Our great sponsors
-
WorkOS
The modern identity platform for B2B SaaS. The APIs are flexible and easy-to-use, supporting authentication, user identity, and complex enterprise features like SSO and SCIM provisioning.
Been running k3s for personal projects etc for some time now on a cluster of raspberry pies. Why pies? Were cheap at the time and wanted to play with arm. I don’t think I would suggest them right now. Nucs will be much better value for money.
Some notes:
Using helm and helmfile https://github.com/helmfile/helmfile for deployments. Seems to work pretty nicely and is pretty flexible.
As I’m using a consumer internet provider ingress is done through cloudflare tunnels https://github.com/cloudflare/cloudflare-ingress-controller in order to not have to deal with ip changes and not have to expose ports.
Persistent volumes were my main issue when previously attempting this, and what changed everything for me was longhorn. https://longhorn.io Make sure to backup your volumes.
Really hyped for https://docs.computeblade.com/ xD
Related posts
- Diskomator – NVMe-TCP at your fingertips
- Difference between snapshot-cleanup and snapshot-delete in Longhorn recurring job?
- Container redundancy with multiple Unraid servers?
- How do I clean up a Longhorn volume? Trimming the volume doesn't work, "cannot find a valid mountpoint for volume"
- Is there a tool to manage the resources of individual containers in an easy, visual way? Currently containers devour every VM resource, when I set 20GB RAM, they will consume 20GB. I want to set limits for each container, doing it in docker compose is inconvenient.