Frequently asked questions
Troubleshooting
Troubleshooting advice can be found on our Troubleshooting Cheatsheet.Configuration
How to Enable KubeSpan
Talos Linux provides a full mesh WireGuard network for your cluster.
To enable this functionality, you need to configure KubeSpan and Cluster Discovery in your Talos Linux configuration:
machine:
network:
kubespan:
enabled: true
cluster:
discovery:
enabled: true
Since KubeSpan encapsulates traffic into a WireGuard tunnel, Kube-OVN should also be configured with a lower MTU value.
To achieve this, add the following to the Cozystack ConfigMap:
apiVersion: v1
kind: ConfigMap
metadata:
name: cozystack
namespace: cozy-system
data:
values-kubeovn: |
kube-ovn:
mtu: 1222
Operations
How to enable access to dashboard via ingress-controller
Update your ingress application and enable dashboard: true option in it.
Dashboard will become available under: https://dashboard.<your_domain>
What if my cloud provider does not support MetalLB
You still have the opportunity to expose the main ingress controller using the external IPs method.
Take IP addresses of the external network interfaces for your nodes.
Add them to the externalIPs list in the Ingress configuration:
kubectl patch -n tenant-root ingresses.apps.cozystack.io ingress --type=merge -p '{"spec":{
"externalIPs": [
"192.168.100.11",
"192.168.100.12",
"192.168.100.13"
]
}}'
kubectl patch -n cozy-system configmap cozystack --type=merge -p '{
"data": {
"expose-external-ips": "192.168.100.11,192.168.100.12,192.168.100.13"
}
}'
After that, your Ingress will be available on the specified IPs:
# kubectl get svc -n tenant-root root-ingress-controller
root-ingress-controller ClusterIP 10.96.91.83 37.27.60.28,65.21.65.173,135.181.169.168 80/TCP,443/TCP 133d
How to cleanup etcd state
Sometimes you might want to flush etcd state from a node. This can be done with Talm or talosctl using the following commands:
Replace nodeN with the name of the failed node, for instance, node0.yaml:
talm reset -f nodes/nodeN.yaml --system-labels-to-wipe=EPHEMERAL --graceful=false --reboot
talosctl reset --system-labels-to-wipe=EPHEMERAL --graceful=false --reboot
How to generate kubeconfig for tenant users
Use the following script:
user=tenant-root
cluster=$(kubectl config get-contexts | awk '$1 == "*" {print $3}')
token=$(kubectl get secret -n "$user" "$user" -o go-template='{{ printf "%s\n" (index .data "token" | base64decode) }}')
kubectl config view --minify --raw > tenant-kubeconfig
kubectl config --kubeconfig tenant-kubeconfig unset users
kubectl config --kubeconfig tenant-kubeconfig unset contexts
kubectl config --kubeconfig tenant-kubeconfig set "users.$user.token" "$token" --set-raw-bytes=true
kubectl config --kubeconfig tenant-kubeconfig set "contexts.$user@$cluster.user" "$user"
kubectl config --kubeconfig tenant-kubeconfig set "contexts.$user@$cluster.namespace" "$user"
kubectl config --kubeconfig tenant-kubeconfig set "contexts.$user@$cluster.cluster" "$cluster"
kubectl config --kubeconfig tenant-kubeconfig set current-context "$user@$cluster"
in the result, you’ll receive the tenant-kubeconfig file, which you can provide to the user.
How to configure Cozystack using FluxCD or ArgoCD
Here you can find reference repository to learn how to configure Cozystack services using GitOps approach:
How to Rotate Certificate Authority
In general, you almost never need to rotate the root CA certificate and key for the Talos API and Kubernetes API. Talos sets up root certificate authorities with a lifetime of 10 years, and all Talos and Kubernetes API certificates are issued by these root CAs.
So the rotation of the root CA is only needed if:
- you suspect that the private key has been compromised;
- you want to revoke access to the cluster for a leaked talosconfig or kubeconfig;
- once in 10 years.
Rotate CA for a Tenant Kubernetes Cluster
See: https://kamaji.clastix.io/guides/certs-lifecycle/
export NAME=k8s-cluster-name
export NAMESPACE=k8s-cluster-namespace
kubectl -n ${NAMESPACE} delete secret ${NAME}-ca
kubectl -n ${NAMESPACE} delete secret ${NAME}-sa-certificate
kubectl -n ${NAMESPACE} delete secret ${NAME}-api-server-certificate
kubectl -n ${NAMESPACE} delete secret ${NAME}-api-server-kubelet-client-certificate
kubectl -n ${NAMESPACE} delete secret ${NAME}-datastore-certificate
kubectl -n ${NAMESPACE} delete secret ${NAME}-front-proxy-client-certificate
kubectl -n ${NAMESPACE} delete secret ${NAME}-konnectivity-certificate
kubectl -n ${NAMESPACE} delete secret ${NAME}-admin-kubeconfig
kubectl -n ${NAMESPACE} delete secret ${NAME}-controller-manager-kubeconfig
kubectl -n ${NAMESPACE} delete secret ${NAME}-konnectivity-kubeconfig
kubectl -n ${NAMESPACE} delete secret ${NAME}-scheduler-kubeconfig
kubectl delete po -l app.kubernetes.io/name=kamaji -n cozy-kamaji
kubectl delete po -l app=${NAME}-kcsi-driver
Wait for the virt-launcher-kubernetes-* pods to restart.
After that, download the new Kubernetes certificate.
Rotate CA for the Management Kubernetes Cluster:
See: https://www.talos.dev/v1.9/advanced/ca-rotation/#kubernetes-api
git clone https://github.com/cozystack/cozystack.git
cd packages/core/testing
make apply
make exec
Add this to your talosconfig in a pod:
client-aenix-new:
endpoints:
- 12.34.56.77
- 12.34.56.78
- 12.34.56.79
nodes:
- 12.34.56.77
- 12.34.56.78
- 12.34.56.79
Execute in a pod:
talosctl rotate-ca -e 12.34.56.77,12.34.56.78,12.34.56.79 \
--control-plane-nodes 12.34.56.77,12.34.56.78,12.34.56.79 \
--talos=false \
--dry-run=false &
Get a new kubeconfig:
talm kubeconfig kubeconfig -f nodes/srv1.yaml
For Talos API
See: https://www.talos.dev/v1.9/advanced/ca-rotation/#talos-api
All commands are like for the management k8s cluster, but with talosctl command:
talosctl rotate-ca -e 12.34.56.77,12.34.56.78,12.34.56.79 \
--control-plane-nodes 12.34.56.77,12.34.56.78,12.34.56.79 \
--kubernetes=false \
--dry-run=false &
Public-network Kubernetes deployment
A Kubernetes/Cozystack cluster can be deployed using only public networks:
- Both management and worker nodes have public IP addresses.
- Worker nodes connect to the management nodes over the public Internet, without a private internal network or VPN.
Such a setup is not recommended for production, but can be used for research and testing, when hosting limitations prevent provisioning a private network.
To enable this setup when deploying with talosctl, add the following data in the node configuration files:
cluster:
controlPlane:
endpoint: https://<MANAGEMENT_NODE_IP>:6443
For talm, append the same lines at end of the first node’s configuration file, such as nodes/node1.yaml.
How to allocate space on system disk for user storage
Default Talos setup assumes that each node has a primary and secondary disks, used for system and user storage, respectively. However, it’s possible to use a single disk, allocating space for user storage.
This configuration must be applied with the first
talosctl apply
or
talm apply
— the one with the -i (--insecure) flag.
Applying changes after initialization will not have any effect.
For talosctl, append the following lines to patch.yaml:
---
apiVersion: v1alpha1
kind: VolumeConfig
name: EPHEMERAL
provisioning:
minSize: 70GiB
---
apiVersion: v1alpha1
kind: UserVolumeConfig
name: data-storage
provisioning:
diskSelector:
match: disk.transport == 'nvme'
minSize: 400GiB
For talm, append the same lines at end of the first node’s configuration file, such as nodes/node1.yaml.
Read more in the Talos documentation: https://www.talos.dev/v1.10/talos-guides/configuration/disk-management/.
After applying the configuration, wipe the data-storage partition:
kubectl -n kube-system debug -it --profile sysadmin --image=alpine node/node1
apk add util-linux
umount /dev/nvme0n1p6 ### The partition allocated for user storage
rm -rf /host/var/mnt/data-storage
wipefs -a /dev/nvme0n1p6
exit
When the storage is configured, add the new partition to LINSTOR:
linstor ps cdp zfs node1 nvme0n1p6 --pool-name data --storage-pool data1
Check the result:
linstor sp l
Output will be similar to this example:
+---------------------------------------------------------------------------------------------------------------------------------------+
| StoragePool | Node | Driver | PoolName | FreeCapacity | TotalCapacity | CanSnapshots | State | SharedName |
|=======================================================================================================================================|
| DfltDisklessStorPool | node1 | DISKLESS | | | | False | Ok | node1;DfltDisklessStorPool |
| data | node1 | ZFS | data | 351.46 GiB | 476 GiB | True | Ok | node1;data |
| data1 | node1 | ZFS | data | 378.93 GiB | 412 GiB | True | Ok | node1;data1 |
Bundles
How to overwrite parameters for specific components
Moved to the Bundles reference.
How to disable some components from bundle
Moved to the Bundles reference.