Just like CRDs extend the API surface of k8s, the YANG schema owned by a certain application on SR Linux extend the NOSes API.

We onboard applications separately one from another, each app comes with its own YANG model, and all models are then joined under a single root

Today was a great day for Clabernetes/containerlab and open-source networking labs.
Carl Montanari and Simon Peccaud introduced Clabernetes to RIPE 88 attendees.

OVHCloud was an early adopter of Clabernetes, which helped them to simulate the WAN network.

Still feels surreal that one can work on preparing 70 VMs for Autocon1 attendees while being 10k feet above the ground 🧐

Pst, I drained the private cloud, no more 6cpu/24GB VMs left, oopsie.

I also went and quickly checked what would it cost me to run 70 8vcpu/32GB VMs in

The way I do my ansible
✅ hermetic reproducibility, same ansible version all the time
✅ zero install harassment. No deps hell
✅ OS agnostic, ~whenever I want to run - it runs
✅ Easy ansible version switch. Change a tag string and you're off

And yes, I am building/maintaining

When the conference season starts you better be ready

Been trying this Oracle Cloud and it's free tier instances to support my workshops with a small and publicly accessible VM.

All worked amazingly well for 2 months until one evening the instance was removed and account deactivated. Apparently Oracle does it often

And while having an accessible image is a first step to make, the other one is making your image consumable.

If a single instances demands 8vcpu/32gb mem and 15 minutes to boot -- you are doomed to have a negative overall experience

In may I'll have my first pair of the real back to back business trips.

Landing in AMS from LAX to only have a a start of another trip to KRK

AMS-SFO
SFO-LAX-AMS-KRK
KRK-AMS
🤤

Ah, sweet, GitHub topics allow you to create a personalized topic header when the topic is used by the community.

Now clab-topo has a welcoming header with the logo and the whole shebang 🥰

Also you did great, we doubled the repo count since the announcement.

We expect to have 50+ Containerlab workshop attendees at Autocon1 next month.

With a crowd that big, there is no way, everyone is going to see clearly what happens on the screen. So I decided to pull a little trick that should work perfectly 🤞

Besides having a normal slide

Hey all,
as you know, containerlab labs belong to Git, and GitHub is the most popular service for hosting public projects.

You all did a great job by creating labs, publishing them and showing what's possible. Thank you 🙏

Now, we wanted to make your labs easily

Systemd timers are better than cron

Don't know why exactly but browsing a network telemetry dashboard on a mobile feels oddly satisfying and somewhat illegal.

We will explore this topology with our Autocon1 workshop attendees

🫡Your infrequent reminder.

We are shaping the product for you, even if you just rant on twitter we listen.

Here is a 10min introduction to a feature that landed in 24.3 and was inspired by a user' response

📺 youtube.com/watch?v=9gr6L4…

This weekend's nights were dedicated to preparing the automation to build the infrastructure for the upcoming containerlab workshop at Autocon1.

I had an ambitious goal to give every attendee a decent-sized, personal VM so everyone has their own sandbox.

Pulumi,

Next week we will talk about the on-box automation and how it compares to the typical off-box approach.

I will present you a case of writing a trivial Discord bot client using SR Linux Event Handler framework.

50 lines of python and you are off to the races.

When you have a hammer, everything is a nail 😜

Topology is courtesy of netreplica