Skip to main content

Python, Flask, Concurrency and Windows


This blog is for those who have written a program in Python and exposed it as API using Flask. You might have written the python algorithm and will be happy to see the output. With small usage and limited users, it still works by making some configuration changes. Problem arises when you host the same API in Production where multiple users are browsing it.

Why it does not work in Production as local environment?
In production, mutiple requests come to server from mutiple locations. If setup is not tuned to cater to concurrent users, either these requests will make the API down or will be throwing exception like 'Internal server error'.

How can we configure the setup to serve concurrent users?
There are few configuration which you have to keep in mind while designing the eco system to accomodate mutiple users.

  1. Make sure Python is installed on server. Download
  2. Download pip from get-pip.py and run command python get-pip.py
  3. Install Flask using pip commands pip install Flask
  4. Open command prompt, run python [filename].py and check your python file is running using Flask.
  5. Install Apache from Apache Lounge and install it as service. Check http://localhost and apache default page should appear. 
    • Open command prompt and goto Apache directory
    • run httpd.exe -k install
    • Open service.msc and look fot Apache service
    • Right click and run it as Automatic
  6. Install mod_wsgi(component for concurrency) using command pip install mod_wsgi
  7. Run command mod_wsgi-express module-config and this will produce output somthing like LoadModule wsgi_module...These are the directives needed to configure Apache to load mod_usgi.
  8. Copy the above output and paste it in [apache home directoty]/conf/httpd.conf at the bottom of the file. 
  9. Modify the Apache confirguration to listen to desired port. Open [apache home directoty]/conf/extra/httpd-vhosts.conf
    • <VirtualHost *:5000>
    •         ServerAdmin admin-name-here
    •         ServerName localhost:5000
    •        WSGIScriptAlias / "C:/Sumit Bajaj/web.wsgi"
    •         DocumentRoot "C:/Sumit Bajaj"
    •         <Directory "C:/Sumit Bajaj">
    •               Require all granted
    •         </Directory>       
    •         ErrorLog "C:/Sumit Bajaj/logs/error.log"
    •         CustomLog "C:/Sumit Bajaj/logs/access.log" common
    • </VirtualHost>
  10. In the above example "C:/Sumit Bajaj" is directory where python file resides. If web.wsgi file does not exists in the folder, create one. 
  11. Copy these lines into web.wsgi file - 
    • import sys
    • sys.path.insert(0, 'C:\Sumit Bajaj')
    • from app import app as application
  12. After all these steps, your app.py file should be running using mod_usgi on Apache.
  13. Even after your setup is not working. See FAQs below.
Python ecoSystem to support concurrency

Frequently Asked Questions(FAQs)

Apache is giving port not found exception 
Open [apache home directoty]/conf/httpd.conf  and search for 'Listen' and make sure your desired port is added to Listen.

Where mod_wsgi directly exists on windows?
It should be in [Python_directory]/lib/. If you are not able to find, search for web.wsgi file and you will get the directory.

Module is not available
While runnig this experiment, if you get execption 'Module is not available', don't get discouraged. Browse the module on internet and install it.

Can we tune flask to run mutiple threads?
Use app.run(host="0.0.0.0", port="5000", threaded=True) rather than default app.run()

After using Flask, mod_usgi and Apache, application is not able to feed concurrent requests

  1. Check your hardware
  2. If server has enough resources to server mutiple request. Check your Python code for improvement. In Python there is concept of Global Interpreter Lock(GIL) is a mutex which prevents access to Python objects for multithreaded environment. You may use threading library and lock the object till its work is completed. Read more on Python threading

Should I run Python on Windows or Linux?
Python and dependent libraries are better supported on Linux OS rather than windows, so I will recommended to use Linux. However, if your exisitng ecoSystem does not support it, nothing to worry. This blog is for you.


Please share your queries in comments and I will put my best to answer. 

Comments

Popular posts from this blog

Cannot alter the login 'sa', because it does not exist or you do not have permission.

Working on projects, it can happen that 'sa' account gets locked. If it is on local machine OR development boxes, onus would be on you to fix it. If scripts and SQL steps are not working, this might help you fixing the issue. Steps to unlock 'sa' account and resetting the password. 1. Open SQL Server Configuration Manager 2. Select SQL Server Services -> 'SQL Server' service. 3. Right click on 'SQL Server' service and click on "Startup Parameters". For 2008, server "Startup Parameters" are inside Advanced tab.   4. Add '-m' in startup parameters as shown above and click on 'Add'. This will put SQL server into 'Single User Mode' and local admin will have 'Super User' rights. For 2008, server you have to add ':-m' in the last of the existing query. 5. Save the settings and Restart the service. 6. Now open the SQL Server Management Studio and connect to database using 'Windows A

Could not load file or assembly 'Microsoft.Web.Infrastructure'

Could not load file or assembly 'Microsoft.Web.Infrastructure, Version=1.0.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35' or one of its dependencies. The system cannot find the file specified. What 'Micorosoft.Web.Infrastructure' does? This dll lets HTTP modules register at run time. Solution to above problem: Copy 'Micorosoft.Web.Infrastructure' dll in bin folder of your project and this problem should be resolved. If you have .Net framework installed on machine, this dll should be present on it. You can search for this dll and copy it in your active project folder.   Alternatively,  you can install this dll using nuget package manager PM> Install-Package Microsoft.Web.Infrastructure -Version 1.0.0 Happy coding!!

Git merge from one repo to another repo

This blog is for those who are looking for merging code from one repo to another repo. Why I will merge code from one repo to another? I forked from one git repo( may be some public git repo ) and did some cutomization on existing code. Occasionaly, features are being introduced in main git branch and I would like to get all those features in my own git repo. In this scenario, I would like to merge latest changes in my git repo. Here are few simple steps which will merge code from one git repo to another- 1. Clone the repo1(source git repo) > git clone https://github.com/org/repo1.git   > git pull 2. Clone the repo2(destination git repo) > git clone https://github.com/org/repo2.git > git pull 3. Goto repo2 (destination git folder) and checkout your prefered branch > cd repo2 > git checkout master>  4. Use below command to see the remote branch associated with your destination git repo > git remote -v  origin https://github.com/org/repo2.g