What's the best approach in python to do something like what happens when you run jupyter notebook, in other words, run a server (for example, with http.server) on some available port if the default one isn't available? Is it common to just catch the error if starting the server fails and try a different port until it works?
You can use port 0 - this will bind your server to some port that is currently known to be available by kernel. However, that makes a problem of service discovery - how your clients will know which port number server is listening on? If that's only you, shouldn't be big deal.
Related
what I have:
I made a simple web application using Flask, which woks great on the localhost.
what I want to achieve:
I want to deploy it so it is visible in the internet.
Because Flask says that the production server should not be used for deployment I installed gunicorn.
I managed to get it working on localhost (running gunicorn server:app).
Then I tried to make it public using the -b 0.0.0.0:5000 option. it said it is running on 192.168....:5000 Now I could access the website using that ip address from my computer and also from my phone connected to the same wifi. however I was unable to connect to the website when I didn't use this wifi.
when I searched about this, I found out that the 192.168.... ip address range is reserved to the local network and cannot be accessed from anywhere else.
how do I need to run the script so it is accessible from everywhere? Do I need to modify the firewall settings? Maybe there is a better way to deploy such an app on the local machine using a different framework?
Just some additional information:
python version: 3.9
I am using a venv (and installed flask and gunicorn into it)
os: macos 11.6.1 (I could run it on an Ubuntu machine if that is easier)
This question isn't anything related with Python or Gunicorn but to networking (so maybe StackOverflow isn't the correct place to ask but other community like ServerFault)
Currently your application is already exposed on your machine in the defined port, so next step would be to forward all the traffic that comes to your router to there.
In order to do that, you will have to configure the firewall of your router to accept incoming traffic through a desired port and finally forward the traffic which comes from that port to your machine in the port 5000.
Also, in the case that you have everything already configured, it will only work if your ISP is providing you a single IPv4, what currently doesn't happen anymore but you actually are sharing that with few more people. Other option would be that you configure IPv6 incoming traffic.
As you can see, this isn't a simple task neither one that should be done without proper care, since you would be literally opening your network to possible attackers.
So, in order to simplify it at the most for you, since you already have Gunicorn, I would recommend you to use any of the resources exposed by other users as Heroku or Netlify which are free for a single application and will fulfill your expectations without requiring high amount of networking knowledge.
You would need to do the following steps:
read up and learn a lot about security for Web servers, then read some more, it is fraught with risks
find your Mac's IP address on your local network and make sure it is fixed, i.e. set as static on your Mac (under "System Preferences->Network") or reserved in your router's DHCP tables (by putting your Mac's MAC address in its allocation tables) so that it always gets the same local address on your internal network when it boots
log into your router and set up "Port Forwarding" to forward external requests (coming from the Internet) for port 5000 (or some other port) to your Mac's fixed IP address and the port 5000 where Gunicorn is serving
log into your router and find your WAN IP address, or go to http://whatsmyip.com to get the address you need to put in your browser, or your friends need to put in their browsers to see your shiny new website
as it stands, this will work until the next time your router reboots when your ISP will likely allocate it a new IP address. If you want it permanent, you need to either 1) ask your ISP for a static IP address, or 2) subscribe a DDNS service (e.g. noip.com or dyndns.com) or 3) tell your friends your new IP address every time you reboot your router
I do not know anything about gunicorn. But what I used to use when needing to deploy a flask app was pythonanywhere. They have a great and totally free hosting service. It's really fast in deploying, needs no resources from your computer and is just great. Also you would have to forward the port 5000 for your computer to enable other devices not in the same network to access your computer/flask app. That comes with security issues. However, you do not need any of that when using pythonanywhere.
There are also other great hosting services like that one. It's just the only one I know and used yet but you'll find others for sure if you don't like that one.
I'm using Python 3.8.5 currently on windows, but the script should run also on linux and on macOS.
I'm creating UDP server and a client and have establish communication between them via multicast.
The client does not know from advanced the IP of the server so it is a "discovery tool" that will let me know information about the server ones it has received the data. -This is working-
On the server I'm using netifaces to get information about the interfaces.
The server sends information to the client such as IP address, mask, gateway. -information that I get from netifaces and is working-.
Example of data I get:
Now to the issue:
The python server can have a static IP or can be connected to a DHCP server -in case that is connected to a DHCP server I will not have access to the DHCP server-.
I wish to know when the DHCP is enabled or disable on the python server so I can send this as a flag to the client together with the rest of the information.
The idea is for the client to know that the IP that he's connecting to can change and he may lose connection and will need to start the discovery tool once more.
Because the user will not know in advanced if the python server is connected to a DHCP server or wil know the IP address of the python server or have access to the python server command line or configuration, using ip addr show, ipconfig, ifconfig and other commands before connecting is not an option.
I have seen that people use scrapy for DHCP communications but it seems that it does not solves what I wish to accomplish.
I do no need to configure the DHCP, just detect if my python server IP is static or dynamic.
I have search for over 4 days and have not found a possible solution.
I'm not asking to have the solution in silver spoon just to be pointed on the right direction.
**EDIT: I forgot to mention that the python server will be running on Ubuntu 16.04 and 20.04.
This will depend on your operating system setup. For example, here's a related question on how to detect whether DHCP is enabled in Ubuntu.
If using Luigi in a server I am connected to with ssh, is it possible to see the progress of tasks (as I can if I use luigi locally by looking up "localhost" in browser)?
Any help appreciated
Short answer: yes
When you run luigid "locally", there is a server that starts on your system, which you, as you mentioned, you can access at http://localhost:8082 (or whatever port you specify). To make that work on a remote server, all you need to do is run luigid in said server, then point your browser at, instead of http://localhost:8082, http://:8082 (or whatever port you configure luigid to listen on.
For example, if I have my minecraft server running on port 25565, I want to have a python script close the port so that all connections will be dropped and no further connections will be made without having to shutdown the service.
I have tried binding a new socket to the same port number and then closing the socket, but it does not have any effect on the server.
I am working in Python 3.3.
Use a firewall for example?
On linux there is the iptables. It's easy to use and powerful.
I'm trying to set up an ipython notebook server on an Ubuntu host machine - but can't seem to access it remotely. I set up the notebook server as per the tutorial, and launch it - everything seems fine. But going to https://my-host-ip:9999/ I get a timeout (error 118) message in the browser.
My intuition is that need to open the appropriate port (9999 in the setup tutorial) on my host. How do I do this (safely) with Ubuntu? More generally, is there a debugging checklist I should go through at this point?
Did you try to run it as public ? (listening on '*' ?)
http://ipython.org/ipython-doc/dev/interactive/htmlnotebook.html#running-a-public-notebook-server
Don't forget to make it over https and with password.
Is the port 9999 open?
I don't know why, but I had to set the IP to the IP of host to make it work.