This tool is a modification of https://github.com/jianz/k8s-reset-terminating-pv
USE IT WITH CAUTION AT YOUR OWN RISK
Reset namespace status from terminating back to active.
When delete a kubernetes namespace by accident, it may stuck in the terminating status due to some finalizer prevent it from being deleted. You can use this tool to reset its status back to active.
You can download the latest compiled binary from here.
If you prefer to compile by yourself:
git clone git@github.com:gmeghnag/k8s-reset-terminating-ns.git
cd k8s-reset-terminating-ns
go build -o resetns
For simplicity, you can name the etcd certificate ca.crt, etcd.crt, etcd.key, and put them in the same directory as the tool(resetns).
The tool by default connect to etcd using localhost:2379
. You can forward the etcd port on the pod to the localhost:
kubectl port-forward pods/etcd-member-master0 2379:2379 -n etcd
--k8s-key-prefix
: Default set to registry
for the community version of kubernetes as it uses /registry
as etcd key prefix, the key for namespace mynamespace is /registry/namespaces/mynamespace
. Set to kubernetes.io
for OpenShift as it uses /kubernetes.io
as prefix and the key for mynamespace is /kubernetes.io/namespaces/mynamespace
.
In the first terminal execute:
oc port-forward pods/$(oc get pods -n openshift-etcd -l app=etcd --field-selector="status.phase==Running" -o jsonpath="{.items[0].metadata.name}") 2379:2379 -n openshift-etcd
Then, in the second terminal execute:
oc get secret -n openshift-etcd etcd-client -o json | jq '.data."tls.crt"' -r | base64 -d > ca.crt
oc get secret -n openshift-etcd etcd-client -o json | jq '.data."tls.crt"' -r | base64 -d > etcd.crt
oc get secret -n openshift-etcd etcd-client -o json | jq '.data."tls.key"' -r | base64 -d > etcd.key
./resetns --k8s-key-prefix kubernetes.io <NAMESPACE_NAME>