Skip to main content


We had our quarterly Hackday at Open Earth Foundation today. It's a day where we don't have any meetings or do any regular tasks. We just get a day to exercise our creativity and make cool things with and for each other.
Some of the projects my colleagues worked on:

- a conference on Open Source and climate change
- a Jira plugin that scores tickets for team effort and mental state
- a dashboard for evaluating biodiversity at a national level
- an AR visualization of climate data on a spinning globe
- a rapid prototyping gap analysis tool
- a documentation section on our wiki for leadership and communications skills
Me, I decided to set up a Mastodon instance for our team.

Total goose egg. I got absolutely nowhere.

I started off making my own K8S layout and after a couple of hours realised I should use a Helm chart instead.

I tested a few Helm charts on my local Docker Desktop, and none worked. So I decided to set up an EKS cluster to run it instead.

Cue two hours of dorking around with IAM JSON. I wasn't ever able to use Helm with the half-broken cluster I got working.
With an hour left, I decided to spin up an EC2 instance, run microk8s, and see if that would work.

I got Helm to run the install, but it kept failing. After multiple passes through values.yml, I ended up with about half the pods running, and all the important ones mysteriously silent.

It sucked!

I'm going to give it another try over the holidays, but jeez. I think I might be doing it wrong.
The only way I've ever managed is to install yunohost and then install mastodon as an app.
it's nice to read where things don't go smoothly.

I can really empathise with the well maybe just put it on a VM 😀
@tomwj I don't yet have the skills to debug Kubernetes and Helm errors. A lot of the time, I'm just looking at a status like "NotReady" and wondering what it means and how I got there and what handle I can jiggle to try to make it stop.
i love cloud-native architecture in principle but in practice, you will invariably spend more time troubleshooting your deployment than working with your actual project
@selfagency it's nice when it just goes tootling along on its on for months on end, though.
I heard that @brianleroux is working on a serverless ActivityPub server. You should bug him to release it 😜
@schizanon@brianleroux neat! I tried to do one on Amazon Lambda a few years ago but ran out of gas.
@schizanon it's much nicer these days but definitely can be a lot to learn (and unlearn from long lived workloads)
Yeah, Mastodon is not AT ALL easy to set up. NOT AT ALL.

It should be, though.
where do we access such goodness?
@helloronaut I think everything is on our GitHub. We usually push Hackday projects even if they're not finished.

https://github.com/Open-Earth-Foundation/
I need to know more about the #dashboard for #biodiversity, I know some tries in this way, id’like to compare.
#Cop15
#speciesextinction
Those all sound wonderful! Do any of these projects happen to have public links to check them out? I study environmental science and would love to see any prototypes!
OEF ever work with contractors?