Delete the global Cluster on Huawei DCS

destroy deletes the global cluster and the Huawei DCS resources that it created, with the credential stored in the global-credential Secret. It refuses while the platform still manages another cluster; for this and the other rules every provider shares, see The destroy Command.

Before You Begin

  • Delete every workload cluster that the platform manages.
  • Detach on DCS any volume attached to the global VMs that you need to keep.
  • Use a machine that reaches the global cluster and the DCS VRM on TCP/7443, such as the bootstrap node. destroy signs in to DCS: repeated failed sign-ins lock the account.

Procedure

On that machine:

<extract-dir>/res/<arch>/ac adm alpha global destroy --kubeconfig <install-dir>/output/kubeconfig

destroy lists each item, then asks you to type the cluster name. Before it stops the VMs, it scales the DCS provider on the global cluster (cpaas-system/cluster-api-provider-dcs-manager) to 0, so that the provider does not start them again. It deletes:

  • Every VM of the cluster, with its system and ephemeral disks.
  • The persistent and raw disks of the DCSIpHostnamePool objects, after detaching them.
  • The DRS rule of controlPlaneHA, and leftover CD-ROM ISOs.

Other volumes attached to the VMs are deleted with the VM unless they are independent; an independent volume is left behind unattached.

It keeps the folders, datastores, switches and port groups, VM templates, IP addresses, and the control plane VIP. A persistent disk that is attached to a VM of another cluster is kept and listed.

If destroy stops, run the same command again; see The destroy Command.