CKS node Instance stuck in "Starting" state (VM lifecycle) when a non-default (explicitly selected) template is used; works with the default SystemVM template

### problem

When creating a CKS Kubernetes cluster and explicitly selecting a non-default node
template (the stock cks-ubuntu-2204-kvm CKS-ready image) in Advanced Settings, the
node Instance stays in the Starting VM lifecycle state indefinitely and never
transitions to Running — even though the libvirt domain is actually up on the KVM host
(the VNC console is reachable and the guest OS boots).

The same cluster creation succeeds when the default SystemVM template
(systemvm-kvm-4.22.0-x86_64) is used for the nodes. The only changed variable between
the working and failing cases is the node template selection, which points at the
Flexible Kubernetes Clusters per-node template selection path rather than at
in-guest provisioning (cloud-init / kubeadm / CNI).

STEPS TO REPRODUCE

1. ACS 4.22.1.0 on KVM, CKS enabled.


2. Register the stock CKS-ready Ubuntu 22.04 KVM template (cks-ubuntu-2204-kvm),
   marked "For CKS".


3. Register a supported Kubernetes binaries ISO/version.


4. Create a CloudManaged Kubernetes cluster (3 control node HA) on a VPC tier network,
   and in Advanced Settings explicitly select cks-ubuntu-2204-kvm as the node
   template.


5. Observe the control node Instance in the UI (Instances → the control node):
   it stays in Starting.


6. Repeat the exact same cluster creation but do not select a template (use the
   default systemvm-kvm-4.22.0-x86_64): the node reaches Running and the cluster
   provisions normally.

EXPECTED RESULTS

Selecting a "For CKS" template in Advanced Settings should behave the same as the
default template path: the node Instance transitions Starting → Running within the
normal VM start window, after which CKS provisioning proceeds and the cluster reaches
Running.

ACTUAL RESULTS

Behaviour depends solely on the node template selection:

In the failing case:

• The control node Instance remains in Starting (VM lifecycle state) indefinitely, so
  the cluster also stays in Starting.


• The libvirt domain is genuinely running on the KVM host: the VNC console is reachable
  and the guest OS boots.


• SSH to the node (port 2222) is not usable; the UI-displayed password is not accepted
  on the VNC console (consistent with the node never being marked Running and/or the
  start/provisioning workflow not completing).

### versions

Hypervisor version: KVM Ubuntu 24.04
ACS version: 4.22.1.0
Management server: Ubuntu 24.04

### The steps to reproduce the bug

_No response_

### What to do about it?

_No response_

首先跟踪 Flexible Kubernetes Clusters 的逐节点模板选择路径，以及 Instances UI 使用的 VM 生命周期转换，并将显式选择的 cks-ubuntu-2204-kvm 模板与默认的 SystemVM 模板进行比较。在 KVM 上复现，并检查 libvirt 域状态和节点启动工作流。当选定的 CKS 模板达到 Running，且集群配置过程像使用默认模板时一样继续进行，即表示完成。

由索引模型根据 Issue 内容生成。