Validation

Validate Core and the required Aligned Extensions after the installation completes and before you continue with post-install tasks. These checks apply to both installation paths. If the global cluster runs Alauda OS, also complete Additional Checks for Alauda OS.

Prerequisites

  • The installation completed.
  • You can access the platform Web UI address. On a non-Alauda OS, it is the address that the installer shows. On Alauda OS, it is global.platform.address from install-config.yaml.
  • You can run kubectl against the new global cluster. On a non-Alauda OS, run it from the installation node. On Alauda OS, use a kubeconfig for the new global cluster. The bootstrap node has no kubectl. To run the commands on this page there, use <extract-dir>/res/<arch>/ac --kubeconfig <install-dir>/output/kubeconfig in place of kubectl.

Validate Web UI Access

After the installation is complete, open the platform Web UI and verify that the platform login page is reachable. On a non-Alauda OS, the installer displays the platform access URL. Click Access to open it. On Alauda OS, open https://<platform-access-address>, where <platform-access-address> is global.platform.address.

Validate Platform Components, Extensions, and Pods

Run the following commands against the new global cluster to confirm the installed state. On a non-Alauda OS, run them on a control plane node.

# All nodes are Ready
kubectl get nodes

# ClusterModule/global reports the base module as deployed and healthy
kubectl get clustermodule global -o jsonpath='{.status.base.deployStatus}{" "}{.status.base.runningStatus}{"\n"}'

# No AppRelease is in a failed state
kubectl get apprelease -A

# Core and Aligned modules report their current phase and version
kubectl get moduleinfo -l cpaas.io/cluster-name=global

# No Pod is stuck outside Running or Completed
kubectl get pod --all-namespaces | awk '{if ($4 != "Running" && $4 != "Completed")print}' | awk -F'[/ ]+' '{if ($3 != $4)print}'

The installation is healthy when:

  • All global cluster nodes are Ready.
  • ClusterModule/global prints Completed Healthy.
  • No AppRelease is in a failed state.
  • The following platform components are installed, and their corresponding ModuleInfo resources report Running at the target version:
    • Alauda Container Platform Web Console
    • Alauda Container Platform Web Console Central
    • Alauda Container Platform Marketplace Web Console
    • Alauda Container Platform Application Catalog
    • Alauda Container Platform Observability Essentials
    • Alauda Container Platform Essentials
    • Alauda Container Platform Cluster Essentials
    • Alauda Container Platform GatewayAPI Plugin
  • Critical Pods in cpaas-system are Running or Completed.

In the Web Console, open Administrator > Marketplace > Cluster Plugins and confirm that the required applications are available and none of them shows an installation failure.

Validate the Platform Image Source

Confirm that ProductBase records the expected platform image source:

kubectl get productbase base \
  -o jsonpath='{.spec.registry.address}{"\t"}{.spec.registry.external}{"\n"}'

For an external registry, the address is cpaas-registry.cpaas-system and the second value must be true. Confirm that this logical registry resolves to the prepared registry:

kubectl -n cpaas-system get registry cpaas-registry \
  -o jsonpath='{.spec.mode}{"\t"}{.spec.endpoints}{"\t"}{.status.phase}{"\n"}'

The expected output is mirror, the prepared registry address, and Ready.

For an authenticated external registry, confirm that the pull Secret exists without printing its contents:

kubectl -n cpaas-system get secret global-registry-auth

Confirm that the global cluster records the same registry address as ProductBase:

kubectl -n cpaas-system get cluster global \
  -o jsonpath='{.metadata.annotations.cpaas\.io/registry-address}{"\n"}'

ProductBase also contains catalog entries for optional or uninstalled Extensions. Do not expect every entry in ProductBase.status.artifacts to be Ready. An Absent entry for an unrelated Extension does not by itself indicate an installation failure.

Confirm that the required Extension catalog entries exist and that no Pod is blocked by an image pull failure:

kubectl get moduleplugins

kubectl get pods --all-namespaces \
  | awk '$4 ~ /ImagePullBackOff|ErrImagePull/'

The Pod command must produce no output. Accept the installation based on the required applications, their ModuleInfo and AppRelease state, and their running Pods. If a required application is installed but unavailable, follow Installation Troubleshooting to tell a missing manifest apart from DNS, routing, firewall, CA trust, pull credential, or registry availability problems.

Additional Checks for Alauda OS

If the global cluster runs Alauda OS, also verify the following with the kubeconfig that ac writes to <install-dir>/output/kubeconfig. On the bootstrap node, use <extract-dir>/res/<arch>/ac in place of kubectl.

When create returns, some platform components, such as Kube-OVN and the authentication components, can still be starting. ClusterModule/global can then briefly report a state other than Completed Healthy. This usually settles within a few minutes. Before you treat a check as failed, run the last command again until it prints Completed Healthy.

kubectl --kubeconfig <global-kubeconfig> get nodes
kubectl --kubeconfig <global-kubeconfig> get clusters.platform.tkestack.io global \
  -o jsonpath='{.status.phase}'
kubectl --kubeconfig <global-kubeconfig> get pods -n cpaas-system
kubectl --kubeconfig <global-kubeconfig> get clustermodule global \
  -o jsonpath='{.status.base.deployStatus}{" "}{.status.base.runningStatus}{"\n"}'

The installation is successful when all of the following conditions are true:

  • ac adm alpha global create has returned, with all eight stages complete. See Monitor the Installation.
  • All global cluster nodes are Ready.
  • Critical Pods in cpaas-system are Running or Completed.
  • ClusterModule/global prints Completed Healthy.

Next Step

After validation succeeds, continue with Next Steps. On Alauda OS, ac has already decommissioned the bootstrap cluster; see Decommission the Bootstrap Cluster.