DevOps Notes

One of my favorite DevOps tools, Chef, is gaining popularity by the day. Github alone has over a thousand Chef cookbooks publicly accessible.

Starting to use Chef is easy. Realizing the full power of Chef is not quite as easy. Many Chef cookbooks that I looked at can be significantly improved. Cookbooks created by developers can be improved by using systems engineering patterns. Cookbooks created by system administrators can be improved by using software engineering patterns.

Opscode does a very good job describing building blocks of Chef cookbooks. Explaining how to write good Chef cooks is something I have not seen done yet. In this post I will go over my cookbook writing process using a syslog-ng cookbook as an example.

This story is from a year ago. I kept telling myself that I got to write a blog post about it and here it is: a continuous deployment troubleshooting case.

One of the application server pools running a LAMP application (Linux / Apache/ MySQL / Memcached / PHP) was showing the following issue. Every time new code was deployed to the pool, Apache processes on 10-15 random servers out of 300+ would deadlock and stop serving requests. The problem was happening for some time but was not perceived as a big issue. Random Apache deadlocks did not impact the end users, and restarts of deadlocked Apache instances returned services to the operational state. I was not very comfortable with this behavior, so I decided to dig in and see what was going on.

With software development teams universally embracing Agile project management methodology, the question about Agile in Operations comes - up - frequently. Many view "agile sysadmin" as a key component of the DevOps movement. Recently, I had to give an answer to the "agile in operations" questions to my coworkers. Here's my take on it.

DevOps Days 2010 had a great pannel discussion on Infrastructure as Code. The responses to one question (asked at 27:30) did surprise me. The question was "Who here goes to kernel.org grabs the sources and compiles them?" No a single panelist raised a hand. I find it shocking that all distinguished DevOps evangelists on the panel are running stock distribution kernels.

Problem

Requests to a LAMP-based Facebook application are load balanced between server nodes by an F5 BIG-IP Local Traffic Manager (LTM). The F5 BIG-IP LTM has two components: an ASIC-based fast-switching component and an AMD Opteron-based software-switching component. Layer 4 load balancing between the application server nodes can be handled either exclusively by the F5 ASIC (Performance L4 mode), or exclusively by F5 software components (Standard mode), or by a combination of F5 ASIC and F5 software components. With the Standard mode turned on, F5 BIG-IP LTM capacity utilization is at 60%. With the Performance L4 mode turned on, F5 BIG-IP LTM capacity utilization is at 20%. However, with the Performance L4 mode turned on, developers report seeing an occasional blank Facebook canvas page returned in response to legitimate application requests. Serving blank Facebook canvas responses at any significant rate is unacceptable. The issue with the blank Facebook canvas responses in Performance L4 mode has to be addressed.

What's a DevOps blog without the discussion about what DevOps is?

If nothing else, DevOps is a buzzword that means different things to different people.

Ted Dziuba thinks that DevOps is a trend of nonsense "where system administrators start writing unit tests and other things to help the developers warm up to them." DevOps is certainly not about system administrators writing unit tests. DevOps is not about Ted Dziuba learning "Operations" (aka UNIX) either. DevOps is about system administrators, Ted Dziuba, and developers working closely together sharing knowledge about unit tests, UNIX, etc. to produce and deliver products.