kubernetes_ How to solve the problem that namespace cannot be deleted

scene ：Kubectl delete ns XXXX , namespace Always in Terminating state .

Use kubectl delete ns test --force=true, It can't be deleted .

  Cause analysis ： k8s Think of the current Namespace Inside , There are still unreleased resources

Actual observation ,delete After the execution of the command , Been waiting , Therefore, some resources have not been released

Solutions ：

1. View all resources under the namespace

kubectl api-resources -o name --verbs=list --namespaced | xargs -n 1 kubectl get --show-kind --ignore-not-found -n test

test All resources under the namespace are No Resources

Be careful ： If the resource under the namespace is not empty , Do something , Resources may continue to run on the cluster indefinitely .

2. take test Name space with json Export as a file

kubectl get namespace test -o json > test.json

3.  take spec Delete the content under ： The purpose of this step is to empty the content , With empty content ns Cover the original ns, inform k8s-api To delete ns Content is empty , The delete command cannot be blocked

4.  Add empty content ns Of namespace , adopt api-server Interface , cover k8s In the cluster

curl -k \
-H "Content-Type: application/json" \
-X PUT \
--data-binary @test.json \
http://127.0.0.1:8001/api/v1/namespaces/test/finalize

# because k8s The master node uses authentication , If you directly use the following command, the connection will be rejected

# Need to use kube-proxy Acting as agent

kubectl proxy --port=8081 

Be careful ： Use kubectl proxy The command does not use daemon mode , Once you Ctrl+c after , The agent is gone

The correct approach open  2 individual k8s Master node terminal , Execute... In a window kubectl proxy, Another window connects api-server

Thank you very much CSDN Blogger 「Jerry_Pan1990」 The original article of
Link to the original text ：https://blog.csdn.net/Jerry_Pan1990/article/details/103633627