Thinking outside of the box with NGINX series - geo command
For the second part of this series let’s talk about another command, this time, let’s talk about the geo command.
The geo command, as the name says, was created to allow geolocation information, but actually, you can think of geo as a “switch case” where you can match network blocks and get a result from it.
Thinking outside of the box with NGINX series - split_clients command
I think it’s time for me to start a series of posts about useful NGINX commands, and different ways of solving common problems using NGINX and to start this series, let’s talk about split_clients.
SSL acceleration using NGINX and backend SSL connection identification
One common trick to optimize SSL termination performance is to have a proxy terminating all the SSL connections and proxying the connection to a plain HTTP backend.
This, in theory, works very well and without any code changes, but in practice, you discover that after doing this your application can’t tell the difference between HTTP and HTTPS clients.
Zero-downtime deployments using NGINX
Hello, everyone! Missed posting here, I should say that I need to start posting again!
As many of you know, some application servers do not support zero downtime deployments or even rolling restarts, and no, I’m not talking about unicorn, unicorn supports this out of the box.
Dublin DevOps Meetup - Nginx
Today, 17th April 2016, I’ve presented in our Dublin DevOps Meetup, that happened on Dublin Zendesk’s Office.
My talk was about my experience using Nginx, what’ve learned over the years, my mistakes and how I solved some of my problems.
I would like to say thanks to everyone that came to Zendesk’s office to see our talks.
The presentation is available at my github
Thanks!
Persistent backend selection using cookies with NGINX
Today a good old friend, Alex, sent me a “Bat signal” asking for help with an unusual NGINX configuration, I wonder why my friends only remembers me during “weird” incidents :P
Basically the requirement was to run a multiple copies of a legacy application on the same URL as it does not support folder prefix or subdomains.
Using map for upstream configuration on NGINX
Hi everyone, sorry for not posting any content here for such a long time.
Some of you should have noticed already that I like using map for a lot of stuff.
Dublin Devops Meetup
Yesterday, 25th March 2015, I’ve presented in our first Dublin DevOps Meetup, that happened on Dublin Zendesk’s Office.
My talk was about my experience using Chef, what’ve learned over the years, my mistakes and how I solved some of my problems.
I would like to say thanks to everyone that came to Zendesk’s office to see our talks.
The presentation is available at my github.
Thanks!
Caching range requests using NGINX at MaxCDN
On the end of 2013 I’ve met two amazing guys Justin Dorfman and David Archibald from MaxCDN, they were looking for a way to cache 206 Partial content.
As some of you may be aware, partial content caching using squid/varnish/nginx does not work as expected, the cache needs to download the entire file before returning partial content, and until the file is not downloaded, it returns 200.