Appendix A: Cheat Sheet
Everything important from Days 0–15 on a few pages. The number in
parentheses is the day on which you practiced it. When in doubt,
ranchermanager.docs.rancher.com,
fleet.rancher.io, docs.rke2.io,
docs.k3s.io and
kubectl explain <typ>.<feld> are
authoritative.
Start of the Day and Training Ground
source ~/rancher-lab/setup/env.sh # LAB, VM addresses from vms.env (0, 2)
export KUBECONFIG="$LAB/kube/rancher.yaml" # local cluster = Rancher (3)
kubectl --kubeconfig "$LAB/kube/werk.yaml" … # target cluster through Rancher (5)
R="https://rancher.$RANCHER.sslip.io"
H="Authorization: Bearer $(cat "$LAB/rancher-token")" # API token (5)| Command | Purpose | Day |
|---|---|---|
rdctl start --container-engine.name=containerd … |
Start Rancher Desktop without a window | 0 |
nerdctl --namespace k8s.io build -t notes-api:0.2.0 … |
Build an image directly for k3s | 1 |
curl -sfL https://get.k3s.io with
INSTALL_K3S_VERSION |
Install k3s | 2 |
k3s ctr images import datei.tar |
Get images into the cluster without a registry | 2 |
helm install rancher rancher-stable/rancher --set hostname=… |
Install Rancher | 3 |
Rancher Objects
| Object | Meaning | Day |
|---|---|---|
clusters.provisioning.cattle.io (fleet-default) |
Create or import a cluster | 4, 5 |
clusters.management.cattle.io (c-m-…) |
Rancher’s view, labels for Fleet, k3sConfig | 4, 9, 12 |
clusterregistrationtokens.management.cattle.io |
Import and node commands, placeholder {token} |
4, 5 |
machines.cluster.x-k8s.io |
One node of a created cluster | 5 |
projects.management.cattle.io |
Project with quota and defaults | 6 |
projectroletemplatebindings.management.cattle.io |
Permissions in the project | 6 |
globalrolebindings.management.cattle.io |
Global role, such as user-base |
6 |
clusterrepos.catalog.cattle.io |
Helm catalog, per cluster | 7 |
gitrepos.fleet.cattle.io,
clustergroups.fleet.cattle.io |
GitOps with Fleet | 8, 9 |
etcdsnapshots.rke.cattle.io |
etcd snapshot of a cluster | 13 |
backups.resources.cattle.io,
restores.resources.cattle.io |
Back up Rancher | 13 |
tokens.management.cattle.io |
API tokens with ttl and clusterName |
14 |
API Calls
# Log in and save the token (5)
curl -sk -X POST "$R/v3-public/localProviders/local?action=login" \
-H 'Content-Type: application/json' \
-d '{"username":"admin","password":"…","responseType":"json"}' | jq -r .token
# kubeconfig for a cluster (5)
curl -sk -X POST -H "$H" "$R/v3/clusters/$ID?action=generateKubeconfig" | jq -r .config
# Create a user (6)
curl -sk -X POST -H "$H" -H 'Content-Type: application/json' "$R/v3/users" -d '{…}'
# Install or upgrade an app (7)
API="$R/k8s/clusters/$ID/v1/catalog.cattle.io.clusterrepos/<repo>"
curl -sk -X POST -H "$H" -H 'Content-Type: application/json' "$API?action=install" \
-d '{"charts":[{"chartName":…,"version":…,"releaseName":…,"values":{…}}],
"namespace":…,"wait":true,"timeout":"600s"}'
# Versions that Rancher knows (12)
curl -sk -H "$H" "$R/v1-rke2-release/releases" | jq -r '.data[].id'Fields in the Cluster Object
| Field | Effect | Day |
|---|---|---|
spec.kubernetesVersion,
rkeConfig.machineGlobalConfig |
Version and RKE2 configuration | 5 |
rkeConfig.registries.configs.<host> |
Login and CA for registries (caBundle: base64 PEM) | 7 |
rkeConfig.upgradeStrategy.workerDrainOptions |
Drain before the upgrade | 12 |
rkeConfig.etcdSnapshotCreate.generation |
Request a snapshot | 13 |
rkeConfig.etcdSnapshotRestore {name, generation, restoreRKEConfig} |
Restore | 13 |
localClusterAuthEndpoint.enabled |
Direct context without Rancher | 14 |
defaultPodSecurityAdmissionConfigurationTemplateName |
Pod Security for the cluster | 14 |
Fleet
# fleet.yaml (8, 9)
defaultNamespace: notes
helm:
releaseName: notes
repo: https://gitea…/api/packages/lab/helm # or chart: oci://…
chart: notes
version: 0.3.1
values: {…}
targetCustomizations:
- name: k3s
clusterSelector: {matchLabels: {provider.cattle.io: k3s}}
helm: {values: {…}}| GitRepo field | Purpose | Day |
|---|---|---|
targets: [{clusterName}] or
[{clusterGroup}] |
Where to | 8, 9 |
clientSecretName, helmSecretName,
caBundle |
Access to Git and Helm repo | 8 |
correctDrift.enabled |
Revert manual changes | 8 |
paused |
Read commits, don’t apply them | 8 |
paths |
Only these folders (without it: the whole repo) | B |
forceSyncGeneration |
Force a sync | B |
Failure Patterns
| Symptom | Cause | Fix |
|---|---|---|
cluster not found or invalid token on import |
{token} not replaced |
Token from Secret crt-token-default-token (4) |
READY true, machines Provisioning |
Column turns green too early | Wait for condition Ready (5) |
Pods missing, serviceaccount "notes" not found |
Platform part missing | namespace.yaml first (1, 2) |
| Namespace invisible to project member | Created without projectId annotation |
Create with annotation (6) |
Quota 0/0 in the new namespace |
Project budget exhausted | Adjust budget or share (6) |
x509: malformed certificate on the ClusterRepo |
PEM instead of DER | openssl x509 -outform der (7) |
| ClusterRepo invisible in werk | Created in the local cluster | Create in the target cluster (7) |
db-0 Pending, Helm timeout |
No StorageClass | local-path or Longhorn beforehand (7) |
GitRepo Modified |
Manual change in the cluster | correctDrift (8) |
| Label has no effect on imported cluster | Only on the provisioning object | Set on the management cluster (9) |
updates to statefulset spec … are forbidden |
StorageClass changed | Default StorageClass, backup/restore (10) |
| Grafana “Page not found” through Rancher | Proxy path | kubectl port-forward (11) |
GitPolling=False commit not found for branch |
Wrong branch | spec.branch (15) |
2/2, but commit has no effect |
Build job failed or paused |
Job log, WaitApplied (15) |
db-0 stuck after etcd restore |
Volume still on the old node | Delete the pod (13) |
ServiceUnavailable through Rancher |
Rancher gone | Context werk-werk1 (14) |
local-path PVC Pending under restricted |
Helper pod not hardened | Longhorn or exemption (14, 15) |