infra/talos/README.md
Graham McIntire 39bedb08d1
renumber infrastructure from 10.0.15.0/24 to 10.0.16.0/22 (hosts in 10.0.19.x)
Proxmox: node1-3 → 10.0.19.101-103
Talos cp1-3 → 10.0.19.1-3, workers → 10.0.19.4-6
K8s endpoint → VIP https://10.0.19.10:6443
Ansible: prom → 10.0.19.31, db → 10.0.19.30
Talos: added VIP block to controlplane.yaml base config
Promtail: Loki URL → 10.0.19.31
Docs: all references updated, talos4 removed
2026-07-18 08:28:47 -05:00

139 lines
3.4 KiB
Markdown

# Talos Home Cluster
## Cluster Information
- **Cluster Name**: home-cluster
- **Kubernetes Version**: v1.36.1
- **Control Plane Nodes**: 3 (10.0.19.1-3)
- **Worker Nodes**: 3 (10.0.19.4-6)
- **API Endpoint**: https://10.0.19.10:6443 (shared VIP)
## Node Details
| Node Name | Role | IP | Proxmox Node | VMID |
|-----------|------|-----|--------------|------|
| talos-cp1 | Control Plane | 10.0.19.1 | node1 | 200 |
| talos-cp2 | Control Plane | 10.0.19.2 | node2 | 201 |
| talos-cp3 | Control Plane | 10.0.19.3 | node3 | 202 |
| talos-worker1 | Worker | 10.0.19.4 | node1 | 210 |
| talos-worker2 | Worker | 10.0.19.5 | node2 | 211 |
| talos-worker3 | Worker | 10.0.19.6 | node3 | 212 |
## Configuration Files
- `secrets.yaml` - Cluster secrets (keep secure!)
- `controlplane.yaml` - Base control plane config (applied with per-node patches)
- `worker.yaml` - Base worker config (applied with per-node patches)
- `patches/*.yaml` - Per-node patches (hostname, disk overrides)
- `talosconfig` - Talosctl client configuration
## Using the Cluster
### Talosctl Commands
```bash
# Check cluster health
talosctl --talosconfig talosconfig health
# Get node list
talosctl --talosconfig talosconfig get members
# View logs from a node
talosctl --talosconfig talosconfig logs -f kubelet
# Check service status
talosctl --talosconfig talosconfig service
# Upgrade Talos (example)
talosctl --talosconfig talosconfig upgrade --image ghcr.io/siderolabs/installer:v1.13.3
```
### Kubectl Commands
The kubeconfig has been merged into your default kubectl config:
```bash
# View nodes
kubectl get nodes
# View all pods
kubectl get pods -A
# View cluster info
kubectl cluster-info
```
## Storage Configuration
VM boot disks are stored on local-lvm on each Proxmox node (no shared Ceph storage). For Kubernetes persistent storage, deploy a CSI provisioner (e.g. Rook-Ceph, OpenEBS, or cloud provider CSI).
## Networking
- **CNI**: Flannel (deployed by default)
- **Service CIDR**: Default Kubernetes range
- **Pod CIDR**: Default Kubernetes range
## Maintenance
### Adding a Node
1. Clone VM from template in Terraform
2. Create a per-node patch in `patches/` with the hostname
3. Apply base config with patch:
```bash
talosctl --talosconfig talosconfig apply-config --insecure --nodes <NEW_NODE_IP> --file worker.yaml --config-patch @patches/<new-node>.yaml
```
### Removing a Node
```bash
# Drain node first
kubectl drain <node-name> --ignore-daemonsets --delete-emptydir-data
# Remove from cluster
kubectl delete node <node-name>
# Reset Talos on the node
talosctl --talosconfig talosconfig reset --nodes <NODE_IP>
```
### Upgrading Kubernetes
```bash
talosctl --talosconfig talosconfig upgrade-k8s --to 1.36.1
```
## Troubleshooting
### Node not joining cluster
```bash
# Check node logs
talosctl --talosconfig talosconfig logs -f kubelet --nodes <NODE_IP>
# Check etcd members
talosctl --talosconfig talosconfig etcd members
```
### Reset a Node
```bash
talosctl --talosconfig talosconfig reset --nodes <NODE_IP> --graceful=false
```
### Reboot a Node
```bash
talosctl --talosconfig talosconfig reboot --nodes <NODE_IP>
```
## Backup
Important files to backup:
- `secrets.yaml` - Required to add nodes or recover cluster
- `talosconfig` - Required for talosctl access
- `controlplane.yaml` and `worker.yaml` - Base node configs
- `patches/` directory - Per-node patches
- Kubeconfig is stored in `~/.kube/config`
Store these securely in a password manager or encrypted backup.