使用Kind删除Kubernetes集群时遭遇只读文件系统错误求助

我用Kind测试Kubernetes，创建了名为nodes-test的集群：

vagrant@vagrant:~$ sudo kind get clusters
nodes-test

执行sudo kind delete cluster删除集群时报错：

Deleting cluster "kind" ...
failed to update kubeconfig: failed to lock config file: open /root/.kube/config.lock: read-only file system
ERROR: failed to delete cluster "kind": failed to lock config file: open /root/.kube/config.lock: read-only file system

但检查对应路径并未找到该文件：

vagrant@vagrant:~$ sudo su
root@vagrant:/home/vagrant# cd /root
root@vagrant:~# cd .kube/
root@vagrant:~/.kube# ls -la
total 20
drwxr-xr-x 3 root root 4096 Sep  8 12:07 .
drwx------ 6 root root 4096 Sep  8 12:26 ..
drwxr-x--- 4 root root 4096 Sep  8 12:00 cache
-rw------- 1 root root 5622 Sep  8 11:59 config

kubeconfig文件内容：

apiVersion: v1
clusters:
- cluster:
    certificate-authority-data: ss==
    server: https://127.0.0.1:44155
  name: kind-nodes-test
contexts:
- context:
    cluster: kind-nodes-test
    user: kind-nodes-test
  name: kind-nodes-test
current-context: kind-nodes-test
kind: Config
preferences: {}
users:
- name: kind-nodes-test
  user:
    client-certificate-data: sss
    client-key-data: sdasd=

当前网络监听状态：

Netid State  Recv-Q Send-Q  Local Address:Port   Peer Address:PortProcess
udp   UNCONN 0      0       127.0.0.53%lo:domain      0.0.0.0:*    users:(("systemd-resolve",pid=10281,fd=13))
udp   UNCONN 0      0      10.0.2.15%eth0:bootpc      0.0.0.0:*    users:(("systemd-network",pid=10265,fd=18))
tcp   LISTEN 0      4096    127.0.0.53%lo:domain      0.0.0.0:*    users:(("systemd-resolve",pid=10281,fd=14))
tcp   LISTEN 0      128           0.0.0.0:ssh         0.0.0.0:*    users:(("sshd",pid=711,fd=3))
tcp   LISTEN 0      4096        127.0.0.1:44155       0.0.0.0:*    users:(("docker-proxy",pid=1109,fd=4))
tcp   LISTEN 0      128              [::]:ssh            [::]:*    users:(("sshd",pid=711,fd=4))          

更新情况 #

指定集群名称执行sudo kind delete cluster --name nodes-test仍报错：

vagrant@vagrant:~$ sudo kind delete cluster --name nodes-test
Deleting cluster "nodes-test" ...
failed to update kubeconfig: failed to lock config file: open /root/.kube/config.lock: read-only file system
ERROR: failed to delete cluster "nodes-test": failed to delete nodes: command "docker rm -f -v nodes-test-worker nodes-test-control-plane" failed with error: exit status 1
Command Output: Error response from daemon: container 4cf8b399150fffb9899ff963adacd255fe3b0b7aebdf6feda11ea1646b541340: driver "overlay2" failed to remove root filesystem: unlinkat /var/lib/docker/overlay2/1ba0d047b313dabc396f61cf1f2b3086d5cad05372de37410326d81e02f6130b: read-only file system
Error response from daemon: container 8b2fa50dc006191add9b64e89b834a0c952344433e61d1b9795b5c2470bf3e05: driver "overlay2" failed to remove root filesystem: unlinkat /var/lib/docker/overlay2/3a35de9a7ad18fbf3358c0be3d01a51ba7c7558ed49242d659a0e365c58798ae: read-only file system

1. 修复只读文件系统挂载 #

只读挂载是核心问题，先解除只读状态：

• 找出系统中的只读挂载分区：

  mount | grep ro
  

• 将对应分区重新挂载为读写模式（示例为根分区，根据实际输出调整）：

  sudo mount -o remount,rw /
  

• 如果/var/lib/docker是单独挂载的分区，同样执行：

  sudo mount -o remount,rw /var/lib/docker
  

2. 手动清理Kind集群资源 #

自动删除失败时，手动清理所有关联资源：

（1）删除集群容器 #

直接用Docker删除集群的节点容器：

sudo docker rm -f nodes-test-worker nodes-test-control-plane

如果集群有更多节点，将所有节点容器名称加入命令。

（2）清理Docker镜像与卷 #

删除Kind使用的节点镜像（可选，释放存储空间）：

sudo docker rmi -f kindest/node:latest

清理无关联的Docker卷：

sudo docker volume prune -f

（3）手动更新kubeconfig #

编辑/root/.kube/config，删除以下内容：

• clusters段中name: kind-nodes-test的条目
• contexts段中name: kind-nodes-test的条目
• users段中name: kind-nodes-test的条目
• 若current-context为kind-nodes-test，可将其改为其他存在的上下文或删除该行

保存文件后，kubeconfig即完成清理。

3. 验证清理结果 #

检查集群是否已移除：

sudo kind get clusters

确认Docker容器已删除：

sudo docker ps -a

内容的提问来源于stack exchange，提问作者user63898