Operate nodes¶
These pages are for operators who install and run chalkos nodes with chalkctl, YAML files and signed images. They apply to every node, apart from the sections on etcd, the API server and kubeconfigs, which need the Kubernetes layer. Their examples come from a Kubernetes cluster. Run Kubernetes sets up and runs the cluster itself, and Run a router covers what router nodes add. The command cheat sheet lists the command and the credential for each common task.
Set up¶
These pages go from a new workstation to the first installed node.
-
Set up your workstation
Get chalkctl, generate the cluster's secrets and choose the credentials each command needs.
-
Write a node's config
Give each node its files and image in the inventory, and check its documents.
-
Install a node
Boot a physical machine from the installer and install a node on it.
-
Install a node in a VM
Run a node under QEMU, libvirt or Proxmox from the kvm image, installed in place.
-
Enrol Secure Boot keys
Put the cluster's Secure Boot keys into a machine's firmware before it is installed.
Day 2¶
These pages cover nodes that run: config changes, new and removed nodes, upgrades, access, storage, registries and the tasks that recur.
-
Change a node's config
Deliver changed YAML to a running node, preview it, and act on a refusal.
-
Add and remove nodes
Add a node to a running cluster, reinstall one, or take one out for good.
-
Upgrade nodes
Roll a new image through the nodes, and act on a node that rolls back.
-
Manage access and certificates
Give and revoke access, renew node certificates, and rotate the OS CA and the node CA.
-
Manage secrets
Generate and rotate the cluster bundle, and source credentials for node services.
-
Add storage volumes
Size VAR, and add encrypted or raw volumes on the system disk or on other disks.
-
Registries and mirrors
Give nodes the credentials of a private registry, and pull images through a mirror.
-
Recurring tasks
What expires when, what to back up, and how to move the cluster to a newer chalkos.
Diagnose and recover¶
These pages read what a node does and bring back a node that fails.
-
Status, logs and the console
Read what
chalkctl statusreports, follow a unit's journal, run the image's diagnostics and reach a node's tools. -
The console on the machine's screen
Read a node at its screen, and reboot, shut down or roll it back there.
-
Recover a node
Bring back a node that does not unlock, boot or renew, or whose secrets file is lost.
-
Troubleshooting
Find the cause of a node's symptom and the page that fixes it.