- hark
- May 10, 2023
-
I'm sleep
|
Getting some raspberry pi's and building a K8s cluster that you can update containers without downtime is the best way to do this. Running through some online courses is probably good enough for what you're looking at though.
do you have any resources for how to do this that you think are particularly good? I have a bunch of rpis but I don't know anything about k8s really and would love to put those to use learning it.
|
#
¿
Oct 27, 2023 19:44
|
|
- Adbot
-
ADBOT LOVES YOU
|
|
#
¿
May 15, 2024 01:58
|
|
- hark
- May 10, 2023
-
I'm sleep
|
for a group of raspberry pi I would use k3s. it has an installation script
https://github.com/k3s-io/k3s
k3s is a magic self extracting archive that has swiss army knife functionality similar to like busybox. k3s is kubernetes, but it's a total rewrite for exactly the kind of home lab stuff you're doing that simplifies initial setup and will get you running in under half an hour
resources for running k3s:
1. install ubuntu 22.04 on your raspberry pi (3 or more ideally, one will be the master/controller, the other two act as HA nodes) then
2. run the script in the github url above and get the three nodes talking to eachother.
3. save off your kube.config file to a safe place and make sure you can access the cluster, you'll need this later
4. install helm
5. find some recent helm charts and get prometheus and grafana installed on your cluster and browse to your node's IP address and try using them
6. come join us in the devops thread https://forums.somethingawful.com/showthread.php?threadid=3695559
I recommend prometheus and grafana because they're sort of the "hello world" of helm charts and also get you started on the right foot with having visibility into monitoring your cluster/performance
more edit: https://medium.com/globant/setup-prometheus-and-grafana-monitoring-on-kubernetes-cluster-using-helm-3484efd85891
fantastic! thanks!
|
#
¿
Oct 27, 2023 20:19
|
|