It's been almost a year since I decided to set up Kubernetes on my home server, and I have to say, the experience has been a real rollercoaster. Let me share some of the highlights (and lowlights) with you.
On my personal microk8s cluster I unfortunately recently had a power failure affecting all nodes. I was using the default settings which create a file on /data/xxx.img. Unfortunately this file was missing from all my nodes.
I got a bus error when the postgres operator was running the bootstrap script. This happens when you have hugepages enabled (which is required to run mayastor). Many people claim you can simply disable huge_pages in postgres but I was not able to get this working. Instead, I simply just added a small allocation of hugepages to the postgres instance:
Over at Fenerum we are currently translating our software into multiple new languages, and while it worked fine for a while editing .po files by hand, it is now becoming a roadblock to getting things done.
I wanted a way to track the various websites I host in a privacy friendly way, so I decided on umami.is. Now, how do you set up umami on k8s?
After deciding to migrate to Kubernetes at Fenerum, I wanted to set up a homelab to play around in a safe environment and at the same time take the opportunity to start hosting various services myself.
Being a startup, it's always essential to decide what you want to spend your time on. For us, it's essential that spend as much time as possible on improving our product and talking to our customers - in short, creating value for our customers. That's why we initially outsourced our entire hosting setup to Heroku, although we have plenty of in-house hosting/ops knowledge. Customers don't care much where we are hosting our solution, as long as it's safe and has a high uptime.
When I previously used to use Postgresql.app I received the error $ psql
psql: could not connect to server: No such file or directory
Is the server running locally and accepting
connections on Unix domain socket "/tmp/.s.PGSQL.5432"? pretty frequently, so I decided to switch to the homebrew version.
While this is not required, I have chosen to re-install my mac from scratch with the new OS X Mavericks. This gave me the opportunity to correct a few less smart decisions i took previously. Here I'll describe the process it took for me, to get OS X Mavericks properly running - skip the reinstall part, if you just want to upgrade using the App Store.
During the sprints at Djangocon I finally had the opportunity to work a bit on django-hvad. The package now supports Django 1.5. This release will probably be the last before a 1.0 release that will drop the 'nani' compatibility module. The 0.3 release is mainly a stability and performance update.