This message was deleted.
# k3s
a
This message was deleted.
c
no. any CLI flag is a valid key in config.yaml. The config file is literally just translated to CLI flags during startup.
so if you look at the server or agent docs, or run
k3s server --help
, any of the flags listed there can go in the config file.
b
gotcha, thanks for the tip 🙂 also is this still valid?
Copy code
kube-controller-manager-arg:
- "address=0.0.0.0"
- "bind-address=0.0.0.0"
kube-proxy-arg:
- "metrics-bind-address=0.0.0.0"
kube-scheduler-arg:
- "address=0.0.0.0"
- "bind-address=0.0.0.0"
etcd-expose-metrics: true
trying to switch from systemd service flags to that config file
Copy code
disable:
- "traefik"
- "local-storage"
flannel-backend: "none"
etcd-s3-endpoint: "<http://s3.us-west-000.backblazeb2.com|s3.us-west-000.backblazeb2.com>"
etcd-s3-access-key: "xxxxx"
etcd-s3-secret-key: "yyyyy"
etcd-s3-bucket: "k3s-etcd"
etcd-s3-region: "us-east-1"
kube-controller-manager-arg:
- "address=0.0.0.0"
- "bind-address=0.0.0.0"
kube-proxy-arg:
- "metrics-bind-address=0.0.0.0"
kube-scheduler-arg:
- "address=0.0.0.0"
- "bind-address=0.0.0.0"
etcd-expose-metrics: true
c
yeah that all looks fine
b
should be, but apparently
address=
is no longer valid but
bind-address=
appears to work
c
I mean, they do have to be valid arguments for those components
you can put whatever args you want in there, k3s doesn’t validate them, just passes them through to Kubernetes
if you try something that doesn’t exist, or give it a bad value, it will probably blow up
b
i did and it does 🙂
btw, thanks so much for the help this is what i was coming from https://github.com/k3s-io/k3s/issues/3619#issuecomment-878501106
c
yeah, be careful copy-pasting stuff out of comments… not everyone that suggests something knows what they’re doing lol
b
a significant amount of my homelab is probably copy paste from github issue comments
but then again my downtime doesnt cost me anything
uuuuuuuh i'm using argocd to manage stuff in my cluster is there a way to codify k3s config in a repo and have argocd apply it?
maybe with system-upgrade-controller?
c
Not really… SUC is more about upgrading the binary.. for host configuration management I think people usually use things like ansible/puppet/teraform?
b
yea, that sounds reasonable