Two years ago I shared my experience on building the AWS Lambda function for a python project of my own. And a few days ago I stumbled upon a nice opensource CLI tool that I immediately wanted to transform in a web service.

Naturally, a simple, single-purpose tool is a perfect candidate for function-as-a-service (FaaS), and since I had past experience with AWS Lambda, this time I decided to meet its Google’s sibling - Google Cloud Function.

In this post we’ll discover how to take a python package with 3rd party dependencies, make a GCP Function from it and deploy it without a single click in the UI - all without leaving the IDE.

[Project’s source code]

I love the tech conferences that share the recordings of the sessions without hiding behind the registration or a pay wall. Luckily the trend to share the knowledge with a community openly is growing, yet you still can find a nice talk hidden behind the above mentioned walls.

Sometimes an easy registration is all it takes, but then, how do you watch it offline? For example, I do love to catch up with with the recent trends and experiences while being on a plane, meaning that I just cant afford to be hooked up to the Internet.

If a talk is published on the YouTube, you are good to go and download it with any web service that pops up in the google search by the term “Youtube download”. But what do we do when the video is hosted somewhere in the CDN and is served as a dynamic playlist of *.ts files?

Here I share with you an easy way to download the videos from an m3u/m3u8 playlist.

While working on the Ipanema Wan Opt VNF integration with Nuage Networks I stumbled upon an interesting case which required to max out the network with FTP traffic. The tricky point there was to create the FTP connection which won’t be limited by the disk IO performance. Especially, considering that the disks were kind of slow in the setup I had.

It turns out, you can use the in-memory devices in the FTP communication path /dev/zero -> /dev/null, ruling out the slowliness that could have been added by the disks. Lets figure out how to do that!

Shortly after I passed AWS CSA exam I went on a hunt for the next certification to claim. Decided to tackle the Openstack COA certification first saving the Docker/Kubernetes certification for a later occasion.

There is a common joke floating around: “Oh, is Openstack still a thing?” - yes, its pretty much still a thing, especially in the Telecom area where VMs are the only viable option for the most of the cases (think VNFs).
Openstack also powers our teams public SDN lab - NuageX - that allows to provision a fully functional Nuage environment in a matter of minutes. So I wanted to get a better operational knowledge of Openstack to be able to support and tune the platform if necessary.

Disclaimer: I took the recently updated COA exam which is based on the Openstack Pike release, although I did not face any Pike-specific questions during the exam. This does not mean that the exam content will stay the same throughout the course evolution, so watch out for the updates.

On May 11th I passed the AWS Certified Solution Architect Associate exam which was harder then I expected. In this memo I will outline how I prepared to this exam, what topics you better pay more attention to and some tips and hints I could give to anyone going for AWS CSA exam.

Disclaimer: I took the original AWS CSA exam, not the one that was launched in Feb 2018; this older version is only available to schedule till August 2018. After that date the newer version of this exam will be the only one available. Watch out, it has a new set of objectives.

Have you ever tried to upload thousands of small/medium files to the AWS S3? If you had, you might also noticed ridiculously slow upload speeds when the upload was triggered through the AWS Management Console. Recently I tried to upload 4k html files and was immediately discouraged by the progress reported by the AWS Console upload manager. It was something close to the 0.5% per 10s. Clearly, the choke point was the network (as usual, brothers!).

Comer here, Google, we need to find a better way to handle this kind of an upload.

Back in the days when I mostly did routing stuff I spent the whole day configuring SROS devices via SSH. And once in a while I saw that SSH session or its server part (or even underlying connection) glitched, resulting in a corrupted lines feeded to the device.

What was also quite common is to make a mistake (i.e. syntax one) in a single line and watch like the rest of config got applied to the wrong context.

These sad facts pushed me to create a rootifier CLI script, that was converting tree-like SROS config into flattented (aka rooted) fashion.

Now I decided to make a web service of that script, that is publicly available at https://rootifier.netdevops.me/

Flask documentation is very clear on where is the place for its built-in WSGI application server:

While lightweight and easy to use, Flask’s built-in server is not suitable for production as it doesn’t scale well and by default serves only one request at a time.

So how about I share with you a Dockerfile that will enable your Flask application to run properly and ready for production-like deployments? As a bonus, I will share my findings discovered along the way of building this container image.

Today I faced a task which required first to establish an SSH tunnel in a background process and later use this tunnel for SSH connection. What seemed like a child’s play first actually had some fun inside.

A problem were hidden right between the moment you spawned ssh process in the background and the next moment you tried to use this tunnel. In other words, it takes literally no time to spawn a process in the background, but without checking that tunnel is ready, you will quite likely receive an error, since your next instructions will be executed immediately after.

Consequently, I needed a way to ensure that the SSH service is ready before I try to consume it.