The easiest way to bootstrap a self-hosted High Availability Kubernetes cluster. A fully automated HA k3s etcd install with kube-vip, MetalLB, and more. Build. Destroy. Repeat.
Currently, it is not possible to use multiple control nodes where the network interfaces to use have different names. This is due to the fact that {{ flannel_iface }} of the first control node will be hardcoded into the kube-vip manifest as the interface that all kube-vip pods must use.
kube-vip is able to pick a matching interface per pod automatically for a "long time".
This patch enables the automatic choice of interface by default and offers users the ability to override the interface if they want to (independently of flannel_iface, as requested in #221.
Checklist
[x] Tested locally
[x] Ran site.yml playbook
[x] Ran reset.yml playbook
[x] Did not add any unnecessary changes
[x] Ran pre-commit install at least once before committing
timothystewart6
commented
1 year ago
Proposed Changes
{{ flannel_iface }}of the first control node will be hardcoded into the kube-vip manifest as the interface that all kube-vip pods must use.flannel_iface, as requested in #221.Checklist
site.ymlplaybookreset.ymlplaybook