./package.sh config.yaml报错Kubekey 镜像导出失败 - quay.io/tigera/operator 下载时出现500错误

[wenzhaojie]

What is version of KubeKey has the issue?

v4.0.3

What is your os environment?

ubuntu22.04

KubeKey config file

apiVersion: kubekey.kubesphere.io/v1
kind: Config
spec:
  cni:
    calico_version: v3.29.7
    cilium_version: 1.18.5
    hybridnet_version: 0.6.8
    kubeovn_version: v1.15.0
    flannel_version: v0.27.4
    multus:
      image:
        tag: v4.2.1
  cri:
    container_manager: containerd
    containerd_version: v1.7.13
    crictl_version: v1.30.0
    cridockerd_version: v0.3.21
    docker_version: 25.0.5
    runc_version: v1.1.12
  dns:
    dns_cache_image:
      tag: 1.22.28
    dns_image:
      tag: v1.11.3
  etcd:
    etcd_version: v3.5.15
  download:
    arch:
      - amd64
      - arm64
  image_registry:
    docker_registry_version: 2.8.3
    dockercompose_version: v2.20.3
    harbor_version: v2.10.2
    keepalived_version: 2.0.20
  kubernetes:
    sandbox_image:
      tag: "3.9"
    helm_version: v3.13.3
    kube_version: v1.30.14
    control_plane_endpoint:
      kube_vip:
        image:
          tag: v0.7.2
      haproxy:
        image:
          tag: 2.9.6-alpine
  storage_class:
    localpv_provisioner_version: 4.4.0
    nfs_provisioner_version: 4.0.18
  image_manifests:
  # kubernetes-v1.30.14
  - quay.io/tigera/operator:v1.36.16
  - docker.io/calico/apiserver:v3.29.7
  - docker.io/calico/cni:v3.29.7
  - docker.io/calico/ctl:v3.29.7
  - docker.io/calico/csi:v3.29.7
  - docker.io/calico/kube-controllers:v3.29.7
  - docker.io/calico/node-driver-registrar:v3.29.7
  - docker.io/calico/node:v3.29.7
  - docker.io/calico/pod2daemon-flexvol:v3.29.7
  - docker.io/calico/typha:v3.29.7
  - registry.k8s.io/coredns/coredns:v1.11.3
  - registry.k8s.io/dns/k8s-dns-node-cache:1.22.28
  - registry.k8s.io/kube-apiserver:v1.30.14
  - registry.k8s.io/kube-controller-manager:v1.30.14
  - registry.k8s.io/kube-proxy:v1.30.14
  - registry.k8s.io/kube-scheduler:v1.30.14
  - registry.k8s.io/pause:3.9
  - docker.io/openebs/linux-utils:4.3.0
  - docker.io/openebs/provisioner-localpv:4.4.0
  - docker.io/library/haproxy:2.9.6-alpine
  - docker.io/plndr/kube-vip:v0.7.2
  # ks-core
  - registry.cn-beijing.aliyuncs.com/kse/ks-apiserver:v4.2.0-20251118
  - registry.cn-beijing.aliyuncs.com/kse/ks-console:v4.2.0-20251118
  - registry.cn-beijing.aliyuncs.com/kse/ks-controller-manager:v4.2.0-20251118
  - registry.cn-beijing.aliyuncs.com/kubesphereio/kubectl:v1.33.1
  - registry.cn-beijing.aliyuncs.com/kubesphereio/redis:7.2.4-alpine
  - registry.cn-beijing.aliyuncs.com/kubesphereio/redis:7.2.8-alpine
  - registry.cn-beijing.aliyuncs.com/kubesphereio/haproxy:3.0.8-alpine
  - registry.cn-beijing.aliyuncs.com/kse/extensions-museum:v11.1.2
  - registry.cn-beijing.aliyuncs.com/kse/ks-console-embed:v4.2.0-20251118
  - registry.cn-beijing.aliyuncs.com/kse/ks-posthog:v2.0.0
  # metrics-server
  - docker.io/kubesphere/metrics-server:v0.7.2
  - docker.io/kubesphere/metrics-server-frontend:v1.0.1
  - docker.io/kubesphere/ks-autoscaling-controller:v1.0.0

A clear and concise description of what happend.

在使用Kubekey的package.sh命令导出集群镜像时，在下载容器镜像步骤中失败。具体错误发生在尝试拉取镜像 quay.io/tigera/operator:v1.36.16 时，quay.io返回500内部服务器错误。

完全按照官网教程来进行，而且已检查网络条件完好（日本AWS EC2新实例）

Relevant log output

Additional information

⠸ [localhost] failed [2s]
10:19:17 UTC [Playbook default/artifact-export-ds6rg] finish. total: 36,success: 33,ignored: 2,failed: 1
Error: task Image | Download container images run failed:
[localhost][executor]: module run failed
[localhost][item=][0]: failed to pull image "quay.io/tigera/operator:v1.36.16" to local dir: failed to push manifest for linux/amd64: failed to copy config: failed to push blob sha256:b0e588a29e9081b400bac0a0ea17b6eab09bb8428d79ec5a9b1eb246606bc175: PUT "https://quay.io/v2/tigera/operator/blobs/uploads?digest=sha256%3Ab0e588a29e9081b400bac0a0ea17b6eab09bb8428d79ec5a9b1eb246606bc175": response status code 500: Internal Server Error
task Image | Download container images run failed:
[localhost][executor]: module run failed
[localhost][item=][0]: failed to pull image "quay.io/tigera/operator:v1.36.16" to local dir: failed to push manifest for linux/amd64: failed to copy config: failed to push blob sha256:b0e588a29e9081b400bac0a0ea17b6eab09bb8428d79ec5a9b1eb246606bc175: PUT "https://quay.io/v2/tigera/operator/blobs/uploads?digest=sha256%3Ab0e588a29e9081b400bac0a0ea17b6eab09bb8428d79ec5a9b1eb246606bc175": response status code 500: Internal Server Errorubuntu@ip-172-31-14-40:~/kubekey$ ./kk version
{"major":"4","minor":"0","gitVersion":"v4.0.3","gitCommit":"fd08c8caab8116f87810b5cde3313a7cbacc53fb","gitTreeState":"clean","buildDate":"2026-01-20T10:40:44Z","goVersion":"go1.25.0","compiler":"gc","platform":"linux/arm64"}

[redscholar]

节点连不上 quay.io 吧

[wenzhaojie]

节点连不上 quay.io 吧

手动docker pull 都能下。而且我是日本aws节点。网络环境不是问题。

[wenzhaojie]

节点连不上 quay.io 吧

@redscholar 我也按照官网教程在配置c
cat >> config.yaml << 'EOF'
cri:
registry:
auths:
- repo: <镜像仓库地址> # 拉取镜像时需访问的镜像仓库，如 docker.io
username: <镜像仓库账号> # 如 docker.io 的登录账户
password: <镜像仓库密码> # 如 docker.io 的登录密码
EOF

可是还是出现报错。请问您知道怎么解决吗？

[zuoxuesong-worker]

我在测试时也发现过这个问题，通过oras的客户端拉取到了镜像后保存至本地文件系统时出现的错误，但是并不能稳定复现，出现概率非常低。请问你是能够稳定复现这个问题吗？