failed to delete node

[lllicg]

Sealos Version

5.0.0-beta5

How to reproduce the bug?

1、run cmd sealos apply -f Clusterfile to create k8s. Clusterfile: apiVersion: apps.sealos.io/v1beta1 kind: Cluster metadata: name: default spec: env:

• criData=/data/cri/lib hosts:
• ips:
  • 10.19.193.224:22
  • 10.19.193.167:22
  • 10.19.193.161:22 roles:
  • master
  • amd64 ssh: passwd: xxxx1 user: root port: 22
• ips:
  • 10.19.193.153:22 roles:
  • node
  • amd64 ssh: passwd: xxxxx2 user: root port: 22 image:
• 10.21.10.10:5000/labring/kubernetes:v1.27.11
• 10.21.10.10:5000/labring/helm:v3.10.3
• 10.21.10.10:5000/labring/calico:v3.26.1

2、modify the Clusterfile and remove ip address 10.19.193.161, to delete node apiVersion: apps.sealos.io/v1beta1 kind: Cluster metadata: name: default spec: env:

• criData=/data/cri/lib hosts:
• ips:
  • 10.19.193.224:22
  • 10.19.193.167:22 roles:
  • master
  • amd64 ssh: passwd: xxxx1 user: root port: 22
• ips:
  • 10.19.193.153:22 roles:
  • node
  • amd64 ssh: passwd: xxxx2 user: root port: 22 image:
• 10.21.10.10:5000/labring/kubernetes:v1.27.11
• 10.21.10.10:5000/labring/helm:v3.10.3
• 10.21.10.10:5000/labring/calico:v3.26.1

3、run cmd sealos apply -f Clusterfile 2024-07-09T11:01:12 warn delete master 10.19.193.161:22 failed cannot get node with ip address 10.19.193.161:22: cannot find host with internal ip 10.19.193.161 2024-07-09T11:01:12 error failed to clean node, exec command if which kubeadm;then kubeadm reset -f -v 0;fi && \ rm -rf /etc/kubernetes/ && \ rm -rf /etc/cni && rm -rf /opt/cni && \ rm -rf /var/lib/etcd && (ip link delete kube-ipvs0 >/dev/null 2>&1 || true) failed, connect error: ssh: handshake failed: ssh: unable to authenticate, attempted methods [none], no supported methods remain 2024-07-09T11:01:12 error failed to clean node, exec command rm -rf $HOME/.kube failed, connect error: ssh: handshake failed: ssh: unable to authenticate, attempted methods [none], no supported methods remain 2024-07-09T11:01:12 info succeeded in deleting master 10.19.193.161:22 2024-07-09T11:01:12 info start to sync lvscare static pod to node: 10.19.193.153:22 master: [10.19.193.224:6443 10.19.193.167:6443] 10.19.193.153:22 2024-07-09T11:01:12 info generator lvscare static pod is success 2024-07-09T11:01:12 info Executing pipeline UndoBootstrap in ScaleProcessor 2024-07-09T11:01:12 error Applied to cluster error: failed to execute remote command /var/lib/sealos/data/default/rootfs/opt/sealctl hosts delete --domain sealos.hub: connect error: ssh: handshake failed: ssh: unable to authenticate, attempted methods [none], no supported methods remain Error: failed to execute remote command /var/lib/sealos/data/default/rootfs/opt/sealctl hosts delete --domain sealos.hub: connect error: ssh: handshake failed: ssh: unable to authenticate, attempted methods [none], no supported methods remain

What is the expected behavior?

No response

What do you see instead?

No response

Operating environment

- Sealos version:v5.0.1
- containerd version:v1.7.14
- Kubernetes version:1.27.11

Additional information

No response

[bxy4543]

It looks like the machine is no longer accessible

[yangxggo]

It looks like the machine is no longer accessible

no, the target node that needs to delete is not existing in the new clusterfile, and it has a different password with others'. In that case, sealos cannot get the right password to connect the target node.

[yangxggo]

/assign