Skip to main content

Your submission was sent successfully! Close

Thank you for signing up for our newsletter!
In these regular emails you will find the latest updates from Canonical and upcoming events where you can meet our team.Close

Thank you for contacting us. A member of our team will be in touch shortly. Close

  1. Blog
  2. Article

Jorge O. Castro
on 6 July 2015


Canonical sent three Juju Ecosystem developers and one Juju core developer to Dockercon 2015 in San Francisco during which we participated in a multi-company hackathon team; threw a marketing event featuring industry luminaries; and made connections with many potential charm authors and partners.

We learned loads, preached a lot of good Juju, were impressed what individuals and companies were doing and the collaborative spirit of the event in general. Here’s a bit more detail on what we got up too….

Hackathon

Rancher sponsored a 24 hour hackathon before the conference got underway. This was a really good experience and the grand prize was ticket to the Dockercon 2015 conference (which most of us had already).

Canonical’s Matt Bruzek, Wayne Witzell III and Whit Morriss participated in a pick-up team that included folks from Seed Digital, Capital One, Monsanto, and Michigan.

The idea

The morning of we all ended up sitting at the same table discussing what we were going to work on separately. Whit was blabbing about a talk by Roy Rapoport from Netflix at Monitorama and canaries, OODA loops, and basic metrics for implementing automated canary systems. The group decide to build a dj mixer for doing blue-green releases based on Rapoport’s suggested metrics.

What we built

All the code is hosted under the system-zoo group on GitHub. We created an ec2 environment using Juju, used it to set up some base architecture and provide our team with ssh access to all the machines involved. From the Juju world we used the following charms:

  • Consul
  • Registrator
  • Docker
  • Elasticsearch
  • Heka-hub, Heka-sub

Outside Juju-land, we used Docker to deploy the bits we were hacking up at the moment:

  • cross-fading service aware scala loadbalancer
  • nginx to serve the static content as well as provide the API proxying and service load balancing for Consul, Heka, and ElasticSearch.
  • the test app which had configurable latency and random failure

We used Git and GitHub for preserving ad hoc config and used Docker hub for delivering the test app and the load balancer.

What we learned about Juju in this situation

While this an extremely ambitious project, we learned that by standing on the shoulders of giants, we could actually do something that surprised ourselves with almost no prior planning and little prior experience working together. Juju’s ability to rapidly create a secure shared environment populated with mostly working basic required services made a big difference here. Granted, we did not try to set up the same infra using Docker only tools, but under the circumstance, having full machine access made debugging much faster when there were hiccups. We’d recommend folks participating in high time pressure events using Juju, it’s a great way to expose usability warts and other issues.

What we learned in general

Hackathon’s are a great place to meet people, talk shop, and learn what’s happening outside Canonical. The vibe was collaborative, super friendly and it was all around a positive experience both as individuals and Canonical representatives.

Evening of Orchestration

Canonical organized an event the night before Dockercon, it was held at Yelp headquarters. Organizers were Whit Morriss, Randall Ross, Jorge Castro, and Julie Nguyen (from Yelp).

The event was well attended despite not being on the official Dockercon roster and being several blocks away from the official hotel. The venue was great and Yelp was an awesome host. Sam Eaton from Yelp MC’ed the lightning talks which include talks from Yelp, Crate.io, Metaswitch, Clusterhq, Glider Labs, Chuck Butler, and Weave.

We presented a panel of leaders in the orchestration industry including:

  • John Willis (Moderator) from Docker Inc previously Socketplane
  • Adrian Cockcroft from Battery Ventures
  • Ben Saller from Canonical
  • Bill Farner from Twitter
  • Brian Grant from Google who worked on Borg and Omega, now k8s
  • Jeff Lindsay from Glider Labs

Ben did a great job of talking about the principles of Juju and the need for a language for modelling infrastructure to a very positive response from the other panelists and the crowd.

Many people who were taking different approaches to orchestration at a high level, the model that Juju is working to achieve was well received.

Summary

The attitude, the content, the crowd, the production were all impressive. For such a big conference, people were all super friendly and inclusive (from community celebs like Hykes on down). There was a really good overall collaborative community vibe and we were definitely recognized and welcomed coming from Canonical.

Related posts


sergiodj
18 November 2024

Profile-guided optimization: A case study

Ubuntu Article

Software developers spend a huge amount of effort working on optimization – extracting more speed and better performance from their algorithms and programs. This work usually involves a lot of time-consuming manual investigation, making automatic performance optimization a hot topic in the world of software development. Profile-guided opt ...


Mita Bhattacharya
6 November 2024

Meet Canonical at KubeCon + CloudNativeCon North America 2024

Cloud and server Article

We are ready to connect with the pioneers of open-source innovation! Canonical, the force behind Ubuntu, is returning as a gold sponsor at KubeCon + CloudNativeCon North America 2024.  This premier event, hosted by the Cloud Native Computing Foundation, brings together the brightest minds in open source and cloud-native technologies. From ...


Andreea Munteanu
1 November 2024

Charmed Kubeflow vs Kubeflow

AI Article

Why should you use an official distribution of Kubeflow? Kubeflow is an open source MLOps platform that is designed to enable organizations to scale their ML initiatives and automate their workloads. It is a cloud-native solution that helps developers run the entire machine learning lifecycle within a single solution on Kubernetes. It can ...