BUG: deploy cluster not using internal registry

[LiaoSirui]

Sealos Version

v4.3.7

How to reproduce the bug?

deploy cluster using

sealos run \
    \
    -e criData=/var/lib/containerd \
    -e registryData=/var/lib/registry \
    -e registryDomain=platform-local.hub \
    -e registryPort=5000 \
    \
    docker.io/labring/kubernetes:v1.28.7-4.3.7 \
    docker.io/labring/cilium:v1.14.7 \
    docker.io/labring/helm:v3.14.1 \
    docker.io/labring/nerdctl:v1.7.0 \
    \
    --port=22 \
    --user='root' \
    --pk='/root/.ssh/id_rsa' \
    \
    --debug=true \
    \
    --masters 10.244.244.20

What is the expected behavior?

deploy cluster using internal registry

What do you see instead?

[preflight] Pulling images required for setting up a Kubernetes cluster [preflight] This might take a minute or two, depending on the speed of your internet connection [preflight] You can also perform this action in beforehand using 'kubeadm config images pull' I0305 21:05:34.300215 107333 checks.go:828] using image pull policy: IfNotPresent I0305 21:05:34.332485 107333 checks.go:854] pulling: registry.k8s.io/kube-apiserver:v1.28.7

Operating environment

- Sealos version: v4.3.7
- Docker version:
- Kubernetes version: none
- Operating system: rocky linux 9.3
- Runtime environment: none
- Cluster size: single node
- Additional information:

Additional information

No response

[LiaoSirui]

maybe same: https://github.com/labring/sealos/issues/4514

[cuisongliu]

using 5.x的sealos

[sealos-ci-robot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

sealos using 5.x

[caijian76]

使用sealos 5可以初始化安装k8s v1.28，但是从v1.27升级到v1.28，依然遇到相同的问题。

[sealos-ci-robot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

You can initialize and install k8s v1.28 using sealos 5, but when you upgrade from v1.27 to v1.28, you still encounter the same problem.