [created]

GH has special trigger to run on Github Releases created

probably on tag creation exists too

https://docs.github.com/en/actions/writing-workflows/choosing-when-your-workflow-runs/events-that-trigger-workflows
Check options there

[rapid sparrow]

we in company use running build creation/deployment of a library on git release published)

Then u will be able to merge your conditions into a single workflow file

on:
  schedule: your cron schedule
  release:
    types: [published]

[alpine horizon]

I dont get it, how can you include everything in the same file? here you are not checking if the commit contains the message? (and if you add the check then it will fail on cronjob ...)

I need smth like

on:
  schedule:
    - cron: '0 7 * * *'
  push:
    branches:
      - master

jobs:
  build:
    if: "!contains(github.event.head_commit.message, 'abcd') or is_cronjob(...)"

[rapid sparrow]

i said do not check if the commit contains the message

use alternative optional triggers

through on Git Tags or on Github Release creation instead

on:
  push:
    tags:
      - v1.**

Git tags are git native solution that u can create with git alones

[alpine horizon]

oh so smth like on: push: tags: - "trigger actions"

[rapid sparrow]

yup. That is assuming u need to build only on releases of your app (besides cron)

Alternatively if your optional trigger stuff is not tied to product releases at all... then just use workflow_call, for human allowed CI trigger from GUI (probably usable from gh cli too)

through githooks it can be triggered then automatically on specific commit message 😁https://git-scm.com/book/en/v2/Customizing-Git-Git-Hooks

[alpine horizon]

I think this is what I'm looking for.
cuz sometimes I want to test the actions after a commit (but not always).
the issue is that on the website if you run a workflow manually it runs on the same commit ref, not the latest one, so it accomplishes nothing in my use case

[alpine horizon]

with git filter-repo, after I run the command to rewrite the history, how do I publish to the remote? (git status says there are no changes )

[rapid sparrow]

git push
if u rewrote then git push --force-with-lease

[alpine horizon]

I'm missing something:

$git push --force-with-lease
fatal: No configured push destination.
Either specify the URL from the command-line or configure a remote repository using

    git remote add <name> <url>

and then push using the remote name

    git push <name>

[rapid sparrow]

okay then set remote first then

git remote add origin [e-mail]:browserpass/browserpass-extension.git for example

u can clarify what u have set with git remote -v

once u configured remote origin, then u can normally use git push (the first one will be with extra flag potentially though)

[alpine horizon]

it says to do git push --set-upstream origin master, but then rejects the push because "the tip of your current branch is behind its remote counterpart"

[rapid sparrow]

git push --set-upstream --force origin master

this will overwrite history of remote

with also setting remote as its remote target

be careful in its usage)

[alpine horizon]

thanks so much it worked

[rapid sparrow]

after u do it once, next time it will be just git push

[alpine horizon]

now I just gotta hope I didnt mess anything up

[alpine horizon]

really thanks for the help

[versed apex]

Please tell me about your favorite plugins you use in pycharm, I want to customize my workflow

[brazen forge]

try asking in #editors-ides, instead

[daring ore]

I'm trying to learn AWS features in prep for an interview, but I don't know how to create a database without costing me. It says it's in free tier, but both options for configuring Dynamo and RDS are saying they'll cost me. Any help?

[thorny shell]

dynamo is so cheap that I wouldn't worry about it.

RDS is a whole different story -- I considered using it for a toy project, and the cheapest I could find was I think $70/month

[brazen forge]

!rule ad

[rancid schooner]

Rules

6. Do not post unapproved advertising.

[heavy knot]

I am setting up elastic cache to use redis service, any recommendations on how to change the settings?

[rapid sparrow]

1. changing settings in AWS with terraform is recommended
2. Serverless is many times more expensive for such types of infra than using persistent server

[haughty burrow]

Hi all, I put together a dead simple pulumi (python flavor) gist to configure daemon sets to multiplex GPUs by configuring the nvidia-device-plugin: https://gist.github.com/MMMarcy/8f8bc3f660e01eb46cf28c654d373a62 .
It's helping me to run several small LLMs and embedding models on the same GPU from my minikube instance. Not sure this is the correct space, but hopefully it's useful for someone messing around with LLMs/K8s like I am doing

[heavy knot]

How I am supposed to access terraform, is it free?

[pseudo cipher]

Yes, there is a community version. There is also OpenTofu.

[heavy knot]

is there any specific settings I should be concerned with changing?

[pseudo cipher]

I am not sure I understand the goal. You can create an ELK stack and a REDIS service in a lab for free. Is this for educational purposes, or is there some other goal?

[heavy knot]

i need the REDIS service to function for an AWS server that is deploying a Django app

[pseudo cipher]

@heavy knot So, you have Django inside an EC2 instance?

[heavy knot]

yes

[pseudo cipher]

If the Redis service is strictly for Django, I would run as a service in your EC2 instance. Then the site can talk to it using localhost. I am not sure how the cost would compare. I would think just a service on the EC2 instance would be less expensive than the Redis service offering. Have you considered that as an option? https://serverfault.com/questions/1127483/how-to-install-and-configure-redis-server-on-amazon-linux-2023-al2023

[heavy knot]

the whole point of getting redis service to work is so that the websocket works for user interaction, how do i test that this installation of the service works by using postman to test the websocket?

[pseudo cipher]

You would need to edit the redis config after installation. You could set permissions and access rules in it. Once you have that defined, you can use whatever tool you want to verify that the endpoint is reachable. https://stackoverflow.com/questions/19091087/open-redis-port-for-remote-connections

[pseudo cipher]

You can limit it by IP on your firewall if you know what IPs need access.

[heavy knot]

8001 is for daphne, 8000 is for the server, is there some port redis service uses?

[pseudo cipher]

By default it is 6379, but you can set it to whatever in the config file in the /etc/ directory. By default it will bind to loopback but you can change it to your IP address or 0.0.0.0 to bind to all interfaces.

[heavy knot]

this explains that i need a config file but does not explain what should be in the config file: sudo nano /etc/redis6/redis6.conf

[pseudo cipher]

Most of the defaults are probably fine, this is the section that configures whether it will listen to an interface other than localhost. Here is some reference docs, too. https://redis.io/docs/latest/operate/oss_and_stack/install/install-redis/install-redis-on-linux/

[heavy knot]

should i use administration, security, or replcation for Redis?

[pseudo cipher]

I don't think you need any of that on a single node install. You can just set the bind address list and start the service with systemctl.

[pseudo cipher]

Maybe this is helpful? My config is just at /etc/redis.conf

[heavy knot]

i found this online, would this work?

[pseudo cipher]

Yeah, try firing it up at the command line with something like this. It will tell you if there is an issue. I gotta step away for a bit, be back in an hour or so. You can find the binary like this...

[pseudo cipher]

So, /usr/bin/redis-server /path/to/config.file, something like that.

[heavy knot]

darn, what did i do wrong with my settings?

#Redis will listen on this port
port 6379

#Bind to all network interfaces. For better security, consider binding to specific interfaces.
bind 0.0.0.0

#Enable or disable TCP keepalive
tcp-keepalive 300

# Number of databases. By default, Redis has 16 databases.
databases 16

# Maximum memory usage. Redis will start evicting keys if this limit is reached.
# Uncomment and set the value to a specific amount of memory.
# maxmemory 2gb

# Eviction policy when memory limit is reached. Options: noeviction, allkeys-lru, volatile-lru, allkeys-random, volatile-random, volatile-ttl
# Uncomment and set the value as needed.
# maxmemory-policy noeviction

# Save the database on disk every 900 seconds if at least 1 key changed
save 900 1

# Save the database on disk every 300 seconds if at least 10 keys changed
save 300 10

# Save the database on disk every 60 seconds if at least 10000 keys changed
save 60 10000

# Append only file (AOF) configuration
# Enable AOF persistence (recommended for durability)
appendonly yes
# AOF rewrite policy, options: always, everysec
appendfsync everysec

# Set the name of the AOF file
appendfilename "appendonly.aof"

# Enable Redis to write a disk snapshot every time it is saved
rdbcompression yes

# Enable Redis to use rdb files for persistence
rdbchecksum yes

# Log level: debug, verbose, notice, warning
loglevel notice

# Path to the log file
logfile "/var/log/redis/redis-server.log"

# Daemonize the Redis server (run in the background)
daemonize yes

# Set the Redis server PID file
pidfile /var/run/redis/redis-server.pid

# Security configurations
# Require clients to authenticate with a password (uncomment to enable)
# requirepass yourpassword

# Use a Unix socket for local connections (optional, enable if needed)
# unixsocket /var/run/redis/redis.sock
# unixsocketperm 700

# Enable protected mode (recommended for security)
protected-mode yes

# Disable commands that could be harmful if exposed to the public
# Uncomment the following to disable dangerous commands
# rename-command CONFIG ""
# rename-command SHUTDOWN ""
# rename-command FLUSHDB ""
# rename-command FLUSHALL ""

# Specify the server's time zone (optional)
# time-zone "UTC"

# Include additional configuration files (optional)
# include /etc/redis/redis-user.conf

# Path to the Redis server binaries
/usr/bin/redis-server /path/to/config.file
./redis-server /path/to/redis.conf

[pseudo cipher]

Change daemonize to no for more output, set loglevel to debug, for troubleshooting, also, try running as root or with sudo to see if it is permissions. One of those should give more insights on the issue.

[pseudo cipher]

Oh, and if the log file that is listed is empty, it could lack write permissions to that or something similar.

[heavy knot]

what should i write to the log file? also what does this warning mean?

[heavy knot]

is it just this but in /var/log/redis/redis-server.log?

logfile "/var/log/redis/redis-server.log"

loglevel debug

[pseudo cipher]

Try this one... this gives the redis user access to the directory it uses to store stuff by default.

[heavy knot]

survey says: "no". hmmm what user besides sudo would there be?

[astral apex]

Sudo is not a user

Sudo changes the currently logged in account from your user to the root account, and then executes the command you gave it

If the redis user doesn't exist, you will need to create it

I'm surprised the installation didn't create it, which makes me wonder if it was installed correctly

[pseudo cipher]

Normally, it gets created during package install, but maybe the Amazon Linux instance has a minimal package of some sort? You can check the service definition to verify that it is trying to run as redis, and if it is, create a redis user using the useradd tool.

[pseudo cipher]

This shows the service definition file

[pseudo cipher]

Command to view contents of file is sudo cat /usr/lib/systemd/system/redis.service

Contents will look something like this, showing the user. You can view users on the system with sudo cat /etc/shadow This gives info on adding a user. https://www.geeksforgeeks.org/useradd-command-in-linux-with-examples/

[heavy knot]

do i already have a user or do i have to set one up?

[pseudo cipher]

Looks like it is redis6, try running that command again with redis6:redis6 instead of redis:redis. I bet it is running as that user.

[pseudo cipher]

Those errors about /etc/nginx being a directory are because cat is for files. You can view what's in a directory with ls.

[heavy knot]

seems to work but i get a similar error?

[pseudo cipher]

sudo cat /etc/systemd/system/redis.service

[heavy knot]

my user is redis?

should i change that?

[pseudo cipher]

Try editing that file and change it to redis6, then run sudo systemctl daemon-reload , then restart the service again.

[pseudo cipher]

Edit with sudo nano /etc/systemd/system/redis.service or whatever text editor you like to use.

[heavy knot]

hmmm

is there something else wrong with my settings

[pseudo cipher]

sudo cat /etc/redis/redis.conf

[heavy knot]

? why can it not mkdir?

[pseudo cipher]

Where is your config file? sudo ls /etc/redis sudo ls /etc/redis6 ?

[heavy knot]

why can i not exit?

i sudo nano in?

[pseudo cipher]

Your service file /etc/systemd/system/redis.service is pointed at /etc/redis/redis.conf, but your config is /etc/redis6/redis6.conf. You could copy the file or update the service, I think I would update the service definition. It seems the installer package had some settings from redis6 and some from a previous version??

[pseudo cipher]

sudo systemctl status redis6 Is there another service called redis6?

[heavy knot]

less errors but differnet?

[pseudo cipher]

sudo systemctl stop redis && sudo sytemctl disable redis

sudo systemctl enable --now redis6 && sudo systemctl status redis6

[heavy knot]

we have green?

[pseudo cipher]

Maybe there was a different version installed by default. not sure, but it looks like you have two versions installed, but that's fine, the other one is disabled now.

[heavy knot]

thank you for everything you have been very helpful!!!

[heavy knot]

your patience knows no bounds

[heavy knot]

if you want to stick around I think I am going to get an app to run right about now 😪

if this does not work back to the chair (coding) for me

[heavy knot]

what should be my daphne settings, my websocket is not connecting?

we need to connect websocket

[pseudo cipher]

Not sure off hand, what do you get when you run this command curl -k http://localhost:8001

[heavy knot]

wait

let me run daphne

oh no

i get an html error

lmk get a paste bin

[pseudo cipher]

What's the error? Or are you sending me a link?

[heavy knot]

https://pastebin.com/qX2wx1V3

[heavy knot]

if that pastebin does not work let me know

the error is large

i could not fit it all on in the paste bin

but you get the idea

html stuff for daphne?

nani?

[pseudo cipher]

sudo dnf install lynx -y && lynx http://localhost:8001

[heavy knot]

we opened a portal to hell!

the portal to hell closed itself?

[pseudo cipher]

So, this fails daphne -b 0.0.0.0 -p 8001 django_project.asgi:application with whatever your app is, I assume at the end?

[heavy knot]

it dose not "fail" it can not connect

if it would work this would say "Connected"

[pseudo cipher]

Lynx is a text web browser, sometimes useful to check html pages in a terminal session. https://en.wikipedia.org/wiki/Lynx_(web_browser)

[heavy knot]

what does the html page know?

[pseudo cipher]

I got you, so we are getting ahead of ourselves. Is the app configured to only respond to certain requests? I mean if you were on the computer that is running the app, should you be able to load the web page? I am trying to make sure it loads locally before I connect over a public IP.

[heavy knot]

i am not sure what the app if properly configured to respond to? I think the server? here is the repo if would like to see: https://github.com/K2GHub/tsupdate.git

[pseudo cipher]

Does it matter what version of Django or Python?

[heavy knot]

i would assume not but maybe

[pseudo cipher]

Alright give me a bit, I will see if I can get it to load.

[heavy knot]

I will have to afk for about an hour brb

[pseudo cipher]

Ran this, then verified it is loading with Lynx internally by runnning lynx http://127.0.0.1:8000/auth I ran on port 8000, and the page loads locally, so that's good.

[pseudo cipher]

So, now, probably all about firewalls, VPC, security groups, etc.

[pseudo cipher]

What's this look like? sudo firewall-cmd --list-all. Oh, and press "q" to get out of lynx when you are done browsing pages.

[heavy knot]

I have returned

[pseudo cipher]

sudo firewall-cmd --add-port=8001/tcp --permanent

[pseudo cipher]

sudo firewall-cmd --reload

[heavy knot]

yay

[pseudo cipher]

That should add port 8001 to firewall exceptions. I believe you are running on port 8001 so you should be trying to connect on http://127.0.0.1:8001 with lynx, not 8000.

[heavy knot]

? am I snytaxing wrong

[pseudo cipher]

I am not familiar with where you are connecting from. Does your ec2 instance have a public IP assigned?

[heavy knot]

yes: PublicIPs: 3.217.114.187

[pseudo cipher]

Have you registered that domain name?

Did you add an allow rule to your security group in AWS?

https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ec2-security-groups.html

If you know where you are connecting from, you can add an exclusion just for your IP, so you don't open to the whole world.

You can get your public IP for your home Internet at https://ipchicken.com

If you add the security group exception, the next test would be attempting to connect via https://3.217.114.187:8001/auth or whatever page for the last part.

If that works, then you can check if your domain name is correct by doing a lookup of the IP for harmonyapp.org

The domain does look correct.

[heavy knot]

nvm, big lag spike. is the format correct for /auth

[craggy mesa]

Hey peeps, sharing with you my Pixi, Flask, Jupyter, and Docker template for your new project https://github.com/mshaaban0/pixi-docker-template.git
(should get you started with one command if you have docker installed) hope it helps someone 🪓

[pseudo cipher]

Looks like it, maybe missing the SG rule? Something is blocking the inbound request.

[pseudo cipher]

There should be a spot in the console where you can make exceptions for inbound traffic. It should look something like this.

[heavy knot]

what should the source be?

[pseudo cipher]

Well, you could set it to 0.0.0.0. That will let the world access that port, or for testing, it's often better to narrow that so you don't get many unwelcome visitors. You can usually determine the IP you use on the public internet by visiting https://ipchicken.com and refreshing it a few times to see if your IP is fairly predictable.

[pseudo cipher]

Also, I'm pretty sure you are hosting on port 8001. If you are unsure, you can also enter a port range. You use the hyphen to denote a range like this. Eventually, you will want to configure SSL and possibly limit to relevant IP ranges, as well as monitor for any security vulnerabilities in Daphne/Django, etc.

[heavy knot]

not only did it not help websocket it stopped the server? I tried 0.0.0.0 and My IP, what is wrong?

[pseudo cipher]

What does your inbound exception look like, and did you try port 8001? I am pretty sure your API is on 8001.

[heavy knot]

oh wait redis is not working:

[pseudo cipher]

Aren't you using the redis6 service?

[heavy knot]

you are right

wait does "active running" mean i should run a command like daphne?

or maybe it is not the problem

[heavy knot]

is inbound exception different than security groups?

[pseudo cipher]

Inbound Exception tab in the Security Group.

[heavy knot]

also daphne is inactivate?

[pseudo cipher]

What do these referenced entries look like?

[heavy knot]

like this?

[pseudo cipher]

Yep, that should definitely open it up. Run the daphne command that starts the service. daphne -b 0.0.0.0 -p 8001 whatever

[pseudo cipher]

It's up..

[pseudo cipher]

You are going to want to restrict that soon. People are going to start trying to hack at it before too long if you leave that open like that.

[heavy knot]

but somehow the websocket does not work?

[pseudo cipher]

Is the websocket an internal connection? Is the app supposed to connect to a service running on the same server? Should it be pointed at localhost? Maybe the API doesn't need to be exposed? It depends on how the application is architected. Also, what is the UI you are using to test? Is that Postman? I am not that familiar with Postman.

[heavy knot]

the websocket and the server are not connected. i do not know the answer to the rest of the questions, here is my repo if you would like to view it: https://github.com/K2GHub/tsupdate.git

[pseudo cipher]

What happens if you try to connect to Google in that tool you are using?

[heavy knot]

oh, am i using it wrong?

[pseudo cipher]

I'm not sure; maybe I will read about it. I will let you know if I find something.

[pseudo cipher]

Some Googling makes it sound like you need to create the Websocket in your code, but I don't see any references to websocket creation in the code. https://developer.mozilla.org/en-US/docs/Web/API/WebSockets_API/Writing_WebSocket_client_applications

[pseudo cipher]

Could be that http:// and https:// are not valid for the Websocket King tool.

[pseudo cipher]

It looks like you need to configure a secure connection if you want to use Websocket King to troubleshoot and send POSTs to the endpoint.

[heavy knot]

new information: www.google or https.google ar enot websockets. websockets are essentialy connections that need to be made betwene users witihn the app using their user ids: their format is:
ws://IPorDOMAIN:PORT/ws/home/94be4db4-c018-43e7-9c44-c0bb8854129d/

therefore they should look like:
ws://3.217.114.187:8001/ws/home/94be4db4-c018-43e7-9c44-c0bb8854129d/

or

ws://harmonyapp.org:8001/ws/home/94be4db4-c018-43e7-9c44-c0bb8854129d/

[pseudo cipher]

If you try to use Websocket King, you still get the error. It sounds like you need to configure SSL before you can use Websocket King for testing.

[heavy knot]

how do i configure SSL? let's encrypt?

also what should my daphne group be? how should I config daphne?: https://pastebin.com/PAJ2PpDS

[heavy knot]

how do i make nginx publicly accessible?

[olive cipher]

Any apps, programs or scripts that can export a USDX format txt file to a midi or kar file?
https://youtu.be/RsPDDozg_PA?si=B4glvmjzgj7Ut49J

"I think it would be fairly easy for someone to make it. "
Says someone who only knows html, php

[gentle solstice]

forward port 80 in your aws security console

and port 443 if you use https

[heavy knot]

Like this? This does not work so idk what to do? I also need to get daphne service functioning, and set up SSL certificate, and correctly select a group for daphne